Month: July 2024

What to do when Windows won’t boot

When Windows won’t boot up, it can be a disturbing experience. But that moment passes quickly, and you must decide how to get past this (hopefully) temporary hiccup.

Fortunately, there are plenty of ways to tackle this. They do vary in the time and effort required to implement, and in the severity of their potential impact on your PC.

Let’s talk about them, then work through the options and activities involved, along with a strategy for approaching such repair. Throughout, there’s special emphasis on Startup Repair and Boot Recovery tools, facilities and commands available in Windows 10 and 11.

What do you see?

When Windows tries but fails to boot, there are several possible displays you might see. These include a “black screen,” which essentially means nothing is on display at all.

win10 black screen cursor

The dreaded “Windows black screen with cursor.”

Ed Tittel/IDG

If you see a black screen

Should you see a black screen, check your PC indicator lights to make sure the device is still powered on. Sometimes when Windows shuts down, it can automatically kill the power, too. In such circumstances, the best possible outcome is powering up, followed by a normal boot sequence all the way into the Windows desktop.

For a black screen with the power on, things in Windows get more interesting. Even so, it’s worth cycling the power and trying again before attempting other repairs. Often, things will return to normal on their own. If you wind up with a second black screen, see my companion story “How to fix a Windows 10 black screen” for more tips, which work as well for Windows 11 as they do for Windows 10.

If you see something else: NOT a black screen

As Windows starts up, it uses a special program called a boot loader to start the process of taking over a PC, before it hands over control to the operating system. If enough “smarts” are present when a problem occurs during boot-up, users might see what’s called a “stop error” or a “blue screen.” Though the color can vary in recent versions of Windows, blue was a constant backdrop for such errors from the earliest Windows versions through Windows 7 and became enshrined in Windows folklore as the “Blue Screen of Death,” a.k.a. BSOD, as shown in Figure 1.

windows blue screen of death for inaccessible boot device error

Figure 1: BSOD example for INACCESSIBLE_BOOT_DEVICE.

Ed Tittel / IDG

Figure 1 shows a BSOD for what might be called a basic no-boot scenario: it’s associated with Stop Code 7B (a hexadecimal number) otherwise labeled INACCESSIBLE_BOOT_DEVICE. In other words, it shows up when the boot loader recognizes that it cannot access the storage device from which it would normally load the Windows OS. What more basic cause for boot failure could there be? None!

On the other hand, if the Windows boot loader can’t load the OS but it can find a bootable Windows Recovery Environment partition, it should load a warning message like the one shown in Figure 2.

windows recovery windows load error

Figure 2: I forced this error, which would normally read “Windows didn’t start correctly.”

Ed Tittel / IDG

Try WinRE Startup Repair

Select See advanced repair options at lower center right in Figure 2. This displays the root-level Windows Recovery Environment (a.k.a. WinRE), as shown in Figure 3.

windows recovery environment menu

Figure 3: This is the root-level menu for the Windows Recovery Environment. We’ll move on to the Troubleshoot selection.

Ed Tittel / IDG

Boot repairs may be handled under the Troubleshoot option in Figure 3. (If you’ve prepared a bootable USB drive with repair tools, you could instead elect to boot to that by selecting Use a device.) When you click Troubleshoot, you’ll see the screen shown in Figure 4.

winre advanced options menu including startup repair startup settings command prompt uninstall updates uefi firmware settings and system restore

Figure 4: Advanced options appear when you select the Troubleshoot option in Figure 3. Note that the first item reads Startup Repair.

Ed Tittel / IDG

Click the Startup Repair button at upper left. Your PC immediately reboots. Against a black screen, the spinning balls appear over a legend that reads “Diagnosing your PC.” In the background, WinRE is running an automated set of startup checks and (where applicable) making repairs.

You’ll either wind up with a bootable desktop, or the UI will pop up a message that reads “Startup Repair couldn’t repair your PC,” as shown in Figure 5.

winre startup repair message saying repair failed

Figure 5: When Startup Repair fails (and it often does), it reports accordingly.

Ed Tittel / IDG

Failure happens more often than most of us would like. Indeed, I have occasionally seen WinRE’s automated Startup Repair work, but perhaps in only one out of three or four attempts.

Not to fear: there are several more repair strategies to try when Windows won’t boot. I’ll go over those in a moment — but first, it’s useful to understand Windows boot mechanisms and its boot configuration data (aka BCD).

Windows boot and BCD explained

Windows 11 creates at least two boot structures on any media (usually SSD or hard disk) from which Windows will boot. By convention, the disk where two special programs reside, known as the Windows Boot Manager and the Windows Boot Loader, is called a “boot disk.” If a runnable version of Windows also resides on that same drive, it’s called a “boot/system disk.”

The boot loader resides in a disk partition allocated for the Basic Input-Output System (BIOS) or Unified Extensible Firmware Interface (UEFI), the software that kicks off system start-up, to use when the system is initially booting itself up. After it goes through hardware and security checks, including device enumeration, the BIOS or UEFI hands over control to the boot loader, which starts the process of reading OS boot information from the Boot partition on the disk.

It then hands over control to the Windows Boot Manager, which takes over the process of starting up Windows and making it ready to run for user login and application support. The Windows Boot Manager is also responsible for handling what’s called Boot Configuration Data (BCD), on the system’s behalf.

When Windows 10 or 11 gets installed, it also creates a separate Recovery Partition at the tail end of the disk, where it keeps a bootable version of the Windows Recovery Environment (a.k.a. WinRE) that can take over if the Windows system/boot partition becomes inaccessible. This is depicted in the layout of the C: drive from a test PC in Figure 6, where EFI, system/boot, and recovery partitions follow one another from left to right.

Ed Tittel / IDG

If you look at the defaults for boot configuration data in Windows 10 or 11, you’ll see information about the Windows Boot Manager and the Windows Boot Loader appear (from the BCDedit command, discussed in the later section “Rebuilding Windows BCD”). Figure 7 shows the results of running the bcdedit command by itself, which displays active boot information currently known to the boot manager.

Ed Tittel / IDG

For the purposes of our discussion in this story, it’s enough to understand that boot configuration data describes where the system should look to find the programs it needs to boot the system. Hopefully, it’s also obvious that this data is essential to starting up Windows. Thus, corruption, damage, and invalid entries in this data can (and does) result in an unbootable Windows installation. Fixing that is what we’re about here.

Repair strategies when Windows won’t boot

While you can sometimes use WinRE to repair Windows boot problems, that may not be the fastest or easiest way to fix things. In my experience, I’ve had better luck doing any or all of the following when Startup Repair didn’t fix boot issues:

  • Restoring a known, good, working OS image (e.g., using Macrium Reflect or a similar disk image tool restore operation).
  • Using a third-party boot repair tool (e.g., Macrium Rescue Media “Fix Windows boot problems” or similar boot repair tool).
  • Rebuilding the Windows Boot Configuration Data (BCD) from the command line in WinRE.

Please note: I present these options in their recommended order for the techniques covered. That’s because of the time, effort, and complexity they involve or entail. This is almost the reverse of what readers may expect. That’s because it starts from a non-boot-related approach, then goes on to explain third-party boot repair tools, and only then concludes with built-in Microsoft commands.

All this assumes you’ve tried the Startup Repair option from Figure 4 without fixing your Windows 11 boot issue, whatever it may be.

1. Restoring a known, good working image

If you’ve got a recent image backup for the non-booting PC, you can boot to its restore tool. Then you can restore that backup to the target disk. Such a restore replaces the entire disk image, including boot configuration data, disk layout, and all contents. As long as the drive itself is working, this is the fastest, safest, and surest way to resolve boot problems I know. You can read all about how to create and restore a Windows image backup in my story “How to make a Windows 10 or 11 image backup.”

Each of the image backup packages covered in that story can create its own bootable media, able to restore backups from some other drive (and often includes boot repair tools as well, as you’ll see in the next section). Those tools, including links to tutorials on creating them, are:

If you’ve been using the Windows Backup facility in Windows 11 (see my story “A new Windows 11 backup and recovery paradigm?” for info on how to use this), you will be able to access copies of files from key folders (and more) that have changed since your most recent image backup. This provides the best of both worlds when it comes to restoration, because an image backup provides immediate access to all your installed apps and applications, while Windows Backup provides access to key recent files via OneDrive.

2. Using a third-party boot repair tool

Most image backup tools mentioned in the previous section include boot repair facilities (MiniTool relies on its free Partition Wizard for repairing Windows boot errors instead). Other notable tools include those covered in the April 2024 Lifewire story “10 Best Free Disk Partition Software Tools,” many of which include boot repair facilities as part and parcel of their partition management capabilities á la MiniTool Partition Wizard (MTPW).

Personally, I’ve used the “Fix Windows Boot Problems” item from the Macrium Rescue Media on many occasions to tackle Windows boot problems. As long as the underlying drive was still working (it can’t fix failing hardware, alas), it has always been able to restore a working Windows boot environment when asked to do so.

I’ve also used the MTPW tool on multiple occasions, and it, too, has shown itself effective. Online reports and forum threads for the other tools mentioned here and in the preceding section indicate that they also enjoy positive ratings from their users.

If you don’t have a current backup to restore, or can’t restore such a backup for some reason, try one or more of these boot repair tools before you move to the Command Prompt in WinRE, as described in the next section. They will often fix whatever ails your Windows boot environment.

3. Rebuilding Windows BCD in WinRE

In the Windows environment, Boot Configuration Data (BCD) identifies programs used to boot the OS, and related settings (configuration) data. When Windows is running, the command line tool of choice for such information is BCDedit. But when you’re running inside the Windows Recovery Environment, having booted from bootable Windows install media (or some equivalent, such as the Microsoft Diagnostics and Repair Toolkit, a.k.a. DaRT), the tool of choice is bootrec.exe because it works on BCD data for the broken Windows image (that is, the one on your system/boot disk that isn’t currently working).

To run this command, select the Command Prompt option shown in Figure 4, then type the bootrec.exe command at the command line, using one of the options described in the next paragraph.

Interestingly, Microsoft’s Bootrec.exe support files haven’t been updated since the days of Windows 7. Even so, they’re still reasonably accurate for Windows 10 and 11. The following options still work in both:

  • /FixMBR: Writes a new BCD store to the system partition, without overwriting existing partition table data. This option can address boot corruption issues, especially when the boot loader can’t read or interpret available BCD info.
  • /FixBoot: Writes a new boot sector to the system partition using a boot sector compatible with the OS in use. This option helps address improper or invalid BCD changes, boot sector damage or corruption, or changes imposed when installing an older OS after a newer one was installed.
  • /ScanOS: Scans all disks for installations compatible with the current OS. Displays all boot sector entries it finds, including those not currently residing in the BCD store. This option is intended to pick up installations not showing in the Boot Manager menu.
  • /RebuildBcd: Scans all disks for installations compatible with current OS. Allows users to select an installation to add to the BCD store. Also rebuilds the BCD store from scratch.

The most common bootrec invocation is to instruct it to rebuild the BCD store — namely:

Bootrec.exe /RebuildBCD

(Note: the .exe is optional.)

If this technique does not result in a bootable Windows, follow the instructions at the end of the Microsoft support page to export and delete the BCD store, then rebuild that store anew. This usually works.

The makers of Ventoy, a terrific bootable ISO management tool, offer a tutorial called “How to Rebuild BCD in Windows Easily.” It even includes detailed bootrec instructions, especially Section 2, “Using the Command Prompt.” It walks through an illustrated version of the same instructions found at the tail end of the aforementioned support page.

You can learn a lot about the way that bootrec.exe actually works by digging into Microsoft’s detailed (and better documented) BCDedit command reference info (see also BCDedit Command-Line Options). Hopefully, you’ll have an image backup of your problem drive handy so you can always restore same should command-line repairs go off the rails.

Getting past the finish line

Ultimately, you’ll find yourself in one of two places. First and best, you’ll restore Windows to working order, including a proper boot. Second and less favorably, you’ll be stuck going nowhere with no boot in sight.

Should this happen, you’ll have to decide whether or not you want to scrub the existing Windows installation and start over. (In the most dire circumstances, this could mean replacing the boot/system drive that simply won’t boot despite all efforts to fix it.) On the other hand, it may be time to consider taking the PC into the shop to get on a professional bench for repair or replacement as their findings dictate.

In my 30-plus years of working with Windows, drive failure has come up twice. In both cases, the drive that wouldn’t boot wasn’t working and needed replacement. If this is a task you can comfortably handle (it’s something I routinely take care of for my fleet of 12-20 PCs), it’s neither terribly difficult nor time-consuming.

And again: if you have a recent backup, you can usually restore that to a new drive the same way you would work with the existing one. Remember: where there’s a will, there’s a way. Good luck!

What is UEM? Unified endpoint management explained

Unified endpoint management (UEM) describes a set of technologies used to secure and manage a wide range of employee devices and operating systems — all from a single console.

Seen as the next generation of mobility software, UEM tools incorporate several existing enterprise mobility management (EMM) technologies — including mobile device management (MDM) and mobile application management (MAM) — with some of the tools used to secure desktop PCs and laptops.

“UEM in theory ties this all together and gives you that proverbial one pane of glass, so you can see the state of all of your endpoints,” said Phil Hochmuth, program vice president at IDC. “It gives you visibility into what people are doing with corporate data, corporate apps, on any conceivable type of device.”

