You’re paying more for Microsoft 365. You’re getting less. And the AI feature Microsoft spent billions hyping? Almost nobody uses it.
After three years, Microsoft 365 Copilot adoption sits below 4.5%. Weekly active users? One percent. One. And yet — somehow, against all logic — prices went up.
Let that sink in. Microsoft is charging you more for a product that its own users are actively ignoring.
The tech industry doesn’t have an AI problem. It has a delusion problem.
Here’s what’s actually happening: Copilot was sold as a revolutionary productivity assistant that would transform how we work. In reality, it’s an entertainment product masquerading as enterprise software. Even Microsoft’s own documentation admits it’s partly for entertainment. You don’t pay $30 per user per month for a chatbot that occasionally summarizes an email and then hallucinates a meeting that never happened.
But buried inside this flop is a twist nobody’s talking about.
There IS a feature users love. It’s called Co-Work. It gives you autonomous agents that handle recurring tasks — the kind of thing that actually saves time, the kind of thing you’d use every single day. One commenter put it perfectly: “Finally something that actually actively helps.”
So what did Microsoft do with the one feature people actually want?
They siloed it. Billed it separately. On a token basis. So the more you use it, the more you pay.
The good AI isn’t included in your subscription. The useless AI is. That’s not a pricing strategy — that’s a hostage situation.
This tells you everything about where enterprise AI is right now. The industry has decided that general-purpose chatbots — the “write me a poem about my quarterly report” kind of AI — should be bundled and forced on everyone. But the genuinely useful automation, the task-specific agents that run in the background and do real work? That gets locked behind a paywall, metered like electricity.
Why? Because Microsoft knows which one is actually valuable. They just don’t want to give it to you for free.
Think about what that means. The AI you’re being sold as transformative is the AI Microsoft knows is disposable. The AI that could actually change your workflow is the AI they’re afraid to include, because once you’ve tasted real automation, you’ll never go back — and they want to charge you for every bite.
Every AI demo you’ve watched in the last two years was a magic show. The real trick was making you pay for the rabbit before you noticed the hat was empty.
And the layoffs? Don’t get me started. Microsoft is cutting divisions while pushing a product that 99% of its users ignore weekly. Imagine running a restaurant, firing your chefs, raising menu prices, and serving a dish that 99 out of 100 customers send back. That’s not a business strategy. That’s a slow-motion car crash with a great PR team.
The broader pattern here matters for all of us. Enterprise software companies are wrapping basic chatbot functionality in billion-dollar marketing, slapping “AI” on the invoice, and hoping nobody notices the emperor has no clothes. Copilot isn’t an outlier — it’s the blueprint.
So here’s what you should take away: the next time a tech giant tells you their AI will transform your workday, ask one question. Is the actually useful version included in the price, or is it the expensive add-on they mention in the fine print?
If the real product costs extra, the free version isn’t a teaser. It’s a warning.
FAQ
Q: Is the 4.5% adoption statistic even reliable?
A: The number surfaced from internal reporting and hasn't been officially confirmed by Microsoft, which is telling in itself. But even if the real number is double or triple that, we're still talking about a product with single-digit adoption after three years and billions in investment. The exact percentage almost doesn't matter — the gap between hype and usage is enormous either way.
Q: So should companies stop paying for Copilot?
A: Audit your actual usage first. If your team falls into the 99% who don't touch it weekly, you're burning money. The practical move: push Microsoft for access to Co-Work or equivalent agent features, and negotiate pricing based on real adoption metrics, not seat licenses.
Q: Isn't this just early-adopter friction? Won't adoption grow over time?
A: No. Three years isn't early adoption — it's a verdict. Products that deliver real value see hockey-stick growth within months. When usage stays flat for years while the vendor raises prices, that's not friction. That's rejection. Microsoft isn't waiting for users to come around; users have already decided.