Claude is becoming an operating system, not just a tool

Claude is becoming an operating system, not just a tool.

Most companies still buy AI the way they buy any other app. Someone picks a chat app, a few teams get licenses, and it sits next to everything else they use.

This week Anthropic made that view hard to hold. In a few days it added spreadsheets, whiteboards, and dashboards that live inside Claude. It let apps like Notion, Box, and Confluence open and be edited right inside a Claude chat. And it opened a beta that lets Claude join Google Meet calls on Team and Enterprise plans and build the notes and documents while the meeting is still going.

When the Sheets, Whiteboards, and connectors news landed on Tuesday, I posted that Claude is quite literally eating the world. Taken together, Claude isn't just a tool anymore. It's a place where the work happens, the way Microsoft Office became the place work happened for most companies.

That changes how you should buy it. If you treat Claude or ChatGPT as one more app, you make small decisions team by team and end up with a patchwork. If you treat it as the platform your company will run on, you ask different questions up front: which one fits the systems you already use, who controls access and data, and which apps it may replace.

What you can do this week: before your next AI purchase or renewal, ask whether you're choosing an app or a platform. If it's a platform, bring IT, finance, and a department head into the decision together.

What Anthropic launched this week

Your next hire may never mention AI.

My son is applying to colleges. He uses AI throughout and never lists it as a skill. It's just how he works.

That's one person, so I'm treating it as a hypothesis, not a trend. But it changed how I think about who's coming into the workforce next.

Most of how companies handle AI today assumes it's a separate skill. We run training to teach it. Candidates list it on a resume to stand out. People in my line of work spend a lot of time explaining what it can do. The usual worry is that people who know AI will pull ahead of people who don't.

The next group may not fit that picture. If you've drafted, researched, and studied with AI from the start, it isn't a skill you picked up. It's part of how you do anything, so you'd never think to write it down.

If that's right, three things change.

Hiring gets harder to read. A candidate who never mentions AI may use it more than one who lists it first. The resume won't tell you, so ask people to walk you through how they did a recent piece of work.

Training needs a different starting point. A course on how to use AI will feel basic to these hires. What they'll need from you is the part only your company can teach: what good work looks like here, and when not to trust the output.

The baseline moves for everyone else. For experienced professionals, using AI stops being an edge. What still sets you apart is what you know about the work.

People follow their own leaders on AI.

Most AI rollouts start at the top. An executive announces it, IT turns it on, and training goes out to everyone. That gets people access. It doesn't show anyone what AI looks like in their own work.

The companies we see having the most success aren't the ones talking about AI the most from the executive suite. They're the ones where leaders use it themselves. They try it on their own work, experiment, and tell their teams what helped and what didn't.

That changes the message. A note from headquarters says the company is doing AI. A leader who used it on this week's work says this is how we work now. People follow what their leaders do far more than what gets announced.

What you can do this week: if you lead a team, use AI on one real piece of your own work. Then tell your team what you used it for and how it went, including what didn't work.

New on the site

The interface is moving into the conversation

For about twenty years, using software meant learning where the menus and buttons were. This post traces how that's changing: documents, spreadsheets, whiteboards, the browser, and even meetings now open inside one AI conversation, so more work starts with a sentence instead of a click.

It also covers what that means for leaders: how to train people, what to ask when you buy software, how your website holds up when a customer arrives through a chat, and which guardrails to set.

Read the post