The ability to manage various device types in one place is increasingly important as businesses face a growing cybersecurity threat, said Tom Cipolla, senior director analyst at Gartner. “We need to patch faster; everybody acknowledges that,” he said. “UEM gives people a consolidated view into their environment and a consolidated patching and configuration management approach.”

The evolution of mobile management – MDM, MAM, and more

At its core, UEM consists of several device management technologies that emerged to help businesses control employee mobile devices. The first iteration of such tools was MDM, which arrived about a decade ago.

Introduced in response to the initial wave of smartphones used in the workplace, MDM was designed to help IT centrally provision, configure, and manage mobile devices that had access to corporate systems and data. Common MDM features included security configuration and policy enforcement, data encryption, remote device wipe and lock, and location tracking.

However, as employee bring-your-own-device (BYOD) schemes became more prevalent in the office — driven first by the iPhone’s popularity, later by the growth of Android — vendors began to offer more targeted management of apps and data. MAM capabilities delivered more granular controls, focusing on software rather than the device itself; features include app wrapping and containerization, and the ability to block copy/paste or restrict which apps can open certain files.

MAM features were soon packaged with MDM and other tools, such as mobile identity management and mobile information management, and sold as comprehensive enterprise mobility management (EMM) product suites. Those suites led to the next stage in the evolution of device management: UEM.

What is UEM?

UEM merges the various facets of EMM suites with functionality typically found in client management tools (CMT) used to manage desktop PCs and laptops on a corporate network. One example is Microsoft’s Intune, which combined its MDM/MAM platform with Configuration Manager (formerly System Center Configuration Manager) in 2019.

UEM platforms tend to have comprehensive operating system support, including mobile (Android, iOS) and desktop OSes (Windows 11, macOS, ChromeOS, and, in some cases, Linux). Some UEM products support more esoteric categories too, including IoT devices, AR/VR headsets, and smartwatches.

Unlike traditional CMT products, UEM tends to be available as a software-as-a-service, cloud-based tool, allowing management and updates of devices such as desktop PCs without connection to a corporate network. 

The emergence of UEM has been partly driven by the inclusion of API-based configuration and management protocols within Windows and macOS, enabling the same level of device management that was already possible with iOS and Android devices.

It speaks to a wider development, too, of the convergence of mobile and traditional computing devices, with high-end tablets often on par with laptops in terms of processing power. “You have a real blurring of the lines between what is mobile computing and what is traditional endpoint computing,” said Hochmuth.

Why invest in UEM tools?

All of these devices — mobile, desktop, Windows, Mac, in the office and remote — require a unified approach to end user device management, an approach that can provide a variety of benefits, say analysts.

Among these is the opportunity for simplified and centralized management. In short, it’s more efficient for one team to provision and manage all devices from a single tool, rather than have separate support teams and tools that were traditionally divided between mobile and Windows or macOS computers. 

“If you have a separate software product or management platform for four different operating systems, that can be cumbersome and expensive,” said IDC’s Hochmuth. “Converging down to one or two is a goal for a lot of organizations.”

UEM products can reduce manual work for IT, with the ability to create a single policy — such as requiring device encryption — that can be deployed to many devices and operating systems. The same goes for patching.  

By ensuring consistent policies across apps, devices and data, UEM tools can reduce risk, with less complexity and fewer opportunities to misconfigure policies. 

There are cost benefits in replacing separate PC and mobile management applications too. “Getting rid of one software platform and all the licensing associated with that is a cost saving. That’s not the primary driver, but it’s definitely a reason to explore UEM,” said Hochmuth. 

The UEM vendor market

The global market for unified endpoint management software is forecast to grow from $5.9 billion in 2023 to $8.9 billion in 2028, according to IDC data. The rate of yearly growth is set to slow, however, from around 16% to 6% during this period. 

There are a variety of vendors, from big-name firms to smaller, more targeted companies. Microsoft (Intune) and VMware/Broadcom (Workspace One) are often considered the UEM market leaders with the broadest offerings and largest market share by revenue. BlackBerry UEM, Citrix Meraki Systems Manager, IBM MaaS360, ManageEngine, Cisco, and Ivanti UEM are also popular products.

“All these companies have roles or verticals or use cases that they address specifically,” said Hochmuth. For instance, BlackBerry is often viewed as strong in regulated markets, such as finance or healthcare, due its focus on encryption, while Microsoft has a more of a “horizontal” product with general business use cases.  

Among the vendors that have taken a more specialized approach is Jamf, which is focused purely on Apple devices running everything from macOS to tvOS, and SOTI, whose products are tailored to certain industries, such as warehouse workers with ruggedized mobile devices.

UEM reaches mainstream adoption

Gartner defines UEM as being “a late-stage maturity market,” meaning “widespread adoption has already occurred,” said Cipolla. 

IDC data indicates that around two-thirds of US businesses have now deployed a UEM tool. That doesn’t mean most organizations will use a single UEM platform, however. 

Among those that have deployed UEM, around 70% have two or more  management products in place, said Hochmuth.   For example, an organization might have one tool to manage certain Windows devices, another for both mobile and macOS devices, and then a legacy PC management tool still in use for another set of Windows devices. “The norm is more the mixed type of organizations that have different tools and multiple UEMs,” said Hochmuth, though the trend in recent years has been towards consolidation of these tools.

What’s on the horizon for UEM? AI and autonomous endpoint management 

An ongoing trend related to UEM is the rise of digital employee experience (DEX) software. DEX tools can provide IT with data and insights into how employees interact with devices and applications, with the ability to measure usage and highlight performance problems. “That’s a growth area that all the UEM vendors are pushing into,” said Hochmuth.

Also coming to UEM tools: the integration of artificial intelligence (AI). “This space in particular, is incredibly ripe for help from an AI product,” said Hochmuth. 

AI could help manage a longtime challenge for endpoint management — scale. That’s because the wide range of devices, vulnerabilities, and configurations that have to be managed.

“The pure amount of data given off by thousands of devices running different operating systems, it’s super chaotic,” said Hochmuth. “That’s a perfect use case for an AI tool that could sift through data, help you find information you need, or even more importantly, automate a lot of the manual patching, updating, configuration – the reactionary type things that people in IT ops do. Anticipating when someone might need a fix before something breaks: AI could really help with that.”

Gartner’s Cipolla points to the emergence of autonomous endpoint management (AEM), a term that describes the combination of UEM and DEX, with additional automation and AI-assistance capabilities. “The idea is to take the human out of the middle doing the research and the leg work, and put them in control of the automation,” said Cipolla.

Several UEM vendors have already begun to incorporate AEM-like functionality into their software, said Cipolla. But it’s still early for the technology, meaning it will likely be at least a couple of years before AEM tools become more fully developed and more widely used by organizations. “It’s not a product yet, it’s a future idea, it’s a concept. As the vendors work on their ideas, it becomes a market,” he said. 

What is UEM? Unified endpoint management explained

Unified endpoint management (UEM) describes a set of technologies used to secure and manage a wide range of employee devices and operating systems — all from a single console.

Seen as the next generation of mobility software, UEM tools incorporate several existing enterprise mobility management (EMM) technologies — including mobile device management (MDM) and mobile application management (MAM) — with some of the tools used to secure desktop PCs and laptops.

“UEM in theory ties this all together and gives you that proverbial one pane of glass, so you can see the state of all of your endpoints,” said Phil Hochmuth, program vice president at IDC. “It gives you visibility into what people are doing with corporate data, corporate apps, on any conceivable type of device.”

The ability to manage various device types in one place is increasingly important as businesses face a growing cybersecurity threat, said Tom Cipolla, senior director analyst at Gartner. “We need to patch faster; everybody acknowledges that,” he said. “UEM gives people a consolidated view into their environment and a consolidated patching and configuration management approach.”

The evolution of mobile management – MDM, MAM, and more

At its core, UEM consists of several device management technologies that emerged to help businesses control employee mobile devices. The first iteration of such tools was MDM, which arrived about a decade ago.

Introduced in response to the initial wave of smartphones used in the workplace, MDM was designed to help IT centrally provision, configure, and manage mobile devices that had access to corporate systems and data. Common MDM features included security configuration and policy enforcement, data encryption, remote device wipe and lock, and location tracking.

However, as employee bring-your-own-device (BYOD) schemes became more prevalent in the office — driven first by the iPhone’s popularity, later by the growth of Android — vendors began to offer more targeted management of apps and data. MAM capabilities delivered more granular controls, focusing on software rather than the device itself; features include app wrapping and containerization, and the ability to block copy/paste or restrict which apps can open certain files.

MAM features were soon packaged with MDM and other tools, such as mobile identity management and mobile information management, and sold as comprehensive enterprise mobility management (EMM) product suites. Those suites led to the next stage in the evolution of device management: UEM.

What is UEM?

UEM merges the various facets of EMM suites with functionality typically found in client management tools (CMT) used to manage desktop PCs and laptops on a corporate network. One example is Microsoft’s Intune, which combined its MDM/MAM platform with Configuration Manager (formerly System Center Configuration Manager) in 2019.

UEM platforms tend to have comprehensive operating system support, including mobile (Android, iOS) and desktop OSes (Windows 11, macOS, ChromeOS, and, in some cases, Linux). Some UEM products support more esoteric categories too, including IoT devices, AR/VR headsets, and smartwatches.

Unlike traditional CMT products, UEM tends to be available as a software-as-a-service, cloud-based tool, allowing management and updates of devices such as desktop PCs without connection to a corporate network. 

The emergence of UEM has been partly driven by the inclusion of API-based configuration and management protocols within Windows and macOS, enabling the same level of device management that was already possible with iOS and Android devices.

It speaks to a wider development, too, of the convergence of mobile and traditional computing devices, with high-end tablets often on par with laptops in terms of processing power. “You have a real blurring of the lines between what is mobile computing and what is traditional endpoint computing,” said Hochmuth.

Why invest in UEM tools?

All of these devices — mobile, desktop, Windows, Mac, in the office and remote — require a unified approach to end user device management, an approach that can provide a variety of benefits, say analysts.

Among these is the opportunity for simplified and centralized management. In short, it’s more efficient for one team to provision and manage all devices from a single tool, rather than have separate support teams and tools that were traditionally divided between mobile and Windows or macOS computers. 

“If you have a separate software product or management platform for four different operating systems, that can be cumbersome and expensive,” said IDC’s Hochmuth. “Converging down to one or two is a goal for a lot of organizations.”

UEM products can reduce manual work for IT, with the ability to create a single policy — such as requiring device encryption — that can be deployed to many devices and operating systems. The same goes for patching.  

By ensuring consistent policies across apps, devices and data, UEM tools can reduce risk, with less complexity and fewer opportunities to misconfigure policies. 

There are cost benefits in replacing separate PC and mobile management applications too. “Getting rid of one software platform and all the licensing associated with that is a cost saving. That’s not the primary driver, but it’s definitely a reason to explore UEM,” said Hochmuth. 

The UEM vendor market

The global market for unified endpoint management software is forecast to grow from $5.9 billion in 2023 to $8.9 billion in 2028, according to IDC data. The rate of yearly growth is set to slow, however, from around 16% to 6% during this period. 

There are a variety of vendors, from big-name firms to smaller, more targeted companies. Microsoft (Intune) and VMware/Broadcom (Workspace One) are often considered the UEM market leaders with the broadest offerings and largest market share by revenue. BlackBerry UEM, Citrix Meraki Systems Manager, IBM MaaS360, ManageEngine, Cisco, and Ivanti UEM are also popular products.

“All these companies have roles or verticals or use cases that they address specifically,” said Hochmuth. For instance, BlackBerry is often viewed as strong in regulated markets, such as finance or healthcare, due its focus on encryption, while Microsoft has a more of a “horizontal” product with general business use cases.  

Among the vendors that have taken a more specialized approach is Jamf, which is focused purely on Apple devices running everything from macOS to tvOS, and SOTI, whose products are tailored to certain industries, such as warehouse workers with ruggedized mobile devices.

UEM reaches mainstream adoption

Gartner defines UEM as being “a late-stage maturity market,” meaning “widespread adoption has already occurred,” said Cipolla. 

IDC data indicates that around two-thirds of US businesses have now deployed a UEM tool. That doesn’t mean most organizations will use a single UEM platform, however. 

Among those that have deployed UEM, around 70% have two or more  management products in place, said Hochmuth.   For example, an organization might have one tool to manage certain Windows devices, another for both mobile and macOS devices, and then a legacy PC management tool still in use for another set of Windows devices. “The norm is more the mixed type of organizations that have different tools and multiple UEMs,” said Hochmuth, though the trend in recent years has been towards consolidation of these tools.

What’s on the horizon for UEM? AI and autonomous endpoint management 

An ongoing trend related to UEM is the rise of digital employee experience (DEX) software. DEX tools can provide IT with data and insights into how employees interact with devices and applications, with the ability to measure usage and highlight performance problems. “That’s a growth area that all the UEM vendors are pushing into,” said Hochmuth.

Also coming to UEM tools: the integration of artificial intelligence (AI). “This space in particular, is incredibly ripe for help from an AI product,” said Hochmuth. 

