Anthropic launched Claude Meetings, Claude Sheets and a collaborative whiteboard on 10 November: note-taking, a spreadsheet, a meeting space. On the same day, OpenAI turned ChatGPT into a directory where verified "business agents" answer customers, and Adobe opened Photoshop and Premiere to outside agents.
Three moves, one direction: having sold access to a model, the AI labs want to become the software where your teams spend the day.
Until now, adopting AI in an organisation meant plugging a programming interface onto existing tools: a budget, a line item, an interchangeable supplier. What is opening up shifts the ground. The suite where your meetings, your spreadsheets and your documents live becomes the place where your data settles and your work habits form.
The signal from the field tempers the excitement. In France, 58% of large companies were already using an AI technology in 2025, but scaling up agents remains tentative. The suites are therefore arriving before most organisations have even learned to run an agent reliably.
The usual order is reversed: we are offered the full environment when many are still trying to scope a first use case.
For a French executive, product director or IT manager, the decision changes in nature. Choosing a model can be revised in one line of code; choosing the suite that hosts your meetings, your files and your exchanges is revised far less easily.
The skill that will matter next year fits in one sentence: knowing which environment you run your work on, and keeping the ability to leave it. That means testing these tools while knowing the price of the exit, not only the price of the subscription, and training teams on the tool as much as on the way out.
The battle for AI at the office will be won on the environment that millions of people open every morning by reflex, far more than on the most brilliant model. Whoever holds that door holds the data and the uses. It remains to be decided who keeps the key.
The Masteria editorial team
