Claude Remembers Who You Are. And Now It’s Inside Excel and PowerPoint

You know that frustrating feeling of having to explain everything from scratch every time you open a new chat with an AI assistant?

"I work in marketing." "I handle events." "Our tone of voice is institutional but warm." "Remember that our main client is in the pharma sector."

Since March 2026, with Anthropic's Claude, that feeling is officially a thing of the past.

The memory that changes everything

Anthropic has rolled out persistent memory for all Claude users: the model now retains context, preferences and relevant information from one conversation to the next.

In practice: the first time, you tell Claude who you are, how you work, what your priorities are. From the second time on, it already knows.

It sounds like a detail. It isn't.

Persistent memory solves one of the most underrated practical problems in enterprise AI adoption: the cognitive cost of context. Every time a professional opens a new session with an AI assistant and has to start over from the beginning, they lose time and focus. After a few days, they stop using it.

With memory, AI starts to behave like a real colleague: someone who already knows how you work, what you have already done and what you are trying to achieve.

Inside Excel and PowerPoint: the paradigm shift

But the news doesn't end there. In the same week it released memory, Anthropic rolled out Claude as a native add-in in Microsoft PowerPoint and Microsoft Excel.

Translated: you open your spreadsheet, and Claude is already there. You open your presentation, and Claude is already there. No need to open another tab, copy and paste, or step out of your flow. AI is inside the tool you already use every day.

This isn't a product update. It's a paradigm shift in how AI enters, or fails to enter, people's working habits. Until yesterday the main obstacle to AI adoption was behavioral: people have to remember to open the AI tool, learn to use it, integrate it into their workflow. The resistance wasn't technological, it was about habit. With Claude inside Excel and PowerPoint, that resistance disappears. The AI tool is already where you are.

What it means for marketing and communication teams

For anyone working in communication, marketing or project management, the implications are concrete:

• Content creation: building a presentation for a client? Claude helps you structure it, refine the copy, adapt the tone, without ever leaving PowerPoint.
• Data analysis: got an Excel sheet with the results of a campaign? Claude can interpret them, suggest visualizations, spot anomalies, right in the cell next door.
• Brand consistency: thanks to memory, Claude knows your tone of voice, your clients, your guidelines. Every piece of text it generates is already calibrated to your brand, without having to repeat the same instructions every time.
• Decision speed: less time switching between tools, less context lost, more time for the work that really matters.

The point no one makes

There is, however, a reflection worth making.

When AI enters your tools and remembers who you are, the stakes change. The data you share with the assistant, information about clients, strategies, budgets, confidential content, is now part of a system with persistent memory.

The question every company should be asking isn't "should we use Claude?" but "what information do we want Claude to remember?" and "in which environment do we run it?".

AI governance, meaning defining what AI can know, which data it can operate on and with what human oversight, isn't a technical topic. It's a strategic choice.

At Media Engineering we work on this every day: building AI systems that bring real value to business processes, with the right architecture of control and oversight. Because a useful AI and a governed AI aren't different things, they are the same thing.