The other day I was asked a question that sounds sensible on the surface.
“If GitHub Copilot is so good, why not just build it directly into Microsoft 365 Copilot for everyone?”
At first glance it feels like the logical next step. One Copilot. One interface. One AI assistant for everything.
The more I think about it though, the more I believe that would be a mistake.
The reason comes down to something that AI vendors often overlook. Different people work in very different ways.
A software developer spends their day creating, testing and refining code. A finance manager lives in spreadsheets. A salesperson works across emails, meetings and CRM records. An executive spends much of their time making decisions based on information coming from multiple sources.
Those aren’t variations of the same job. They’re fundamentally different disciplines.
The Context Is More Important Than The AI
What makes GitHub Copilot valuable isn’t simply that it can generate code.
It’s that it understands the environment a developer works within.
It sits inside development tools. It understands repositories. It helps navigate large codebases. It assists with testing, debugging and software creation workflows. The value comes from the context surrounding the AI.
Microsoft 365 Copilot has a completely different context.
When I ask Copilot in Outlook to draft a response, or use Copilot in Teams to summarise a meeting, the AI is working with conversations, documents, calendars and organisational knowledge. It is focused on helping people communicate, collaborate and make decisions.
Trying to merge these two worlds into a single experience risks making both less useful.
A productivity worker doesn’t need a sophisticated coding assistant appearing in every interaction. In the same way, a developer probably doesn’t want meeting summaries and document drafting cluttering up their coding environment.
Specialisation Usually Wins
We’ve seen this pattern before.
Nobody expects Excel to become Visual Studio.
Nobody asks Word to become a network troubleshooting platform.
Microsoft succeeds because its tools are specialised while still working together.
I think AI will follow the same path.
GitHub Copilot should continue evolving as the specialist assistant for software creation. Microsoft 365 Copilot should remain focused on helping knowledge workers get through meetings, emails, documents and decision-making processes faster.
The mistake many organisations make is assuming every AI capability needs to be available to every employee.
In reality, most people only need the capabilities that relate directly to their role.
Adding more functionality doesn’t automatically create more value. Sometimes it just creates more complexity.
The Real Opportunity Is Connection
Where I do see value is in integration.
Imagine a project manager using Microsoft 365 Copilot to prepare for a software review meeting. The assistant could pull summaries of development activity from GitHub.
A developer could use GitHub Copilot to understand changes in a repository while still accessing relevant business context from Microsoft 365.
That’s very different from forcing both experiences into a single product.
The future isn’t one giant AI assistant trying to do everything.
It’s a collection of specialised agents and copilots that share information when necessary while remaining focused on their primary role.
We’re already starting to see this approach emerge across Microsoft’s ecosystem.
Keep The Tool Sharp
One lesson I’ve learned over many years working with Microsoft technologies is that the best tools are usually the ones with a clear purpose.
A hammer becomes less useful once you start turning it into a screwdriver, a saw and a wrench at the same time.
The same principle applies to AI.
GitHub Copilot should remain the coding expert. Microsoft 365 Copilot should remain the productivity expert.
The goal shouldn’t be to squeeze every AI capability into a single interface. The goal should be to put the right assistant in front of the right person at the right time.
That’s where the real productivity gains will come from.