AI could help manage a longtime challenge for endpoint management — scale. That’s because the wide range of devices, vulnerabilities, and configurations that have to be managed.

“The pure amount of data given off by thousands of devices running different operating systems, it’s super chaotic,” said Hochmuth. “That’s a perfect use case for an AI tool that could sift through data, help you find information you need, or even more importantly, automate a lot of the manual patching, updating, configuration – the reactionary type things that people in IT ops do. Anticipating when someone might need a fix before something breaks: AI could really help with that.”

Gartner’s Cipolla points to the emergence of autonomous endpoint management (AEM), a term that describes the combination of UEM and DEX, with additional automation and AI-assistance capabilities. “The idea is to take the human out of the middle doing the research and the leg work, and put them in control of the automation,” said Cipolla.

Several UEM vendors have already begun to incorporate AEM-like functionality into their software, said Cipolla. But it’s still early for the technology, meaning it will likely be at least a couple of years before AEM tools become more fully developed and more widely used by organizations. “It’s not a product yet, it’s a future idea, it’s a concept. As the vendors work on their ideas, it becomes a market,” he said. 

Apple agrees to open up Apple Pay in Europe

As Apple faces continued waves of regulation, Apple Pay is about to open up in Europe, allowing rival payment services to gain access to the NFC chips inside iPhones to enable one-click payments.

The motivation behind forcing Apple to open up is to stimulate competition in the mobile payments space. It should enable rival services to offer mobile payments and settles a long-running dispute between Apple and the European Commission. 

What this means to Apple Pay

Under the arrangements, Apple will allow third-party wallet providers access to the NFC chip inside iOS devices without requiring them to use Apple Pay or Apple Wallet. It means rivals can now compete directly with the Apple service, and in theory means customers can choose a payment system they prefer. This relies on an extensive number of commitments, captured in a 36-page document published today.

What Europe says

“From now on, competitors will be able to effectively compete with Apple Pay for mobile payments with the iPhone in shops,” Margrethe Vestager, executive vice president in charge of competition policy, said in a statement. “So, consumers will have a wider range of safe and innovative mobile wallets to choose from.”

EC authorities have put some steel around the agreements. They will by law remain in force for 10 years and apply throughout the EEA. “Their implementation will be monitored by a monitoring trustee appointed by Apple who will report to the Commission for the same time period,” the European Commission said.

In the event Apple fails to keep its commitments, it faces a fine of up to 10% of its total annual turnover without having to find an infringement of EU antitrust rules, or a “periodic penalty” payment of 5% per day of its daily turnover for every day of non-compliance.

How will it work?

A look at the 36-page agreement suggests how the new system will work. First, developers of payment systems will need to obtain entitlements to access a series of APIs Apple will make available to support rival payment systems, but only those operating in the European Economic Area. 

The company will also work to support evolving standards; developers will be subject to developer fees, but no fees related to the use of the NFC system. That sounds like Apple will not receive a cut of payments made.

For consumers, it will be possible to choose a preferred payment system (including Apple Pay) with a new section in Settings. The iPhone will also maintain a register of installed payment apps that want NFC access, and you’ll be able to select which one to use, rather like rifling through payment cards in your real wallet.

You’ll also be able to use Apple Pay on Apple Watch and choose another system for your phone.

What about disputes?

If a developer/payment provider thinks they aren’t getting fair treatment from Apple, they will be able to submit a written complaint to the monitoring trustee. Appointed and reimbursed by Apple and approved by the European Commission, the trustee will be an independent party who monitors the company’s compliance to the agreement.

The trustee may recruit a support team of up to three advisors, and there are strict controls in place to prevent trustees running off to work for Apple or its competitors within a certain time frame. There will also be an Appeal Board to adjudicate in the event a dispute requires independent oversight. 

What about the DMA?

Apple’s decision to reach a constructive settlement concerning Apple Pay in Europe could yet turn out to be a harbinger of similar future détente regarding Europe’s Digital Markets Act. While recent statements from Vestager suggest she has little empathy for Apple’s arguments, the company has already revised some of the arrangements it proposed to bring its business practises into line with the DMA or similar rules looming in other nations.

There’s no reason to think it won’t continue to reach a constructive, if unenthusiastic, dialogue. It does remain open to question whether the agreements will go far enough for Europe or for some of the company’s loudest critics. 

But for the next decade, at least, you’ll be able to use whatever payment system you like across the European bloc as easily as you may already use Apple Pay.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.

With the arrival of AI, Slack adds a new chapter to its story

It’s been 10 years since Slack launched its popular chat application and ushered in an era of fast-paced and more casual business communications. While the email inbox hasn’t yet been consigned to the past, the effect Slack has had on office work is clear, making it easier (at times, too easy) to share information and interact with colleagues, regardless of where they are. 

For the company’s new CEO, Denise Dresser, the introduction of AI-based tools is an opportunity for the company to continue to shape the way work gets done. “I could not be more optimistic about what the future of AI is going to bring to the future of how we all work,” Dresser said. “We celebrated our 10th anniversary in February and I feel like Slack was made for this moment of generative AI…, for Slack to again lead the next decade of this AI-powered future of work.”

The launch of Slack AI earlier this year is one of bigger changes to Slack’s application in recent years. A revamped user interface rolled out in 2023 sought to retain ease of use even as new functions were added. The changes ranged from canvas documents to lightweight video and voice calls and a task management tool, with automation continuing as a major focus via Workflow Builder

There have been some major changes in personnel, too. Co-founder and Stewart Butterfield announced his departure in 2022, a year after Slack’s $27.7 billion acquisition by Salesforce, and other senior leaders have since moved on. Butterfield’s successor, Lidiane Jones, was CEO for just a year before taking over at dating app company Bumble. That makes Dresser, who joined in November 2023, the third boss in a little over a year. 

Among her priorities are plans to bring Slack’s new native capabilities — such as the recently launched lists tool — to customers in a “broader way,” while continuing to build AI into the platform after the general availability launch of Slack AI in February

Another focus has been to more deeply integrate Slack into the Salesforce ecosystem in terms of both product and customer sales strategy. Dresser’s background at Salesforce — where she has held several senior executive roles since 2011 — should help align the two businesses, said Will McKeon-White, senior analyst at Forrester. Her appointment will help in “creating better joint go-to-market motions, in all the rationalization and operationalization that needs to happen with any of these motions — I’m quite a fan of that,” he said. 

Slack’s headwinds

Dresser takes over at a time of slowing growth for the business. Quarterly revenue growth during FY2024 and into FY2025 has reached between 16% and 20% year over year, roughly half as high as quarterly growth shown in Slack and Salesforce earnings reports between 2020 and 2023. 

“Slack has been facing more headwinds recently,” said McKeon-White, pointing to internal challenges such as integration efforts after the Salesforce acquisition, a fast-changing competitive environment (with a wider range of rivals such as Zoom competing more directly), and a shift in customer purchasing post-pandemic.

After businesses scrambled to roll out communication software during the COVID-19 outbreak to facilitate remote work at scale, many later sought to reduce the number of applications they use. The global market for collaboration software continued to see double digit growth, according to IDC data for 2022, when the market was valued at $33.9 billion, though the rate of increase slowed as the pandemic eased. 

Slack appears to have felt the change more acutely, said McKeon-White, due to a formidable competitor: Microsoft’s Teams, which launched in 2016 as a response to Slack’s runaway workplace success. 

For customers invested in the Microsoft 365 suite, it made sense to use what they were already paying for. “Our research shows — and I think the market shows — that a fair amount of companies have gone in that direction and said Teams is ‘good enough,’” said Irwin Lazar, president and principal analyst at Metrigy.

Microsoft has now unbundled Teams from M365 for new subscribers (following an antitrust battle with European regulators), but that’s unlikely to benefit Slack in a significant way, analysts have said.  And yet, many organizations support both apps, said McKeon-White, as businesses seek to deploy multiple communication tools to meet employee needs. 

“So, while there has been that gradual attrition and centralization, there’s now an emerging counter movement to that,” he said.

“There is competition between Slack and Teams, but when they’re used together, when they’re integrated, there’s also a synergy,” said Wayne Kurtzman, IDC’s vice president of social, community and collaboration. “So additional growth may actually come from the synergy of having both in the enterprise.”

In a crowded field, still room to grow

Despite the challenges, Slack remains in a strong position to grow, say analysts. Efforts to add functionality to the platform have paid off, making the application even more useful to customers. “The enhancements to the platform are leaning into their strengths, which is as a center of collaboration and automation in an organization…,” said McKeon-White.  

Dresser argued that the value of Slack is clear and cited the company’s own customer survey data; it indicatea a 47% productivity increase, a 36% increase in win rate for sales users, 32% faster case resolution time in customer service, and a 37% acceleration for decision making in marketing.

Said Dresser: “I find it’s not hard to make the case [to customers]; it’s focusing on the business outcome of the platform itself. Slack is where work gets done and our results and outcomes really speak to that.”

The clearest opportunity for growth lies in selling Slack to Salesforce customer organizations, said McKeon-White, though this remains a work in progress. “That is a ready-made pipeline for them, effectively, but will require some joint go-to-market efforts and additional contract value…. That might be something like platform discounts and other similar motions,” he said.

Slack hasn’t moved as aggressively to integrate with Salesforce as it might have, though the launch last year of Sales Elevate, which makes Salesforce data more easily accessible in the collaboration app, is a sign of an improvement. “I think that’s where there’s a huge opportunity to make Slack the front-end of Salesforce,” said Lazar.  If I’m a salesperson or sales manager, or if I’m using Salesforce marketing campaigns, then I can manage all the different Salesforce features within Slack, and I have the ability to collaborate,” he said.

McKeon-White also sees potential for Slack to further tailor its app to specific job roles and industries. Features like lists and Workflow Builder enable Slack to be tailored to internal use cases, such as procurement, for example, or IT, and there are  opportunities to cater to specific verticals such as a healthcare or retail organization more intently.

Slack can also increase revenues from existing customers, said Lazar, as it continues to evolve. “Most of their growth is going to happen within their existing customer base by adding new feature functionality and adding higher-level licenses, or converting people over to the Enterprise Grid product,” he said. 

Slack’s AI future

A major focus for the company, as with all vendors in the collaboration and productivity software space, is the addition of generative AI (genAI) tools. 

Slack AI launched earlier this year, with three features:

  • AI powered search. This provides personalized answers to questions based on an organization’s knowledge base. Slack AI helps users locate subject matter experts, or find information on anything from work projects to understanding unfamiliar acronyms.
  • Channel recaps. This highlights key discussion points for a Slack user after a period away from the app, or for those who have recently joined a channel.
  • Thread summaries. This feature recaps faster-moving discussions, provides thread summaries, and offers an overview of long conversations, with links to sources in each summary that enable users to check information where necessary.

Slack AI’s advantage lies in its ease of use, with little or no training required, Dresser said.

Slack AI search

Slack AI search allows users to more quickly find information that could be buried in channels and chats.

Slack

“One of our product principles is ‘don’t make me think’ and that’s a key part of how we’re thinking about AI,” she said. That means ensuring Slack is embedded in “the most logical places that drive immediate productivity, and maybe a little bit of joy and delight in the process.” She points to the AI recap feature. “I love starting my day out with ‘recap,’ so that when there are channels that I don’t necessarily read all day long, I get a quick recap of what happened and I’m on with my day.”

Slack, like all tech companies, is still working to overcome some of genAI’s limitations. Hallucinations are an inherent problem for large language models(LLMs), particularly in a workplace context where accuracy is vital. Dresser said Slack attempts to mitigate the impact of hallucinations with citations that link back to the original source of information. “It allows people to feel that it is less of a black box,” she said. “They can actually see the specific conversation that led to the summarization of that result. It’s little things like that that provide the transparency that helps you build trust.”

Slack CEO: Trust matters

Trust around the use of customer data is a hot topic, too. Slack users recently vented frustrations at terms of service that some interpreted as the company seeking to use customer data to train its AI models. While Slack explained that the terms related to the use of “traditional” machine learning algorithms for relatively benign purposes (channel and emoji recommendations, for instance) rather than using messages to train LLMs as some had feared, the situation underlined the tensions around access to customer data. 

“We did hear from customers that we needed to be more clear, so we immediately updated our language on the website, so customers know exactly where we stand,” Dresser said. “Trust is our top priority. When we built generative AI natively into Slack, it was a huge area of our focus. 

“We do not develop LLMs or other generative AI models using customer data, full stop.”

Slack is not alone in tackling genAI’s various difficulties. “This is like the pre-game show for AI,” said Kurtzman. “It is the very beginning. Things are not where we imagine they should be. Slack is doing well with AI that’s tuned to identify content within a conversation and identify value within the conversation. But everyone’s AI is continually improving.”

Despite widespread interest in the technology, there’s still a long way to go in terms of broad adoption. A recent Slack survey showed that only 32% of respondents have accessed AI in their jobs, with half doing so on a weekly basis. 

Part of that is because of cost, part of it is uncertainty about whether generative AI can deliver value, given the additional cost to users. Slack AI costs an additional $10 per user each month — that’s less expensive than others, but still a significant outlay as AI assistants become widely available.

“For organizations who have used it [Slack AI], they seem to be very happy with it,” said McKeon-White. “But getting the budget together in order to justify another internal AI experiment is fairly difficult today: It turns out AI is expensive, especially if you try to do it for all of your organization.” 

