Microsoft announced their new Microsoft 365 E7 last week. It costs $99 per user per month. 65% more than E5. And it includes Copilot, identity management, and an entirely new product called Agent 365 for managing a company’s AI agents.
It’s an interesting move. Not because the bundle is revolutionary, but because it reveals something about where the market actually is right now.
Here’s the number that really matters: Microsoft CEO Satya Nadella said in January that they have 15 million paying Copilot users. Sounds like a lot. But Microsoft has roughly 450 million commercial users. That’s 3%.
Three percent. And now they’re launching an even more expensive product that assumes companies aren’t just using AI as a chatbot, but have AI agents they need to manage and monitor.
It’s like selling a high-end sound system to people who still have the radio turned off.
Microsoft’s strategy makes perfect sense from a business perspective. They’ve invested over $100 billion in data centres in the past year. They need to show returns. E7 is an elegant way to package everything into one invoice instead of six separate add-ons.
But for companies, it’s a mirror more than an offer. How many have actually implemented the AI tools they’re already paying for? How many have a strategy for how AI agents should function in their organisation? And who has thought about the security model when it’s no longer just humans making decisions in the systems?
That last point is worth sitting with. Forbes published an analysis last week on how agentic AI is challenging fundamental assumptions in enterprise security. The core is simple: most security systems are built on the assumption that humans make the decisions. When AI agents start acting autonomously, that assumption breaks down.
This isn’t a future scenario. It’s now. Microsoft is building Agent 365 because they can see companies already experimenting with agents. But most are doing it without governance, without a security model, without a plan for what happens when something goes wrong.
The gap between what’s available and what’s actually being used keeps growing. And it’s not a technology problem. It’s a leadership problem.
The companies that do best with AI aren’t the ones buying the most expensive subscription. They’re the ones that start by understanding what they actually want to achieve, and then choose the tools. Sometimes that’s Microsoft. Sometimes it’s something entirely different. But it always starts with a strategy.
Microsoft’s E7 is a signal that AI agents are going mainstream. That’s good. But the signal should be read as a question to your own organisation: Are you ready for it? Not technically. Organisationally.
Because the cost of waiting isn’t the $99 per user. It’s that your competitors figure it out first.
