I admit it: I use GenAI tools on a daily basis.

I also admit the following: if I wasn't doing consulting work in the Microsoft ecosystem, I would not choose to pay for Copilot on my own. I’d most likely use the $30/month for some other AI tools.

That's the problem Microsoft and other tech vendors have faced: people at large aren't choosing to pay for new AI features in existing apps. Which leaves them seemingly only one option: force the AI features onto everyone, no questions asked.

Today, I’m not here to argue about what Copilot technically can & can't do. It’s just software - it sometimes works and often doesn't do what is advertised. I’ve spent most of my professional life trying to make the needs of customers and the abilities of tech vendors meet each other, as transparently as possible. I know the struggle and I write about it because I care about the end results for the whole user community.

What I do want to argue today is that the way Microsoft has decided to force Copilot onto especially its consumer customers is a big mistake. One that may cost them dearly in the long run. Not merely due to the adoption rate of AI tools from Redmond, but also via the eroding trust that is apparent among the most loyal advocates. Several writers that have been covering MS technology for decades are sharing the sentiment expressed by Ed Bott here:

The audience of paying customers has seen this trifecta of “reimagining” from Microsoft in the recent weeks/months:

1. Changing the name of the tool.

2. Changing the price of the tool.

3. Changing what & how the tool does.

Each round is more impactful than the prior one. In combination, they add up to a wave that deserves to be documented in one place. Those who follow me on social media may have seen my takes on these topics earlier, often in the form of meme pictures pointing out the absurdity of it all. I feel like it needs to be preserved also here in my newsletter, so that we may reflect on it after the dust has settled.

1. Changing the name: killing Office one more time

Do individual product names matter? Maybe not that much, but brands most definitely do. Back in 2020 the decision to give up on the iconic Microsoft Office brand felt like an awkward move. There was a certain logic in changing the commercial packaging to be called Microsoft 365, due to everything that was bundled into it. Still, witness the persistent use of Office as the term actively used in MS ecosystem (and even in product features) five years later. That tells you something about the power of branding.

The brand “Microsoft 365” was a bit boring, with just the name of the corporation used in there. A bit “meh” but easy to understand - which is not a bad outcome if you look at MS branding efforts in general.

Then 2024 came along. What’s the one name that Microsoft today loves even more than its company name? That’s right! We’ve yet to see the rebranding of the corp itself to “The Copilot Company”, but adding the Copilot postfix to the M365 app sure is a step down that same road. This is the official messaging to software admins:

First of all: that icon. Really?

Nothing says “well-polished branding strategy” quite like slapping what looks like a black & white mugshot board on top of an existing Copilot product icon and calling it a day. If the young Bill Gates was still working at the company, he could have claimed to having inspired the design:

People have noted that the Microsoft 365 Copilot app icon is almost impossible to decipher on low-resolution screens. If there had been any UI and accessibility experts involved in the process, Microsoft could have noticed it, too.

Then onto the name. On the business customers side, there had already been a “free” Copilot available inside the Microsoft 365 web portal for quite some time. That didn’t stop MS from making a splash by announcing “Copilot for all” in mid-January:

By “for all” surely you mean that the M365 Copilot that was previously $30/user/month is now available to all paying M365 users?

Hah! Where did you get that idea from! No, we are instead calling the Copilot UI inside M365 “Copilot Chat”, which is yet a new name for a thing that existed already before. Kind regards, MS Product Marketing.

One might imagine that the intention with such marketing activities would be to increase the sales of premium products among the customer base. Now, how does that goal align with the fact that you just renamed the existing app to the exact same name of the premium $30 app you wish customers would buy?

Can you imagine what it’s like when customers realize that the app that was just renamed to “Microsoft 365 Copilot” won’t let them use all the advertised AI features - unless they buy a license called “Microsoft 365 Copilot”? I can, it’s gonna be something like this:

The real cherry on top can be found from the Microsoft 365 app transition documentation page. Guess what happens to MS customers who happen to live in a region where the AI services for Copilot are not available? They get an app called “Microsoft 365 Copilot” that does not have Copilot features - because consistency.