“On the whole, we believe that pricing will eventually be baked into everything as AI becomes ubiquitous,” said Kurtzman. “But for today, the [additional] pricing generally returns value fairly quickly.” 

The initial Slack AI feature such as conversation summarization are useful, but can make it hard to justify the cost. “I think initially it’s a tough sell,” said Lazar, at least until Slack AI can integrate a wider range of data sources from third-party apps, which could significantly increase its capabilities.

Still, early Slack AI customers have already noted its utility, said Dresser; an internal analysis of pilot customers indicated it saves users an average 97 minutes a week, for instance. “We’re still in the very early days…, but the results are really positive. Starting in the right places, in a trusted manner, right in the flow of work, will be the way that I think the world begins to adopt…AI,” she said.

SAI Group buys Get Well; aims to use AI for better patient engagement

Investment firm SAI Group this week announced it has acquired Get Well, a 24-year-old company that provides digital patient engagement technology to 1,000 healthcare organizations.

The financial terms of the deal were not disclosed.

SAI said the purchase of Get Well adds to its portfolio of AI healthcare companies. SAI plans to integrate its own generative AI (genAI) platform – GPT 4.0-powered RhythmX AI — “into the patient experience inside and outside the hospital.”

(RhythmX is also the name of SAIGroup’s subsidiary company.)

GetWell’s own digital patient engagement platform — Get Well 360 — already interacts with more than 10 million patients annually, offering them online point-of-care engagement and “guided care,” among other modules. The RhythmX platform offers patients prescriptive actions and recommendations doctors can drill into using a generative AI-enabled natural language interface and AI-native copilots.

“As part of SAIGroup, Get Well’s mission to enable the best patient experience will undergo a rapid transformation with AI to a full precision care platform for hospitals and ambulatory centers,” SAIGroup CEO Romesh Wadhwani said in a statement. “This strategic investment underscores SAIGroup’s commitment to innovative AI-driven solutions in healthcare and highlights our confidence in Get Well as a leader in the digital patient engagement space.” 

GetWell’s competitors in the Healthcare Management System arena include EPIC, Cerner, and eClinicalWorks.

Through mergers and acquisitions, SAIGroup has grown into a company with a massive trove of healthcare data from 300 million patients, 4.4 billion annual claims, and information on more than 1.8 million healthcare professionals, according to its own reports.

“Experience, which is often where engagement falls, continues to be the top outcome sought from digital investments,” but many organizations are still falling short of goals set by their executive leadership, according to Faith Adams, a Gartner senior director analyst.

As in most other industries, healthcare providers face a massive shortage of AI-skilled employees and IT pros needed to integrate new automation tools. Healthcare also faces a shortage of clinicians, which automated patient interactions could help address, according to Adams.

A 2024 survey by online education company Pluralsight showed more than 80% of IT pros think they can use AI, but just 12% have the skills and expertise to do so. That same survey showed 97% of firms that have deployed AI have benefited from it, citing increased productivity and efficiency, improved customer service, and reduced human error.

““The biggest part of the story is the shortage of AI tech experience, and patient engagement experience,” Adams said. “One of the bigger opportunities we see here is bringing together SAI’s AI expertise with GetWell’s patient engagement expertise.”

AI platforms can serve as digital tools to bolster patient access to personalized medicine and health literacy — the ability to obtain, read, understand, and use healthcare information to make appropriate health decisions and follow treatment instructions. AI tech can also help patients with their “digital literacy,” allowing them to better find, evaluate, and communicate information through digital media platforms.

In other words, instead of struggling to contact clinicians, online query and answer engines powered by AI can give patients answers based on their own health record information and clinical recommendations.

Gartner coined the phrase “Intelligent Health” last year to describe what it sees at the future of digital transformation in healthcare and the life science industries. Intelligent Health refers to the harnessing of the ever-growing volume and variety of patient and clinical data to offer providers and patients a better and more precise healthcare experience.

Intelligent Healthcare

Gartner Inc.

“Given the complexity of healthcare patient journeys, there is really no one-size-fits-all, and this is where technology can help better support personalization [and] precision using data and insights,” Adams said. “Intelligent health is interoperable by default, relying on continuous data to deliver experience through the unification of digital and in-person care delivery that is precise, equitable and ethical.”

Every patient needs to be approached differently to drive behavioral changes, according to Adams. For example, if a patient needs to lose weight or eat healthier to lower their cholesterol and/or blood pressure levels, AI-based technology can assess their history and make recommendations.

“Patients continue to demand more from their experiences, and they have more choice now than ever. Each patient type needs to be approached differently to drive behavioral change.  This [AI tool] simplifies it,” Adams said.

“There are other factors that can influence it, too, but this is always a good starting point to show the no-one-size-fits-all approach will drive behavior change and engagement.”

SAI Group buys Get Well; aims to use AI for better patient engagement

Investment firm SAI Group this week announced it has acquired Get Well, a 24-year-old company that provides digital patient engagement technology to 1,000 healthcare organizations.

The financial terms of the deal were not disclosed.

SAI said the purchase of Get Well adds to its portfolio of AI healthcare companies. SAI plans to integrate its own generative AI (genAI) platform – GPT 4.0-powered RhythmX AI — “into the patient experience inside and outside the hospital.”

(RhythmX is also the name of SAIGroup’s subsidiary company.)

GetWell’s own digital patient engagement platform — Get Well 360 — already interacts with more than 10 million patients annually, offering them online point-of-care engagement and “guided care,” among other modules. The RhythmX platform offers patients prescriptive actions and recommendations doctors can drill into using a generative AI-enabled natural language interface and AI-native copilots.

“As part of SAIGroup, Get Well’s mission to enable the best patient experience will undergo a rapid transformation with AI to a full precision care platform for hospitals and ambulatory centers,” SAIGroup CEO Romesh Wadhwani said in a statement. “This strategic investment underscores SAIGroup’s commitment to innovative AI-driven solutions in healthcare and highlights our confidence in Get Well as a leader in the digital patient engagement space.” 

GetWell’s competitors in the Healthcare Management System arena include EPIC, Cerner, and eClinicalWorks.

Through mergers and acquisitions, SAIGroup has grown into a company with a massive trove of healthcare data from 300 million patients, 4.4 billion annual claims, and information on more than 1.8 million healthcare professionals, according to its own reports.

“Experience, which is often where engagement falls, continues to be the top outcome sought from digital investments,” but many organizations are still falling short of goals set by their executive leadership, according to Faith Adams, a Gartner senior director analyst.

As in most other industries, healthcare providers face a massive shortage of AI-skilled employees and IT pros needed to integrate new automation tools. Healthcare also faces a shortage of clinicians, which automated patient interactions could help address, according to Adams.

A 2024 survey by online education company Pluralsight showed more than 80% of IT pros think they can use AI, but just 12% have the skills and expertise to do so. That same survey showed 97% of firms that have deployed AI have benefited from it, citing increased productivity and efficiency, improved customer service, and reduced human error.

““The biggest part of the story is the shortage of AI tech experience, and patient engagement experience,” Adams said. “One of the bigger opportunities we see here is bringing together SAI’s AI expertise with GetWell’s patient engagement expertise.”

AI platforms can serve as digital tools to bolster patient access to personalized medicine and health literacy — the ability to obtain, read, understand, and use healthcare information to make appropriate health decisions and follow treatment instructions. AI tech can also help patients with their “digital literacy,” allowing them to better find, evaluate, and communicate information through digital media platforms.

In other words, instead of struggling to contact clinicians, online query and answer engines powered by AI can give patients answers based on their own health record information and clinical recommendations.

Gartner coined the phrase “Intelligent Health” last year to describe what it sees at the future of digital transformation in healthcare and the life science industries. Intelligent Health refers to the harnessing of the ever-growing volume and variety of patient and clinical data to offer providers and patients a better and more precise healthcare experience.

Intelligent Healthcare

Gartner Inc.

“Given the complexity of healthcare patient journeys, there is really no one-size-fits-all, and this is where technology can help better support personalization [and] precision using data and insights,” Adams said. “Intelligent health is interoperable by default, relying on continuous data to deliver experience through the unification of digital and in-person care delivery that is precise, equitable and ethical.”

Every patient needs to be approached differently to drive behavioral changes, according to Adams. For example, if a patient needs to lose weight or eat healthier to lower their cholesterol and/or blood pressure levels, AI-based technology can assess their history and make recommendations.

“Patients continue to demand more from their experiences, and they have more choice now than ever. Each patient type needs to be approached differently to drive behavioral change.  This [AI tool] simplifies it,” Adams said.

“There are other factors that can influence it, too, but this is always a good starting point to show the no-one-size-fits-all approach will drive behavior change and engagement.”

Microsoft 365 Copilot explained: genAI meets Office

Initially called Microsoft 365 Copilot when it launched in November 2023, the renamed “Copilot for Microsoft 365” brings a range of generative AI (genAI) features to office productivity apps such as Word, Outlook, Teams and Excel. 

In a blog post announcing the tool, Microsoft CEO Satya Nadella described it as “the next major step in the evolution of how we interact with computing…. With our new copilot for work, we’re giving people more agency and making technology more accessible through the most universal interface — natural language.”

At launch, Microsoft explained that the Copilot “system” consists of three elements: Microsoft 365 apps such as Word, Excel, and Teams, where users interact with the AI assistant; Microsoft Graph, which includes files, documents, and data across the Microsoft 365 environment; and the OpenAI models that process user prompts, such as the ChatGPT-4 large language model and DALL-E 3 model for image generation.

With the tool, Microsoft aims to create a “more usable, functional assistant” for work, J.P. Gownder, vice president and principal analyst at Forrester’s Future of Work team, told Computerworld in fall 2023. “The concept is that you’re the ‘pilot,’ but the Copilot is there to take on tasks that can make life a lot easier.” 

The Copilot for M365 is “part of a larger movement of generative AI that will clearly change the way that we do computing,” he said, noting how the technology has already been applied to a variety of job functions — from writing content to creating code — since ChatGPT-3.5 launched in late 2022.

Forrester report last year predicted that 6.9 million US knowledge workers — around 8% of the total — would be using Copilot for M365 by the end of 2024.

Nadella talked up the effectiveness of the M365 Copilot during a 2023 earnings call, claiming customers had seen productivity gains in line with that of the GitHub Copilot, the AI assistant aimed at developers that launched two years ago. (For reference, GitHub has previously claimed developers were able to complete a single task 55% quicker thanks to the GitHub Copilot, while acknowledging the challenges in measuring productivity.)

Even priced at $30 per user per month, there’s potential to deliver considerable value to businesses, assuming the Copilot delivers on its promise over time. Said Gownder: “The key issue is, ‘Does it actually save that time?’ because it’s hard to measure and we don’t really know for sure. But even conservative time savings estimates are pretty generous.”

The Copilot for M365 is billed as providing employees with access to genAI without the security concerns of consumer genAI tools; Microsoft says its models aren’t trained on customer data, for instance. But deploying the tool represents significant challenges, said Avivah Litan, distinguished vice president analyst at Gartner.

There are two primary business risks, she said: the potential for the Copilot to ‘hallucinate’ and provide inaccurate information to users, and the ability for the Copilot’s language models to access huge swathes of corporate data that’s not locked down properly.

“Information oversharing is one of the biggest issues people are going to face in the next few months, or six months to a year,” said Litan. “That’s where the rubber is going to hit the road on the risk — it’s not so much giving the data to Microsoft or OpenAI or Google, it’s all the exposure internally.”

Copilot for Microsoft 365 features: How do you use it?


Copilot interactions within apps can take a variety of forms, depending on the application. In many cases, users will interact with it via the chat interface available in a sidebar; Copilot functionality is also built more directly into some apps, such as a pop-up in a Word document or Outlook email, for instance. 

Here’s how the Copilot works in some M365 apps.

In a Word doc, it can suggest improvements to existing text or let users create a first draft from scratch. To generate a draft, a user can ask Copilot in natural language to create text based on a prompt, and can upload additional files and sources of information to guide the AI assistant. Once created, the user can edit th document, adjust the style, or ask the Copilot to redo the whole thing. A Copilot sidebar provides space for more interactions with the bot, which also suggests prompts to improve the draft, such as adding images or an FAQ section, or summarize the text. 

During a Teams video call, the Copilot provides a recap of what’s been discussed so far, with a brief overview of conversation points in real time. It’s also possible to ask the AI assistant for feedback on people’s views during a call, or what questions remain unresolved. Those unable to attend a particular meeting can send the AI assistant in their place to provide a summary of what they missed and action items they need to follow up on. 

Copilot Word draft

Copilot can help a Word user draft a proposal from meeting notes. 

In PowerPoint, Copilot can automatically turn a Word document into draft slides that can then be adapted via natural language in the Copilot sidebar. It can also generate suggested speaker notes to go with the slides and add more images. 

These are just some examples. Other apps that feature Copilot integration include Excel, Outlook, OneNote, Loop, and Whiteboard.

The other way to interact with Copilot is via a separate chat interface that’s accessible via Teams. Here, the Copilot works as a search tool that surfaces information from a range of sources, including documents, calendars, emails, and chats. For instance, an employee could ask for an update on a project, and get a summary of relevant team communications and documents already created, with links to sources.

