Copilot is shifting from experimental to essential.
1 Like
I work at a company where M365 Copilot is the only-approved AI platform we can use. It has definitely come a long way, and I find myself missing my personal Claude app less and less. What needs improving IMO is how they are handling third-party integrations and custom Copilot chatbots. They are essentially taking the shaky foundations of PowerApps and the Power Platform and reworking it into their Copilot Studio product. My team is struggling to build anything of use within this framework due to a buggy interface and confusing implementation of MCP support.
Nothing is close to the ChatGPT app, the co-pilot app doesn’t touch it and Gemini doesn’t even have one.
Having an app that read things like the terminal without copying and pasting is great.
We have access to many of the AI service provider tools. I mainly use OpenAI Enterprise at work, and have a personal Gemini AI account (the reasonable $20 one, not the crazy expensive one). We’ve been testing Copilot Chat and Copilot M365 for a while. It was horrible at first, but they have caught up well. Their secret sauce, of course, is being integrated deeply as a first party into tools we use every day. The ability to have it review all my emails and Teams messages, along with my calendar items for the week, and provide me a review is powerful. On the other side, having it prep for the week ahead. That said, I don’t think the AI ecosystem is going to be “won” - I expect to use multiple providers (similar to our multi-cloud approach to use different CSPs) as each will provide expertise in different areas (coding, productivity, etc).