It has been awesome to see that these branding antics Microsoft has pulled are seen as masterfully dumb by basically everyone who has the courage to tell what they feel, without fearing retaliation from their employer. Although this article seems to have been “rewritten for clarity” after being published, the original version still found in the Internet Archive had the most descriptive headline imaginable:

Another newsletter author who had previously worked at Microsoft for 26 years posted not just a delightful header image on the topic “Adios, Office” but also included a reaction video that’s perfect for pretty much every branding decision at MSFT.

Okay, so everyone who just wanted a convenient shortcut to access all their Office docs now had their Microsoft 365 app renamed. Still, is that a big deal for the general population? The folks who just go straight to Word on their PCs, are they gonna notice anything different?

Oh yes they are. Let’s enter enforcement level 2: pricing.

2. Changing the price through an offer you can’t refuse

The price of business M365 subscriptions has not yet been raised. Consumers who have subscribed to Microsoft 365 Personal or Family have been informed that the price they’re paying will go up around 30-40%. For a detailed list, see the Microsoft 365 price increases around the world article from Office Watch:

When you’ve got bad news to tell your customers, like a price increase for their existing subscription, you’d normally want to be very careful in how it is communicated. But Microsoft is no normal company - they are now The Copilot Company! Meaning, some of the communication may contain hallucinations and the recipient should always check the information validity before doing anything with it.

On January 22, customers across Europe (possibly also elsewhere) began reporting a strange prompt that their OneDrive/Word/any other Office app on their Android phones had shown to them. The message text was about the coming changes to the subscription package, noting that starting Feb 20 there would be a new price automatically billed from them. Only the new price reported wasn’t €10 as expected. It was as much as 100x that:

It would have been a hilarious error, if only the operating costs of AI services wouldn’t be a globally hot topic already. I have yet to see an explanation of how the price announcement got messed up so badly. Users who received this note from “Google Play Store” say they have not bought the M365 subscription through it - and as a result, there was no place to go and validate the info in their profile either. You would think online commerce would be a fairly robust process for tech giants like Microsoft (or Google), yet here we are.

Invisible pricing of Microsoft products has been a topic widely circulated on social media recently, already before that 100x blunder. People who have voiced their anger at the M365 price change due to forced AI features have received the tip that a “secret deal” for holding on to the old price level and feature set is actually available. The trick is to go and attempt to cancel your plan first, at which point an offer for Microsoft 365 Classic may (or may not) be shown to you:

Isn’t Microsoft scared that they’ll end up losing money with this price increase if they customers see it as the final straw to cancel their personal/family M365 subscription altogether? In his awesome article titled “The Microsoft 365 Copilot launch was a total disaster”, Ed Bott asked Copilot to explain the numbers behind the business case for making the price change. Since the price hike is so large, it would still be a financially solid outcome for MS if one in ten customers decided to cancel their subscriptions for good:

When you’re a global hyperscaler like Microsoft, even small changes can have a big economic impact. A 40% change in a fee for tens of millions of customers using your cloud service can lead to massive revenue increase. It even made Copilot ask Ed in its response: “What are you planning to do with all that extra revenue?”

Microsoft’s answer, for the time being, appears to be: give customers more of the AI features they aren’t asking for.

3. Changing the tool into something customers don’t want to use

The fact is that many subscription prices go up over time. Especially with the recent inflation felt in many countries, the M365 price hike might have been just another everyday event for folks. A bit like Netflix notifying you of yet another price update.

The online crowd who follows tech news had a chance to learn about the subscription price going up already in advance. What they didn’t have a chance to mentally prepare for is the arrival of Copilot inside the productivity toolkit they depend on for getting work done.

Even if it’s a personal/home subscription, those folks who have chosen to pay for Word, Excel & co. rather than using Google’s tools on the browser for free are likely doing it for a good reason. To do something of importance, rather than just casually opening docs. It has been a trusted toolkit, despite of not carrying the trusted Office brand anymore.

In the past few days, it has been evident in social media channels that the biggest outrage comes not from the branding or pricing change, but from injecting AI into tools that used to be clear from it. So many customers are desperately looking for a way to disable Copilot - and they’re not finding a satisfactory answer for it.

Some aren’t doing it merely for the sake of preserving their own productivity. They may have real concerns about how these forced AI tools are interfering with their private data and potentially resulting in unintended contractual breaches in more regulated fields of work.