Microsoft will extend Copilot’s reach into other apps workers use via “plugins” — essentially third-party app integrations. These will allow the assistant to tap into data held in apps from other software vendors including Atlassian, ServiceNow, and Mural. Fifty such plugins are available, with “thousands” more expected eventually, Microsoft said. 

How much does Copilot cost — is it worth $30 per user, per month?

The main Microsoft 365 Copilot is available for enterprise customers on E3, E5, F1 and F3 plans, as well as Office E1, E3, E5, and Apps for Enterprise. It’s also available for smaller business customers on the following plans: Businesses Basic, Business Standard, Business Premium, and Apps for Business.

In each case, the Copilot for Microsoft 365 costs an additional $30 per user each month.

It’s a significant extra expense given that M365 subscriptions start at $6 per user each month for Busines Basic and go up to $55 per user each month for E5. Part of this due to the cost of the high computing costs of the Copilot incurred by Microsoft, said Raúl Castañón, senior research analyst at 451 Research, a part of S&P Global Market Intelligence.

“Microsoft is likely looking to avoid the challenges faced with GitHub Copilot, which was made generally available in mid-2022 for $10/month and, despite surpassing more than 1.5 million users, reportedly remains unprofitable,” said Castañón.

In addition to the core Copilot for M365, job role-specific Copilots are available as paid add-ons. Sales and service Copilots each cost an additional $20 per user each month, while a finance Copilot is currently in preview.

The pricing strategy reflects Microsoft’s confidence in the impact that genAI will have on workforce productivity.

Per Forrester’s calculations in the “Build Your Business Case For Microsoft 365 Copilot” report, an employee earning $120,000 annually — roughly $57 per hour — might save four hours a month on various productivity tasks; those four hours would be worth around $230 a month. In that scenario, it would make sense to invest in Copilot for an employee earning even half that amount, and that’s leaving aside less tangible benefits around employee experience when automating mundane tasks.

There are, as the Forrester points out, other costs to consider beyond licensing — employee training, for instance, as employees learn the new technology. Gartner also predicts that enterprise security spending will increase in the region of 10% to 15% in the next couple of years as a result of efforts to secure genAI tools (not just M365 Copilot).

Businesses are likely to take a cautious approach to deploying the Microsoft tool, at least at first. Microsoft expects revenue related to M365 Copilot to “grow gradually over time,” Microsoft CFO Amy Hood said during the company’s Q1 2024 earnings call. On the same call, Nadella noted that Copilot will be subject to the usual “enterprise cycle times in terms of adoption and ramp.”

Even if the pace of adoption is gradual, there appears to be plenty of interest in deploying it. Forrester expects around a third of M365 customers in the US to invest in Copilot in the first year. Companies that do so will provide licenses to around 40% of employees during this period, the firm estimated.

(Note: while not actually branded as Copilot, Microsoft also makes some genAI features available in Teams Premium. This includes AI-generated notes, AI-generated tasks and live translations in video calls, all of which are powered by ChatGPT AI models. For businesses that are mostly interested in AI assistant features for meetings, this offers a cheaper option than a full Copilot for M365 subscription.) 

What are Microsoft’s other Copilots?

Microsoft’s Copilot is embedded in a wide array of products. Beyond the M365 suite, there are Copilots for Dynamics, Power BI, GitHub, and Microsoft’s security suite.

And then there are Copilots aimed primarily at consumer, rather than business, users. 

Microsoft launched Copilot Pro in January 2024, a $20 a month subscription that provides individuals with similar functionality to the Copilot for M365. Copilot Pro customers gain access to Copilot chatbot and genAI image creation, as well as AI assistant features in free web versions of apps such as Word, Excel, PowerPoint, and Outlook (though not Teams). Those with Microsoft 365 Personal and family subscriptions can also access the Copilot in desktop apps. 

There’s also a free version of the Copilot with access to chatbot functionality only.

The Copilot chat interface is accessible in several ways by both paid and free users. There’s a dedicated web page, a mobile app, and a chatbot built into the Windows operating system, Edge browser, and Bing search engine.

How are early customers using Copilot?

There are two basic ways users will interact with Copilot. It can be accessed directly within a particular app — to create PowerPoint slides, for example, or an email draft — or via a natural language chatbot accessible in Teams, known as Microsoft 365 Chat. 

Interactions within apps can take a variety of forms, depending on the application. When Copilot is invoked in a Word document, for example, it can suggest improvements to existing text, or even create a first draft.

To generate a draft, a user can ask Copilot in natural language to create text based on a particular source of information or from a combination of sources. One example: creating a draft proposal based on meeting notes from OneNote and a product road map from another Word doc. Once a draft is created, the user can edit it, adjust the style, or ask the AI tool to redo the whole document. A Copilot sidebar provides space for more interactions with the bot, which also suggests prompts to improve the draft, such as adding images or an FAQ section. 

During a Teams video call, a participant can request a recap of what’s been discussed so far, with Copilot providing a brief overview of conversation points in real time via the Copilot sidebar. It’s also possible to ask the AI assistant for feedback on people’s views during the call, or what questions remain unresolved. Those unable to attend a particular meeting can send the AI assistant in their place to provide a summary of what they missed and action items they need to follow up on. 

In PowerPoint, Copilot can automatically turn a Word document into draft slides that can then be adapted via natural language in the Copilot sidebar. Copilot can also generate suggested speaker notes to go with the slides and add more images. 

The other way to interact with Copilot is via Microsoft 365 Chat, which is accessible as a chatbot with Teams. Here, Microsoft 365 Chat works as a search tool that surfaces information from a range of sources, including documents, calendars, emails, and chats. For instance, an employee could ask for an update on a project, and get a summary of relevant team communications and documents already created, with links to sources.

Microsoft will extend Copilot’s reach into other apps workers use via “plugins” — essentially third-party app integrations. These will allow the assistant to tap into data held in apps from other software vendors including Atlassian, ServiceNow, and Mural. Fifty such plugins are available, with “thousands” more expected eventually, Microsoft said. 

copilot business chat

Copilot can synthesize information about a project from different sources.

How are early customers using Copilot?

Prior to launch, many businesses accessed the Copilot for M365 as part of a paid early access program (EAP); it began with a small number of participants before growing to several hundred customers, including Chevron, Goodyear, and General Motors. 

One of those involved in the EAP was marketing firm Dentsu, which began deploying 300 licenses to tech staff and then employees across its business lines globally. The most popular use case so far is summarization of information generated in M365 apps — a Teams call being one example.

“Summarization is definitely the most common use case we see right out of the box, because it’s an easy prompt: you don’t really have to do any prompt engineering…, it’s suggested by Copilot,” Kate Slade, director of emerging technology enablement at Dentsu, said.

Staffers would also access M365 Chat functions to prepare for meetings, for instance, with the ability to quickly pull information from different sources. This could mean finding information from a project several years ago “without having to hunt through a folder maze,” said Slade.

The feedback from workers at Dentsu has been overwhelmingly positive, said Slade, with a waiting list now in place for those who want to use the AI tool.

“It’s reducing the time that they spend on [tasks] and giving them back time to be more creative, more strategic, or just be a human and connect peer to peer in Teams meetings,” she said. “That’s been one of the biggest impacts that we’ve seen…, just helping make time for the higher-level cognitive tasks that people have to do.”

Use cases have varied between different roles. Denstu’s graphic designers would get less value from using Copilot in PowerPoint, for example: “They’re going to create really visually stunning pieces themselves and not really be satisfied with that out-of-the-box capability,” said Slade. “But those same creatives might get a lot of benefits from Copilot in Excel and being able to use natural language to say, ‘Hey, I need to do some analysis on this table,’ or ‘What are key trends from this data?’ or ‘I want to add a column that does this or that.’”

How does Copilot compare with other productivity and collaboration genAI tools?

Most vendors in the productivity and collaboration software market have added genAI to their offerings at this point.

Google, Microsoft’s main competitor in the productivity software arena, launched DuetAI for Workspace in 2023, and rebranded to Gemini Enterprise ($30 per user each month) and   Gemini Business ($20 user each month). Google’s AI assistant can summarize Gmail conversations, draft texts, and generate images in Workspace apps such as Docs, Sheets,and Slides. 

Slack, the collaboration software firm owned by Salesforce and a rival to Microsoft Teams, launched its Slack AI feature in February. Other firms that compete with elements of the Microsoft 365 portfolio, such as Zoom, Box, Coda, and Cisco, have also touted genAI plans. 

Meanwhile, Apple announced that it will build generative AI features into its range of productivity tools.

Then there are the AI specific tools, such as OpenAI’s ChatGPT, as well as Claude, Perplexity AI, Jasper AI and others, that provide also provide text generation and document summarization features. 

Copilot has some advantages over rivals. One is Microsoft’s dominant position in the productivity and collaboration software market, said Castañón. “The key advantage the Microsoft 365 Copilot will have is that — like other previous initiatives such as Teams — it has a ‘ready-made’ opportunity with Microsoft’s collaboration and productivity portfolio and its extensive global footprint,” he said. 

Microsoft’s close partnership with OpenAI (Microsoft has invested billions of dollars in the company on several occasions since 2019 and has a large non-controlling share of the business), likely helped it build generative AI across its applications at faster rate than rivals. 

“Its investment in OpenAI has already had an impact, allowing it to accelerate the use of generative AI/LLMs in its products, jumping ahead of Google Cloud and other competitors,” said Castañón. 

What are the genAI risks for businesses? ‘Hallucinations’ and data protection

Along with the potential benefits of genAI tools like the Copilot for M365, businesses should consider risks. These include the hallucinations large language models (LLMs) are prone to, where incorrect information is provided to employees.

“Copilot is generative AI — it definitely can hallucinate,” said Slade, citing the example of one employee who asked the Copilot to provide a summary of pro bono work completed that month to add to their timecard and send to their manager. A detailed two-page summary document was created without issue; however, the address of all meetings was given as “123 Main Street, City, USA” — an error that’s easily noticed, but an indication of the care required by users when relying on Copilot.

The occurrence of hallucinations can be reduced by improving prompts, but Dentsu staff have been advised to treat outputs from the genAI assistant with caution. “The more context you can give it generally, the closer you’re going to get to a final output,” said Slade. “But it’s never going to replace the need for human review and fact check.

“As much as you can, level-set expectations and communicate to your first users that this is still an evolving technology. It’s a first draft, it’s not a final draft — it’s going to hallucinate and mess up sometimes.”

Tools that filter Copilot outputs are emerging that could help here, said Litan, but this is likely to remain a key challenge for businesses for the forseeable future.

Another risk relates to one of the major strengths of the Copilot: its ability to sift through files and data across a company’s M365 environment using natural language inputs.

While Copilot is only able to access files according to permissions granted to individual employees, the reality is that businesses often fail to adequately label sensitive documents. This means individual employees might suddenly realize they are able to ask Copilot to provide details on payroll or customer information if it hasn’t been locked down with the right permissions.

2022 report by data security firm Varonis claimed that one in 10 files hosted in SaaS environments is accessible by all staff; an earlier 2019 report put that figure — including cloud and on-prem files and folders — at 22%. In many cases, this can mean organization-wide permissions are granted to thousands of sensitive files, Varonis said.

In many cases, the most important data, around payroll, for instance, will have strict permissions in place. A greater challenge lies in securing unstructured data, with sensitive information finding its way into a wide range of documents created by individual employees — a store manager planning payroll in an Excel spreadsheet before updating a central system, for example. This is similar to a situation that the CTO of an unnamed US restaurant chain encountered during the EAP, said Litan. 

“There’s a lot of personal data that’s kept on spreadsheets belonging to individual managers,” said Litan. “There’s also a lot of intellectual property that’s kept on Word documents in SharePoint or Teams or OneDrive.”

“You don’t realize how much you have access to in the average company,” said Matt Radolec, vice president for incident response and cloud operations at Varonis. “An assumption you could have is that people generally lock this stuff down: they do not. Things are generally open.”

Another consideration is that employees often end up storing files relating to their personal lives on work laptops.

“Employees use their desktops for personal work, too — most of them don’t have separate laptops,” said Litan. “So you’re going to have to give employees time to get rid of all their personal data. And sometimes you can’t, they can’t just take it off the system that easily because they’re locked down — you can’t put USB drives in [to corporate devices, in some cases].

“So it’s just a lot of processes companies have to go through. I’m on calls with clients every day on the risk. This one really hits them.”

Getting data governance in order is a process that could take businesses more than a year to get sorted, said Litan. “There are no shortcuts. You’ve got to go through the entire organization and set up the permissions properly,” she said.

In Radolec’s view, very few M365 customers have yet adequately addressed the risks around data access within their organization. “I think a lot of them are just planning to do the blocking and tackling after they get started,” he said. “We’ll see to what degree of effectiveness that is [after launch]. We’re right around the corner from seeing how well people will fare with it.”

The Copilot for M365 pros and cons

Pros:

  • Boost to productivity. GenAI features can save time for users by automating certain tasks. 
  • Breadth of features. Copilot for M365 is built into the productivity apps that many workers use on a daily basis, including Word, Excel, Outlook and Teams. 
  • Responses generated by the Copilot for M365 are anchored in the emails, files, calendars, meetings, contacts, and other information contained in Microsoft 365. This means the Copilot for M365 can arguably offer greater insights into work data than any other generative AI tool. 
  • Enterprise-grade privacy and security controls. Unlike consumer AI assistants, Microsoft promises that customer data won’t be used to train Copilot models. It also offers tools to help manage access to data in M365 apps.   

