![]()
For the past few years, Microsoft 365 Copilot and Copilot Pro have assumed the role implied by their names. That is, they were assistants that worked alongside the user, helping them with classic Office apps. Tied to this interaction model, these offerings were monthly bills for customers that were separate from the underlying Microsoft 365 commercial or consumer subscriptions one had to also pay for.
Those who follow these things closely recognize these Copilot implementations as examples of the “beside applications” AI app structure that Stevie Bathiche discussed during his famous Build 2023 talk. These people also know that Stevie discussed two other AI app models, “inside applications” and “outside applications.” But with a few years of experience and improvements in the technologies and tools, a more nuanced view of how we will use AI is emerging.
Many probably assumed that those three AI app structures would progress in order and that they collectively represented an evolution of sorts. And that is true to some degree: Each app structure, in turn, is more sophisticated than its predecessor, and each requires more and more new code, not just on the AI side but on the app (or service) side. But it’s now obvious that these app structures won’t replace their predecessors, in turn, in waves of new code. Instead, they will exist side-by-side. The mix will shift in time, but this mix of app structures will always be with us.
We know this to be true for two reasons. Common sense, in that not all apps (and services) will be rewritten or modernized and will thus still benefit from “besides application” solutions. And experience, because this happens all the time. For example, your Windows 11 PC has classic desktop apps, modern apps, mobile apps, web apps, and other types of apps all running alongside each other today.
Watching Microsoft evolve its Copilot-based solutions is a full-time job and understanding it all requires a team of people. This creates a day-to-day “fog of war” effect that’s exacerbated by having to deal with the opposing sides of this incredible shift in how we interact with personal computing products and services. On one side are the AI deniers and doomsayers, different camps, yes, but with a shared view of AI as a threat. And on the other are the naive Pollyannas who believe that everything happens for a reason and that AI will deliver us into a Star Trek-like future nirvana.
There’s a lot of noise out there with the chaos.
Without overtly judging those who land on the extremes, I’ll just say that I almost always lean toward the center. I see both sides, disagree with a lot of it, and agree with some of it. But I do view AI as both inevitable and beneficial overall. The only questions in my mind are how and when. To date, the “how” is up in the air, though there are some early winners like OpenAI that may or may not emerge victorious in the end. And the “when” has typically been immediately and constantly: We’re still getting new AI advances on a weekly basis and sometimes even more quickly than that. We do live in interesting times.
Keeping this to Microsoft and its Copilot offerings, I have stressed over what appears to be chaos and unclear, ever-changing strategies. Microsoft gets points for kicking this whole thing off, and even before that for seeing the promise of OpenAI and locking them into a partnership that is both mutually beneficial and mutually destructive. But we can’t look to Microsoft for a clear vision of the future. This company has changed directions more times than I can count. It’s pretty clear that it has been metaphorically throwing everything it can against the wall to see if anything sticks.
By all accounts, little has.
While Copilot has its defenders, oddly, my wife among them, most of the data we’ve seen suggests that Microsoft, despite its first to market advantages and OpenAI partnership, is floundering. Few businesses and consumers are using it, let alone paying for it. According to the most recent report on this topic, between 1.7 and 2.7 percent of all commercial Microsoft 365 seats also pay for a Microsoft 365 Copilot subscription, and most of those are getting a deep discount from the software giant. This business of selling AI directly to customers has not been successful. So something had to change.
Something did change. Today, Microsoft announced Microsoft 365 Premium, a new, um, premium tier of its consumer-focused productivity service and app suite. This $19.99-per month subscription is aimed at “solopreneurs, professionals, and achievers of all kinds” and it combines everything in Microsoft 365 Family (which is also getting some updates) with higher Copilot usage limits and exclusive Copilot features.
If you’re cynical (or just hilarious) you might argue that “more Copilot” is the new “more cowbell.” Or that this shift is tied to Microsoft’s failures in selling Copilot as standalone AI. I certainly see this side of it, and I look at throwaway lines in Microsoft’s announcement—like “Microsoft 365 Premium offers a greater value than ChatGPT Plus”—as leaning a little too obviously into the company’s perceived bitterness over OpenAI’s successes in a field it intended to dominate.
But we should also look at this through two other lenses, both of which matter more than our cynical natures: Why this matters strategically to Microsoft and how or whether this improves things for its customers.