The reason this is happening is that Microsoft chose to build it this order. First, develop a way to force Copilot onto all users of Word, Excel, PowerPoint, OneNote, Outlook. Then, listen to the feedback coming from enraged customers. Only after that, acknowledge the need for off switches and start building them. Leaving your tens of millions of customers relying on rumors that they read and exchange online.

“Big deal, new features pop up all the time in software these days” you may think when reading online posts from people that have flagged forced Copilot in Office apps as an issue. We must not think about it only from a software development perspective, no matter how deep into tech we are. There are professions in this world, such as lawyers, who have a duty to pay attention to aspects such as client data confidentiality. They use Word, and now they are rightfully concerned about whether they are suddenly violating rules that could cost them their jobs:

It is at this moment when the patience of the most loyal MS advocates can quickly run out. The number of rants and parodies seems to be only going up, as people have begun searching for answers. Then, when the answers found make them realize the whole 1-2-3 process of Copilot enforcement, they’re gonna feel… something. Some will settle for a few angry comments on how M$FT suxx. Others will put in more effort, to express their sentiment through the creative channels and content formats familiar to them.

One beautiful example being Atomic Shrimp, who illustrated this episode of forced Copilot rollout with a video built around the story of a software vendor introducing the feature of a screaming goat:

Go and watch the full video, you’ll love it:

Not only is the frustration experienced through unwanted AI features described in a relatable way. Atomic Shrimp also goes on to explain how in reality there was no price increase for the existing Microsoft 365 product. Rather, it has been an elaborate scheme where everyone is automatically upgraded to the new, more expensive tier with Copilot, while preserving the existing SKU as-is but just putting it behind dark UX patterns and the “Classic” branding.

Which leads us to the most fundamental question of them all. If Copilot has to be forced onto all customers via such marketing tricks and shady practices, is the product any good?

Up until this point, we’ve mostly heard from Microsoft and their partners on what are the multitude of benefits that Copilot can deliver to users. Some early adopters have of course used it and I bet many have found ways to get actual value from the AI features. Quite possibly not on the same level as advertised - yet that wouldn't yet make it a failed product.

What can make it a failure, though, is the public’s perception of the product.

Will they ever love you, Copilot?

Even if you are working in the domain of business software and technology bought by organizations, you should never, ever underestimate the importance of individuals. The humans who are part of not just formal decision-making processes but also those who act as influencers insider the work community. Their feelings and reactions matter.

Who brought iPhones into the office first? Tip: it wasn’t the IT. The whole concept of BYOD for born from the fact that people don’t like being forced to use tools. Those who spent their own money in buying an iPhone were not against the use of smartphones as a tool - on the contrary. They saw it as far too important and personal to be subjected to the Windows Mobile devices chosen by corporate IT.

Back then, it didn’t help Microsoft that they had a massive distribution channel of existing corporate customers in place for Windows PCs and Office software. When the mobile computing revolution took off, launching devices with Windows Phone software and Nokia hardware was no longer a formula that played to the advantages of either of the incumbents. It was perceived as a continuation of forced, suboptimal choices made by someone else high up in the corporate hierarchy. Apple, on the other hand, offered a choice for the individuals themselves to make.

Just like there’s been Shadow IT at work since the rise of SaaS, we’re also in the middle of the Shadow AI era now. Only this time it’s not going to be as clear to identify what exactly is the object lurking in the shadows. Unlike devices or apps, AI is already an omnipresent forcefield found inside all technology products. I don’t believe we yet know whether it will ever turn into specific products to be sold, in the way that Microsoft is attempting with Copilot. The recent changes in M365 products and pricing show it’s hard to find the right price to charge for AI.

Choosing to bundle AI into existing products and then force it upon your customer base is prone to cause damage to the brand image. Now that many private users of Office tools have experienced Copilot through these various events described above, what do you reckon they’ll think about it if/when corporate IT decides to give Copilot to them at work? Will they see it as an iPhone, or a Windows Phone?

Putting all technical features aside - which in the case of AI chatbots are incredibly hard to illustrate compared to apps, I suspect the general audience may not consider Copilot to be the most desirable AI tool out there. They may well want to hold on to ChatGPT, even if the underlying foundational models would be the same. Because they honestly don’t care about the chips inside their phones or the models behind the chatbot. They care about how it makes them feel.