Cons:

  • Price. GenAI isn’t cheap and M365 customers are required to pay a significant additional fee each month for access to Copilot features. An individual employee might not need access to Copilot in more than a couple ofM365 apps.
  • Need for employee training. Getting the most out of genAI tools will require guidance around effective prompts, particularly for employees that are unfamiliar with the technology — an additional cost businesses must factor in.
  • Accuracy and hallucinations. LLMs are notoriously unreliable, confidently offering answers that are incorrect. This is a particular concern when it comes to business data, and users must be on the lookout for errors in Copilot outputs.
  • Data protection risks. The ability for Copilot for M365 to access a wide range of corporate data means businesses must be careful to ensure that sensitive documents are not exposed.
  • The Copilot functionality in Excel is limited at this stage.

MacStadium brings Orka Desktop for Devops

Enterprise developers might not be the biggest audience, but they’re a really important one — so the news that MacStadium has introduced a tool that lets them create and manage multiple macOS virtual machines locally matters. It’s a free addition to the company’s existing suite of virtualization tools, which it introduced with Orka Workspace a year or two ago.

The time is right for tools like this. After all, developers are building more Mac, iPad, and iPhone apps than ever before, reflecting the growing market share of Apple products in the enterprise. The snag is that testing is required when building apps both for consumer/customer-facing solutions and also proprietary tools for internal use. 

This is where MacStadium’s new solution comes in. 

Orka Desktop, for free local virtualization

Orka Desktop is a free macOS virtualization tool that lets you deploy virtual Macs locally. What this means to developers, of course, is that they can more easily test apps on virtual Macs, which should help accelerate and democratize the development process.

This is particularly handy as single-use VMs provide trustworthy test results because they act as fresh installs, and subsequent testing takes place on a brand new VM. Use of such ephemeral VMs is becoming an industry standard approach in DevOps.

“Developers regularly use virtualization tooling, but most aren’t aware of the fundamental differences and optimizations that tools can provide for Mac,” said Jason Davis, MacStadium’s chief product officer.

There is a real need for these kinds of tools. One report on Apple use in the enterprise found that three-quarters (76%) of businesses are now using more Apple devices, so it is inarguable that a market for these tools exists.

What does Orka Desktop do?

The developers have put together a consumer-simple, deceptively capable user interface to support the virtual machines; it makes it easier to make changes to those machines from local machines to the cloud.

“Orka Desktop provides the tools to allow developers to create many different macOS image versions locally, commit those into a shared repository, and pass them around for collaboration,” MacStadium CTO Chris Chapman said in a statement. 

Orka Desktop is available now. Useful and powerful features within the system include the following:

  • You can start, stop, pause, and configure VMs easily from within the Orka Desktop admin panel.
  • Free local macOS virtualization provides developers with tools to create, package, and distribute VMs easily without cost.
  • Users get near-native performance on Apple Silicon, with virtualization overhead as low as 5%.
  • Developers can build and share Open Container Initiative (OCI) images, which helps improve collaboration and sharing on projects.
  • The ability to wrap and compress virtual machines using OCI means the size of the disk image can shrink, so a 90GB image becomes a 15GB equivalent, which makes it easier to move the images around. OCI also allows teams to work with most standards-based registries.
  • An easy-to-use GUI interface supports actions on VMs from local machines to the cloud, enhancing efficiency and flexibility.

Existing Orka customers can use Orka Desktop as an extension of their current MacStadium Orka Cluster, which makes it easier to switch from local to global development clusters and product deployment. This enhances team collaboration and offers safer, more reliable testing environments.

Why use VMs?

There are lots of reasons to use of VMs in development. Not only are they inherently easier to secure, but they can be adapted to simulate different hardware configurations — far more cost effective than using actual hardware test machines. They also provide distribution and simulation testing benefits for collaborative development groups.

Of course, this solution isn’t there for Windows or Linux VMs, (for that you may end up working with Parallels or Mac cloud services such as those from MacStadium and/or Amazon), but for Mac and iOS development Orka Desktop makes a lot of sense. Apple also seems to recognize the need to run VMs and recently introduced iCloud support to VM macs.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.

Copilot for Microsoft 365 deep dive: Productivity at a steep price

The AI age has come upon us more quickly than anyone imagined. In just a year and a half, OpenAI’s generative AI tool ChatGPT and its offspring Microsoft Copilot went from a fad to must-have business tools in which the companies are investing billions.

Now the genAI frontrunner, Microsoft is building Copilot into its full product line. There’s a free version of Copilot in Windows and in the Edge browser. There’s a paid Copilot Pro subscription for individuals. There’s a Copilot for Security, a Copilot for Sales, a Copilot for Finance, and many more.

Several months ago, the company released the most anticipated Copilot of them all, the subscription-based Copilot for Microsoft 365, which integrates AI features into the business versions of Microsoft’s office suite. That signaled the end of the hype phase of genAI. It’s now time to see how much it can help businesses in the real world. The rubber’s finally met the road.

Is Copilot for Microsoft 365 right for your business? Is it right for anyone’s business? To find out, I put it through its paces, testing how well it works (or doesn’t work) in Word, Excel, PowerPoint, and Outlook, the core Microsoft 365 apps. I also tested its ability to help you get big-picture overviews that combine information from all those apps.

Based on those results, I made recommendations for whether it’s ready for prime time for businesses, and if so, which companies could benefit from it. Read on to see what I found.

In this article:

Copilot in Word

One way or another, many people’s work revolves around drafting text documents — reports, memos, planning documents, marketing materials, budget suggestions…the list is endless. So odds are that a significant amount of time you spend in Microsoft 365 will be in Word.

How much Copilot will help you depends not only on how much time you spend in Word, but also on the complexity of the documents you write and how comfortable you are with writing. Some people — think of them as the lucky few — sit down at the keyboard, and words start flowing in a well-organized way, everything phrased succinctly and to the point, with very little editing or rewriting required, even in complicated documents. For them, Copilot might not be a tremendous time-saver or productivity booster.

But then there’s the rest of the world. Those who stare forlornly at the keyboard when they’ve got to write a memo. Who feel angst when confronted with a deadline for a project proposal or a marketing document. Who live by the saying: “Writing is easy. You just open up a vein and bleed.”

Those are the people for whom Copilot in Word is designed. And while Copilot won’t solve all their problems, it does quite a good job making efficient writers of most people.

Creating document drafts from a description

Copilot’s help starts the moment you want to create a document. Press Alt-I and the Draft with Copilot screen appears. Describe the document you want to create, including an outline or notes if you have them. Copilot goes to work right away and produces a draft for you.

IDG

Don’t expect anything flashy or unique; Copilot won’t wow anyone with its writing style. It’s workaday and often pedestrian. But for many people in many jobs, workaday and pedestrian is fine, as long as all the information is there, the writing is clear, the document is organized well, and there are no obvious grammatical errors. And I found every time — whether drafting a project proposal, marketing document, or sales pitch — Copilot turned out exactly that kind of document.

Making it more useful is that after it generates a draft, it asks if you want to change it in any way — for example, to make it more or less formal or to make it shorter or longer. You can keep iterating it like this until it reads the way you want it.

srcset="https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?quality=50&strip=all 963w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=300%2C187&quality=50&strip=all 300w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=768%2C479&quality=50&strip=all 768w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=270%2C168&quality=50&strip=all 270w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=135%2C84&quality=50&strip=all 135w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=770%2C480&quality=50&strip=all 770w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=578%2C360&quality=50&strip=all 578w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-02-word-revise-draft.png?resize=401%2C250&quality=50&strip=all 401w" width="963" height="600" sizes="(max-width: 963px) 100vw, 963px">

After you draft a document using Copilot, it allows you to refine the draft in any way you’d like.

IDG

And keep in mind that the hardest part of writing is often just getting a first draft written; you can add flash and pizazz afterwards if that’s what you want. So even experienced writers can find some use for it, because it can quickly generate initial drafts.

Beware of ‘hallucinations’

Copilot at times will do research on its own about the topic you’ve asked it to write about, and include that information in the draft. That can be both a good thing and a bad thing. It’s good when the additional information reflects what you’ve asked it to do. But it can be seriously problematic when it makes up information on its own — especially when that information is incorrect. In my tests, it did that on several occasions.

For example, I asked it to write an email memo to the Director of Data Engineering of an imaginary company I created (Work@Home, which I say sells home office furniture) complaining about data inaccuracies I’ve found. I didn’t provide specific details about what those inaccuracies were.

Copilot, on its own, wrote that I had found “missing values, incorrect labels, inconsistent formats, and duplicate records.” It added in the draft, “I have attached a spreadsheet with some examples of the data errors I have found, along with the sources and dates of the data.”

None of that was true, and I certainly had no spreadsheet that contained the imaginary errors.

I also found that Copilot for Word wrote different drafts for me when I asked the same question on different days. The second time I made the request I detailed in the previous paragraph, it made up far more details than the first, citing nonexistent problems such as “many rows with missing values for important variables such as customer ID, purchase date, and product category… incorrect labels for some variables, such as gender. Some values were labeled as M or F, while others were labeled as Male or Female.”

The draft also complained about outdated information, such as old prices. (Again, none of this existed.) It even added a section with recommendations about how to fix the problems — recommendations that I never made.

The lesson here: You need to very carefully review whatever Copilot creates for you. Copilot, like all generative AI software, is subject to what researchers call “hallucinations” – that is, making up things, and doing what we would call lying if a person did it.

So as always when using genAI tools, it’s important to check Copilot’s output for factual errors, whether you’re using it in Word, PowerPoint, or other apps. Even when it’s creating a draft based on an existing document, it may introduce new material that is incorrect.

I found in my tests that when I included a great deal of detailed information in my request, Copilot tended not to hallucinate like this. So keep that in mind when using it. Asking Copilot to list the sources where it got its information from can also help mitigate hallucinations.

I found another oddity with Copilot for Word. When you create a new document in Word, there are two ways to make a Copilot request: You can click the Copilot icon on the Ribbon or else press Alt-I. When you click the icon, a Copilot pane slides in from the right, and you type in your prompt there. The draft appears in that pane as well. You’ll have to copy and paste it into a Word document manually.

When you press Alt-I, there’s no Copilot pane. Instead, a small “Draft with Copilot” screen appears, and you type your request there. In that instance, the draft is created in the document itself. You’ll be much better off pressing Alt-I, not just because it simplifies your work by creating the document directly in Word.  When do it that way, you can also ask Copilot to use an existing file as a starting point for your draft. (Details about how to do that follow.) You can’t do that using the Copilot right pane.

In addition, I found that when you use Alt-I, you get a more comprehensive draft. Although that’s a good thing, I found that Copilot is much more likely to make up information on its own when using this method to create a draft.

Creating drafts based on existing materials

As I mentioned above, there’s another even more useful feature when creating a new file — having Copilot reference an existing file as the basis for creating a new one. You can feed it several Office file types, including DOC, DOCX, FLUID, LOOP, PPT, PPTX, and XLSX, as well as OpenOffice ODP and ODT files, RTF files, PDFs, and many image files, including GIF, JFIF, PJPEG, JPG, PNG, and WebP.

When you do this, Copilot can take existing information, reorganize and rewrite it, using new information that it finds. For example, I used my initial brief, disorganized notes about this review and created a file from it. Then I fed it into Copilot, which did a credible job of writing a brief description of what Copilot is, listed its pros and cons, and summarized what I wrote.

The more detailed the information in the file you give to Copilot, the better the new document will be. You can, for example, tell Copilot to write a sales pitch based on a marketing document. That’s exactly what I did: I had it write a sales pitch for buying office furniture for those working at home, based on a marketing document — a document that I had previously asked Copilot to create based on my suggestions. In the new sales-pitch document it did an excellent job of highlighting the product’s benefits and making a well-targeted sales pitch. As with everything Copilot creates, the prose wasn’t scintillating, but it did the job.

IDG

This sequence of events showcases how Copilot, when used properly, can be a tremendous productivity booster. It took only about five minutes of typing in notes for me to have Copilot create the original marketing document, and once it was done, only a few minutes of my time to create an accompanying sales pitch. Add in another twenty minutes for checking the drafts and rewriting, and I came away with two well-done pieces of sales and marketing materials. All that took less time than it would take most people to write even a single initial draft of a marketing document.

There is one minor problem with the feature. When you click the “Reference a file” button to choose your existing file, you see only the three most recent files you’ve opened in Word. There’s no way to navigate to others. So if you haven’t opened the starting file recently, you’ll have to open it, then close it. That’s not a tremendous problem, but it’s annoying enough to notice.

Also, keep in mind the source file has to be stored in OneDrive, either locally on your PC, or in the cloud in your OneDrive or someone else’s in your business.

Summarizing documents

Copilot in Word does more than create new documents. You can also use it to edit or summarize existing ones. Open a file, then click the Copilot icon on the upper right of the screen. Copilot’s right pane appears. It will have suggestions for what you can do to the file — for example, summarize it, check it for a call to action, and so on. But you’re not just limited to that. You can also ask Copilot to rewrite it in a more or less formal tone, to reduce its length, and so on.