Strategically, this is Microsoft moving nimbly to take on OpenAI, a company that is both partner and competitor, friend and foe. And the gloves are off: Microsoft repeatedly references ChatGPT Plus, another $20 per month offering, in its announcement. It notes that ChatGPT Plus is a “standalone” offering that forces users to copy and paste between apps, extra work that is not required with Microsoft 365. And that ChatGPT Plus doesn’t cover the cost of all the productivity apps you get with Microsoft 365 Premium (on desktop, web, and mobile), doesn’t include the 1 TB of cloud storage per user, and doesn’t offer advanced security. (In a curious nod to OpenAI as a partner, Microsoft does admit that ChatGPT Plus has two features, custom GPTs and limited access to Sora, that Microsoft 365 Premium lacks.)
This is important because Microsoft cannot afford to lose in AI. CEO Satya Nadella sees this as a potential extinction moment for Microsoft, as it will become less and less relevant and necessary if its customers move to other AI productivity solutions. I still see Microsoft as being successful in a world in which it hosts other AI solutions on Azure, but I also see big risks in its insane datacenter buildouts at ever-escalating costs and with no clear path to profitability. But I do appreciate how definitive Microsoft is here, and even the speed at which it now moves and adapts.
But the bigger strategic move here, to me, is Microsoft moving Microsoft 365 apps, things like Word, Excel, and PowerPoint that are inarguably as legacy as any apps can be, into the “inside applications” structure. Instead of using Copilot exclusively from a sidebar, a literal copilot scenario, it is now “embedding AI capabilities like Copilot deep into the familiar apps people use every day.” This is Microsoft playing to its historic strengths, meeting customers where they are, and bringing everyone forward using familiar tools that businesses, especially, prefer and not requiring anyone to learn new apps.
For individuals (customers), Microsoft says it is “simplifying the lineup,” and while replacing Copilot Pro with a new Microsoft 365 Premium subscription doesn’t seem like a net gain, it is. Previously, one had to pay for Microsoft 365 Personal or Family and then pay another $20 per month just to get access to Copilot Pro. Now, one can pay $20 per month for Microsoft 365 Premium and get it all. Not to mention the various improvements.
This doesn’t solve the single biggest issue with Microsoft 365 Family when it comes to Copilot and AI: As with that offering, only the Microsoft 365 Premium account holder gets access to these Copilot/AI features. And yes, that sucks. But this is still a step in the right direction, and if there is a family of people somewhere who were paying for multiple Copilot Pro subscriptions for some reason, it will be cheaper now to pay for multiple Microsoft 365 Premium subscriptions instead. This is simple math, people.
And speaking of math, I know there are those out there who viscerally and negatively react to any news about AI creeping more and more into the apps and services they use. Microsoft 365 Personal and Family can still be used a before. You don’t have to pay more for Microsoft 365 Premium. You could even buy standalone versions of Office if you like paying for software. You have choices. (Including looking elsewhere, of course.)
Actually, there’s a third lens through which we might view this change. And this one is a big deal for me, personally. A little over a year ago, I opined that I will not pay for AI. This was widely misunderstood, and I never meant that I would not use AI. I meant what I wrote: I will not pay for AI directly.
In other words, I will pay for things that offer AI capabilities as part of a broader set of features. Because to my mind, AI is not a thing, it is a million things that are expressed as individual features in the apps and services I do use. For example, I pay for Affinity Photo, which has some number of AI-based features. But I will not pay Affinity separately for those or other AI features in addition to paying for the app itself. That’s ridiculous.
In the Microsoft space, my issue with things like Microsoft 365 Copilot and Copilot Pro was that the features and functionality they delivered are the types of things subscribers used to just get “for free” as part of their ongoing paid Microsoft 365 subscriptions. This always felt like nickel and diming customers, and still does. It’s not right.
But now I can make a case that Microsoft 365 Premium as an offering maps to how customers have traditionally consumed what I still think of as Office. So where I will not pay for AI, I will pay for an office productivity suite with cloud storage and other advantages, just as others would and do. And I will pay even more for that suite if there is a more powerful and useful version that benefits me. That its advantages over the existing offerings come largely from AI is neither here nor there. Of course they do. AI isn’t a thing you buy. It’s a feature of the things you buy. In this case, many hundreds or even thousands of features.
It’s not perfect. It’s not the answer to all our complaints and concerns. But it is a step in the right direction.
With technology shaping our everyday lives, how could we not dig deeper?
Thurrott Premium delivers an honest and thorough perspective about the technologies we use and rely on everyday. Discover deeper content as a Premium member.