I was surprised at how well it worked. Summaries were succinct and on-target, it followed my directions for rewrite well when I asked it to make a more or less formal document. It even correctly identified the call to action in a marketing document.

Keep in mind that because it’s a chatbot, you’re not limited to pre-created actions. Ask it to do anything you want. The worst that could happen is it will balk at doing it.

There’s another use for Copilot in Word as well — not just working on your own documents, but getting information from one sent to you by a colleague. You can ask it to summarize the document, its most important points, and so on. You can also ask Copilot specific questions about the document, such as finding a particular data point. I found it surprisingly accurate.

Copilot in Word: The verdict

Copilot in Word surprised me — it was far more useful than I expected. It drafted documents according to my specification, did a very credible job creating new documents based on existing ones, and was quite useful when I asked it to rewrite. Those who have trouble writing will find it exceptionally useful. Even experienced writers may be able to reduce tasks they find unpleasant — for example, if they need to provide summaries of a document to others.

However, there is a significant caveat here. You need to carefully review everything it creates in case it includes inaccurate information. That brings up a larger point for enterprises. If they decide to deploy Copilot for Microsoft 365, they should offer serious training to their employees with a significant focus on how to detect Copilot-created errors, and how to use Copilot in a way that makes it less likely it will create errors in the first place.

Copilot in Excel

Word may be the most used application in Microsoft 365, but still, most of us need to create an Excel spreadsheet at one time or another. For people who aren’t spreadsheet jockeys this can be particularly problematic — how, again, do you create charts? How do you insert formulas? How can you make sense of complicated spreadsheets?

That’s where I hoped Copilot would help. In my tests, though, it didn’t. Unlike in Word, I wasn’t able to invoke Copilot when I created a new spreadsheet, describe what I wanted done, and have Copilot do it for me. Instead, I had to create a spreadsheet as I normally would and face the dispiriting vision of its emptiness laid out in front of me.

And I soon discovered another limitation when I clicked the Copilot icon on the upper right of the screen. The Copilot pane slid into the right part of the screen but said only: “You need an Excel table in this sheet to continue. If you want to see an example, I have one ready for you.” If I didn’t know how to create a table, I’d have to figure that out before I could start using Copilot in Excel.

srcset="https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?quality=50&strip=all 685w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=217%2C300&quality=50&strip=all 217w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=505%2C697&quality=50&strip=all 505w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=122%2C168&quality=50&strip=all 122w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=61%2C84&quality=50&strip=all 61w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=348%2C480&quality=50&strip=all 348w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=261%2C360&quality=50&strip=all 261w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-04-excel-copilot-pane-start-table.png?resize=181%2C250&quality=50&strip=all 181w" width="685" height="946" sizes="(max-width: 685px) 100vw, 685px">

Copilot in Excel initially wouldn’t do anything unless your data was in table format.

IDG

However, Microsoft recently announced some improvements to Copilot in Excel. The biggest one: Copilot will no longer be limited to use in tables. If you format your data with a single row of headers on top, you’ll be able to use Copilot on it.

This change is gradually rolling out to users, and I don’t have it yet, so I’m currently unable to test it. But it remedies one of the major limitations I encountered with Copilot in Excel.

Creating charts

Continuing my testing, I created a table that included revenue data, including monthly revenue. I asked Copilot the most basic of requests: “Chart revenue by month.” I wanted to see whether Copilot would choose the right kind of chart (a line chart), and whether it would then create it for me.

The answer was both yes and no. It knew enough to create a line chart and to select and chart the proper data. However, Copilot created the chart in its side pane, not in the spreadsheet itself. I couldn’t find a way to have Copilot insert the chart into the spreadsheet. I had to ask Copilot to create a new spreadsheet, and then asked Copilot to insert the line chart there by clicking “Add to new sheet.”

At that point I had a spreadsheet that had only the line chart and the small amount of data accompanying it. But I wanted the chart in the original spreadsheet. So now I had to manually copy the chart back to the original spreadsheet. I ended up spending far too much time doing something very simple: creating a basic line chart. I could have saved myself a lot of time by creating the chart myself. Copilot had made me less efficient and productive, not more.

IDG

Again, though, a fix appears to be coming. Among the improvements being rolled out, Microsoft says, are “more conversational and comprehensive answers to a wide array of Excel-related questions.” So you should be able to do things such as getting step-by-step instructions for accomplishing certain tasks in Excel. These instructions can include formula examples. Copilot, Microsoft claims, will also be able to correct and explain formula errors it finds.

Because I don’t have the new features yet, I can’t say how well the improved Copilot lives up to these claims — but Microsoft is clearly working toward addressing a major shortcoming of the initial version of Copilot in Excel.

Finding data insights

Even testing its earlier iteration, however, I found one very useful application for Copilot in Excel. Instead of asking Copilot to do a specific task for me — say, create a specific chart — I asked it to find insights that I might not have found myself in an existing table. With a table opened in a spreadsheet, I clicked “Show data insights” in the Copilot pane. It created a bar chart that showed the total number of days that were required to finish each type of task in the table. I most likely would never have found that on my own — I wouldn’t have even thought of looking for it.

It was, however, quite a useful insight. It could help me do a number of things: more accurately put together a project schedule, identify bottlenecks in projects, examine tasks that took the most time to see if they could be streamlined.

I kept clicking “Can I see another insight” and saw other charts— some useful, some not. I still couldn’t directly insert the charts next to my original table — each chart was placed, along with its data, into a new spreadsheet on a new tab, and I would have to copy and paste them into a new tab or spreadsheet if I wanted to use them in other ways. Still, it’s a powerful tool that does a very good job of mining tables for useful, real-world insights.

Copilot is open-ended enough that you should be able to endlessly query data for useful insight and information in this way.

Microsoft claims Copilot can also help with creating complex, advanced formulas. However, I’m not a spreadsheet jockey, so can’t judge its ability to do this.

Copilot in Excel: The verdict

Copilot in Excel falls short in several important ways, especially for people who don’t have advanced spreadsheet skills. It can’t create spreadsheets from scratch based on a person’s needs, which can be the most difficult and time-consuming task for many people. And when you ask it to create charts, it can’t insert them into your existing spreadsheet — it can only insert them into a new spreadsheet in a new tab.

In its initial iteration, Copilot in Excel would only work with data in tables, and it was less than helpful when asked to perform basic spreadsheet tasks. Microsoft is gradually rolling out changes that address these limitations; you may or may not yet have them in your instance of Copilot for Microsoft 365.

All that said, Copilot in Excel can also be quite useful, notably in identifying insights that you might otherwise miss, no matter your level of expertise. And it may be able to create advanced formulas that could be quite useful. Overall, it seems more suited to people who are comfortable with spreadsheets, rather than those who don’t use them on a regular basis.

Copilot in PowerPoint

Could there be anything more dispiriting than encountering an empty PowerPoint screen when you’ve got to build a presentation for a product launch, marketing plan, or any other reason? A blank Word document is bad enough — but there, you’ve only got words to work with. With PowerPoint, there are visuals, slides, transitions, maybe even multimedia…what could be more intimidating?

Copilot in PowerPoint can help. It can draft an entire presentation for you from scratch, or by pointing it at a document. Will the presentation be perfect from the get-go? No. But it’ll give you a solid starting point.

At first glance, there appears to be no way to get Copilot to create a presentation. When you launch PowerPoint, the Copilot icon on the upper right is grayed out, so there appears no way to use it. Pressing the Alt-I key combination, as you do in Word to have Copilot draft a new document for you, won’t work either.

Instead, you have to create a new presentation or else open an existing one, then click the Copilot icon. When you do that, you’ll be asked if you want to create a new presentation from an existing document, create a new slide in the presentation you’ve opened, or create a new presentation based on your description of what you want done.

IDG

Creating presentations from a description

Asking Copilot to create a presentation from scratch proved to be significantly problematic. Using the Copilot prompt, I asked it to create a marketing document in presentation form for my imaginary Work@Home business selling home office furniture. I didn’t provide a document as a starting point. Instead, I wrote a brief description of what I wanted: “Create a sales presentation for Work@Home home office products.”

Within minutes, Copilot created a comprehensive 15-slide presentation with a remarkable amount of granular detail and accompanying graphics. One slide, for example, touted the Work@Home line of chairs due to their “Ergonomic design,” “Variety of Styles and Colors,” and “Adjustable Features,” along with detailed, paragraph-long descriptions of each of those benefits.  Other slides did the same thing for Cable Management, Desks, Lighting, Optimal Illumination, Monitor stands, and many more.

It was impressive, given the bare-bones instruction I had given Copilot. Unfortunately, it was all purely a hallucination. Copilot had gone out to the web, done research on its own, and created a presentation that had nothing to do with reality.

In subsequent attempts, I provided more details, including the specific number of slides and what each should say. That led to a presentation without hallucinations. But it didn’t save me much time, if at all. I might as well have created the presentation myself without Copilot’s help.

Creating presentations from existing materials

A better approach is to feed an existing document with the appropriate information into Copilot for PowerPoint. (Copilot in PowerPoint can handle the same file types as Copilot in Word, listed above.) If you’ve got the right document as a starting point, that’s the way to go.

When I used the Work@Home marketing document that Copilot had created for me in Word, I got exactly what I wanted. Within minutes, Copilot built a well-organized, seven-slide presentation, complete with graphics and speaker notes, that closely mirrored the marketing document.

IDG

The presentation flowed the way it should, starting with an opening slide and the tagline: Work@Home: The Perfect Solution for Your Home Office. The next slide was an agenda, the next was an introduction, and after that were several slides with the top selling points, another detailing how to order the furniture, and a conclusion. Copilot hit all the high points, made everything succinct, and chose suitable, if quite bland, graphics.

The speaker’s notes left much to be desired — they were just a rehash of what’s on the slide. The graphics were generic and dull. And there were no transitions between slides. Still, it was more than a solid start. For many purposes, you could consider it 75% ready to go; it just needed some polish, new graphics, and animated slide transitions.

As with Word, there was nothing unique here, nothing to knock anyone’s socks off. But it was serviceable. More important, it was a great jumping off point to a solid presentation.

Enhancing presentations, summarizing them, and more

Copilot does more than just create a draft of a presentation. You can also ask it to create new slides, after you describe what you want them to say. Particularly impressive is that it can also gather information by itself about a new slide. I asked it to create a slide explaining that a business might pay for someone’s home office furniture if they work from home. It provided details that I hadn’t suggested, such as asking HR for help in getting reimbursement.

You can also ask Copilot to improve a presentation, for example, asking if there are any slides that should be deleted. I found it was only partially helpful. When I added a nonsensical slide titled “Time to Buy a New Horse” into the presentation, Copilot flagged it immediately. However, when I added one that was off-topic, although peripherally related — “How to Improve Your Posture at Your Desk” — it didn’t flag it.

And as in Word, you can ask Copilot in PowerPoint for a summary or other information about a presentation shared with you by a co-worker in OneDrive.

Copilot can also offer design advice. When prompted, it will have Microsoft’s design tool Designer suggest several design alternatives.

IDG

There are limits to what Copilot can do in PowerPoint, though. I asked it to create transitions between slides and it responded, “I’m sorry, but I can’t do that. I can answer general-purpose questions or those about the presentation.” I asked it to find royalty-free images I could add to the presentation without violating copyright laws. Here, too, it said that was beyond the scope of what it could do, but it did point to several web sites that it claimed were good places to find those kinds of images.

In general, I found a thin line between what Copilot will and won’t do in PowerPoint. My recommendation: Ask it to do anything you want, just in case it can do it. You can also type “View prompts” to get a sample of things it can do that you might not have thought of.

Copilot in PowerPoint: The verdict

Copilot can be a powerful tool for creating presentations, notably by creating presentations from scratch based on an existing document. It can be a tremendous time-saver, giving you a first draft of a presentation in minutes with surprising accuracy. Slides follow a logical flow. It also does quite a good job of creating new slides for an existing presentation when you describe what you want added.

That’s not to say it will do everything you want. It won’t add transitions between slides, for example, and it won’t find new graphics. Also, don’t expect knock-your-socks-off creativity. Rather, you’ll get a solid straightforward starting point, which you can then customize and improve.

Keep in mind that if you ask Copilot to create a presentation via your typed-in description rather than by using a document, you may end up with a presentation riddled with misinformation. So whenever possible, use an existing document. And the same caveat applies here as in Word: Double-check your presentation very carefully for errors and misinformation.

Copilot in Outlook

American may run on Dunkin’, if you believe the coffee chain’s sales pitch, but businesses run on email. So Microsoft made sure to integrate Copilot into Outlook to handle the many problems email causes for people — notably drafting emails, responding to emails, and keeping track of complex, often-rambling email conversations.

Composing emails

I started testing Copilot in Outlook by asking it to help me create new emails. To do it, click the Copilot icon on the Outlook ribbon, select Draft with Copilot, and describe what you want done. If you’d like, select the Generate options icon (it’s at the bottom left of the screen and looks like two sliders) to customize the length and tone of the email. You can choose short, medium, or long and a tone of direct, neutral, casual and formal. (There’s also an option to “make it a poem” that clearly is unsuitable for business communications.)

srcset="https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?quality=50&strip=all 924w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=300%2C205&quality=50&strip=all 300w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=768%2C524&quality=50&strip=all 768w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=246%2C168&quality=50&strip=all 246w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=123%2C84&quality=50&strip=all 123w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=703%2C480&quality=50&strip=all 703w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=527%2C360&quality=50&strip=all 527w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-09-outlook-compose-options.png?resize=366%2C250&quality=50&strip=all 366w" width="924" height="631" sizes="(max-width: 924px) 100vw, 924px">

Choosing options when drafting a new email.

IDG

After it generates the draft and you’ve reviewed it, you can make changes yourself, ask Copilot to discard or retry creating a draft, or ask it to make specific changes — for example, to make it shorter or longer, more direct, or more casual.

I found in every instance that Copilot did precisely what I asked it to do, creating well-organized drafts that laid out all the points I requested. Copilot didn’t hallucinate or add errors at any point, even when I wrote an email complaining about data errors someone’s team had made — essentially the same request I had made of Copilot in Word. However, you should always double-check any email crafted by Copilot for errors.

 Although the emails were clear and to the point, they also tended to be quite stilted and came across as being written by an AI chatbot rather than a human being. This happened even when I asked that Copilot use a casual tone. So you’d do well to spend a few minutes editing the draft to make it sound more human and like yourself.

For my ultimate test of drafting new emails, I set Copilot a task every manager dreads: sending a stern message to a team member warning that his work has been subpar, he comes late to work too often, and he misses deadlines. I also asked that he come to my office for a meeting.

Copilot did an excellent job, not only accurately portraying the issues I asked it to address, but getting the tone right — in fact, likely better than I might have done. It’s easy in pressure-filled situations like these to let one’s emotions seep into the email. Copilot, being an AI, doesn’t have emotions, and so the message it sent was less confrontational and likely to be a better received than one that I might have sent. For example, Copilot’s email concluded, “I look forward to our discussion and finding a solution to these issues.” The language may sound stilted, but it was spot on for the task.

Replying to emails

Copilot shines when replying to emails as well, at least if you use it correctly. There are two ways to respond to an email: you can give Copilot specific suggestions about what you want the email to say, or you can allow it to automatically respond on its own, without your input.

I tested the feature by responding to an email sent to me that had a recommendation for new lines of business to help revive Work@Home’s flagging business. First, I had Copilot respond automatically.

It gave me four choices for general ways to respond: “Working on memo,” “Memo attached,” “Need more time,” and “Custom.” The first three ways resulted in a generic, stilted response whose recipient would know it was drafted by AI. I’d recommend never using the first three automated responses, because it could come across as almost offensive to the recipient that you didn’t bother to take the three or four minutes yourself to draft a simple email response.

The fourth choice, Custom, allows you to describe the draft you want Copilot to create. This choice is a winner. I asked Copilot to thank the sender, summarize his recommendations, and ask for a more detailed memo of 1,500 words by the following Tuesday.

IDG

Within about 15 seconds, it created a draft that was right on target. Not only did it follow my request exactly, but it put in all the niceties that can make business communications flow smoothly rather than being misunderstood or fraught with tension. It noted that the recommendations “all seem like promising options,” for example. And its phrasing was similarly thoughtful throughout — for example, “I would appreciate your insights and further thoughts on how we can move forward.”

Yes, it’s true those are timeworn phrases and can feel like throwaway clichés. But there’s a lot to be said for simple politeness and directness in business communications, and in this instance, Copilot did a very good job of getting the tone right, even if the phrasing felt at times a bit wooden. That’s why I recommend that all email communications created by Copilot not just be checked carefully for accuracy, but also edited to make each message sound like a human being composed it, not a chatbot.

Before sending any email message drafted with Copilot, whether it’s a new email or a response to received mail, Outlook flashes a warning: “Check the message. AI-generated content may be incorrect, so be sure to check the message before sending it.” Make sure to follow the advice. In my testing, Copilot didn’t make errors, but that may have only been luck.

Copilot also warns you if you send a message without a subject, which can be useful for people who draft so many emails in a day they forget at times to do that.

Email draft coaching

There’s another Copilot in Outlook feature that I found quite useful: having Copilot analyze a draft of an email you’re planning on sending and suggesting recommendations to improve it. To use it, select Coaching by Copilot from the drop-down menu that appears when you click the Copilot icon in the toolbar.

I wrote a rambling, somewhat incoherent, clearly offensive email full of typos critiquing someone’s work, and had Copilot offer its recommendations. It was absolutely on target and pulled no punches, finding serious problems with the tone and clarity of the writing, as well as how the recipient would likely feel after receiving it. It went well beyond generic advice and offered concrete suggestions for improving individual words and sentences.

For example, about the tone it wrote: “Be more respectful. The email is very informal and casual and might make the reader feel like their work is not taken seriously. The email also uses language that is vague, dismissive, and insulting, such as ‘pretty good,’ ‘kind of,’ ‘suck’ and ‘do lot’s better.’”

IDG

Summarizing email threads

Outlook also does a good job of helping solve one of the most problematic and time-consuming of email tasks — reading through a threaded conversation made up of multiple emails, and identifying the gist of what occurred.

In my tests, I created a thread between two people, one whom asked the other for recommendations on how to revive the Home@Work business, which was starting to flag because fewer people were working remotely from home. Copilot succinctly summarized every important part of the conversation, including mentioning that a spreadsheet with projected new revenue figures was sent in one email.

IDG

Copilot was also able to separate the important topics from less important ones. Within the email thread, I had the two people discuss a concert by the guitar player and songwriter Richard Thompson, including a description of some of the finer points of Thompson’s technique. Copilot recognized that it was a side issue and included that information as a minor aside at the end of the summary. I also included typos in the emails to see whether Copilot would repeat them in its summary. It didn’t; it corrected them each time.

There’s one important thing that Copilot doesn’t do, though — look inside the attachments people send to one another and summarize the data therein. It’s a vital shortcoming, because frequently, more succinct detailed information is available in attachments than in the threads themselves. In this instance, each of the people sent revenue projections to each other. That’s important information that Copilot ignores.

Still, even without that, summarizing email threads is a useful and powerful time saver and productivity booster. For people who don’t need help drafting emails, it will be the most important Copilot feature.

Copilot in Outlook: The verdict

Used properly, Copilot can help you draft new emails and responses more quickly, be more precise in your use of language, and help you choose the right writing tone. It also does a good job of summarizing complex email threads. That being said, its emails generally have a stilted, somewhat artificial tone. You’d do well not to send them as is, but make them sound more natural and personalize them in some way.

Copilot for Microsoft 365: Getting big-picture information

One of Copilot for Microsoft 365’s more intriguing features is its potential to give you a big-picture view of your projects, and then let you drill down into any of them to get more granular information — for example, a specific spreadsheet that has revenue projections for the next five years for new lines of business.

It does this by looking through all your emails and files (including those shared with you by co-workers) based on your request, analyzing what it finds, and then giving you a succinct summary of it all. Based on that, you can ask for more detailed information or follow one of its suggested prompts that recommends data you might want to see.

You do this in the Copilot app itself, rather than starting in Word or another Microsoft 365 app. Make sure you’re signed into your business account in the Edge browser. (If you’re signed into your personal account rather than your business account, it will search the web, not your files and emails.) Then click Edge’s Copilot icon on the upper right of the screen and type in what you’d like it to do. (Note: Your organization may have set up this search differently, so you should check with your IT department when trying to do this.)

I tested out Copilot’s summarization powers with a simple prompt: “Tell me more about Work@Home.” It responded by giving me a clear and accurate description of the imaginary business I had created, and then noted, “There have been discussions on how to revive Work@Home in a post-COVID world,” and briefly and accurately summed summarized a memo I had written about how to do that.

srcset="https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?quality=50&strip=all 447w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=170%2C300&quality=50&strip=all 170w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=396%2C697&quality=50&strip=all 396w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=95%2C168&quality=50&strip=all 95w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=48%2C84&quality=50&strip=all 48w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=273%2C480&quality=50&strip=all 273w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=204%2C360&quality=50&strip=all 204w, https://b2b-contenthub.com/wp-content/uploads/2024/07/copilot-m365-13-copilot-app-business-summary.png?resize=142%2C250&quality=50&strip=all 142w" width="447" height="787" sizes="(max-width: 447px) 100vw, 447px">

Copilot does a stellar job of giving you big-picture information and granular data by looking through all your documents and emails.

IDG

So far, so good. It was an accurate summation. Next, I asked it to show me any emails that have details about new lines of business for reviving Work@Home. Again it was on target. It summarized the email and included a link to it. All I needed was to click the link to get to it.

I then asked for any presentations related to Work@Home, and once again it succeeded. It summarized the presentation, identified the presentation’s author and when it was last modified, and provided a link to it.

It wasn’t always perfect, though. I found that I had to offer as precise a description as possible when asking for documents, or it might not find them. But overall, the results were impressive.

Getting big-picture information: The verdict

Using Copilot to gather and summarize information from multiple documents and emails could eventually prove to be one of the most powerful uses of Copilot for Microsoft 365. In a typical workplace, you’ll be involved with multiple projects at a time, and can’t possibly know detailed information about all of them. An ability to quickly get updates on projects and find the specific information, email, or document could be a tremendous time saver and productivity booster.

Copilot for Microsoft 365 extras

Is Copilot for Microsoft 365 right for your business?

Copilot for Microsoft 365 can be an impressive productivity tool, especially for those who often use Word and PowerPoint. In Word it can draft entire documents merely by your prompting it or based on existing documents. The documents are generally well-organized and well-thought-out, and are excellent for first drafts, although editing and rewrites will be required. It’s particularly powerful for those who have a hard time getting started on creative content such as marketing documents. However, it has a tendency to hallucinate, and all its work needs to be very carefully vetted as a result.

It’s helpful in PowerPoint in much the same way as it in Word. You can create first drafts of an entire presentation by merely pointing PowerPoint to a Word document. It can also create new slides based on your prompts. Once you have a presentation, however, it won’t perform tasks for you such as creating transitions. Also, if you ask Copilot to create a presentation from scratch, rather than from an existing document, you may end up with one filled with hallucinations and misinformation. In all cases, as with Word, carefully vet Copilot’s output.

It’s not nearly as impressive in Excel, at least in its first iteration. For those who don’t frequently use spreadsheets, it won’t offer much help — and it has limited capabilities even for spreadsheet jockeys. It does, however, do a fine job of mining spreadsheets for insights you might otherwise have missed, and new features being rolled out may fix its most serious shortcomings.

Copilot for Microsoft 365 will likely be a productivity booster for almost anyone using Outlook, by helping to draft new emails and to respond to existing ones. As with Word, you’ll need to check it for errors, and you should rewrite its sometimes stilted language.

In my tests, Copilot also proved to be surprisingly helpful in providing big-picture information by searching through emails and files. It was equally useful when looking for a specific file or piece of information.

As for the extras, such as transcribing Teams meetings and summarizing notes in OneNote, those are nice-to-haves, not must-haves.

Also note that Copilot currently supports only about 25 languages. Microsoft says more are planned.

Keep in mind that, like the company’s other subscription-based products, Copilot for Microsoft 365 frequently gets new features. In addition to the important improvements to Copilot in Excel already discussed, upcoming Copilot enhancements include prompt autocomplete and rewrite features, as well as a Catch Up feature that can update you about meetings and shared documents. And Microsoft recently announced that users will soon be able to have Designer create images for them in both Word and PowerPoint.

So should your business use Copilot for Microsoft 365? If it were free or included in a Microsoft 365 subscription, the decision would be a no-brainer. It clearly boosts productivity, and for some people will boost it quite a bit.

But it’s not free — far from it. It costs businesses $30 per user per month, and that needs to be paid on an annual basis, not monthly. There’s also no trial version, so there’s no way to test it out for a few months and see whether it’s worth the price.

That $30 per month can increase the monthly outlay by more than 50% for subscribers to the top Microsoft 365 enterprise plan — and by a whopping 600% for subscribers to the entry-level small-business plan. The Copilot add-on fee adds up quickly, even for large enterprises. For a business with Microsoft 365 subscriptions for 500 people, it means an additional $15,000 per month — and enterprises typically have well over 500 seats.

All that said, Copilot for Microsoft 365 includes enterprise-grade privacy and security protections, including Microsoft’s promise that your company’s data and your users’ prompts are not used to train the large language models used by Copilot. These protections are not offered by consumer genAI chatbots.

For most businesses, it probably makes sense to start small, offering subscriptions to targeted users or departments that might get the most out of Copilot — for example, people or groups who most need help in drafting documents and presentations. Based on the results, a company can roll it out to others who might get good mileage out of it.

As for going in whole hog and buying it for everyone who has a Microsoft 365 subscription in your business, that’s an expensive proposition and likely won’t come close to paying for itself in productivity increases, at least for now.

Should you decide to try it now or sometime in the future, there is one caveat that I feel compelled to repeat: Copilot sometimes makes up false information and presents it as fact. You’ll need to very carefully examine everything it does before using its output, and will need to balance that step against any productivity increases you expect.  

You should also consider launching a formal training program for anyone in your business who uses Copilot, targeted primarily at making sure people are aware of potential hallucinations and errors, and teaching best practices for creating documents so they won’t be rife with errors.