HomeLearnCoursesWebinars
PM × AIReading the latest with you
LearnCoursesServicesWebinars
עב
PM × AI

Taught by Ofer Regev — a PM who ships with AI.

Join WhatsApp group
  • LinkedIn
  • YouTube
  • GitHub
  • theaipmhub@gmail.com
Work with me
  • Live courses
  • Team workshops
  • Consulting
Free content
  • Learn: the Map
  • Articles
  • Videos
  • Skills
  • Webinars
  • WhatsApp group
© 2026 The AI PM Hub · Built for product managers
AboutContactTermsPrivacyAccessibility
  • The Map
  • Articles
  • Videos
  • Skills
  • I'm new to AI. Where do I start?
  • I have a product idea. How do I get to an MVP?
  • How do I work with Claude and Claude Code?
  • I'm job hunting or preparing for interviews.
  • Rules, Commands or Skills? A PM's Guide to Picking the Right One
  • Stop Blasting the Same CV: An 8-Step Job Search System You Can Run With Claude
  • You Built a Prototype. Here's What Stands Between It and a Real App
AI Basics
Claude
Claude Co-work

So You Got Claude at Work. Now What?

4 min readPublished June 30, 2026By Ofer Regev

You got access. Congratulations — welcome to the club everyone keeps telling you to join.

Your manager wants you using it. Your manager's manager wants a slide about it. Your teammates are swapping prompts in Slack. Even your mom asked if you've tried "the Claude thing" yet.

And like with any tool, "use it because everyone says so" is the worst possible reason to start. Adoption for adoption's sake produces exactly what you'd expect: a half-used tool, a guilty conscience, and a slide nobody believes.

So how do you actually start — in a way that sticks?

The one-question test

Here's the filter I use, and it's simpler than most "AI adoption frameworks" you'll find:

Pick something from your actual week. Then ask yourself one question:

Could I hand this off to someone else and trust them to do it?

If yes — keep going. If no, it's not a starting task. Some things genuinely require your judgment, your relationships, your context. Don't force those onto an AI tool just to "use it." That's how trust gets burned in week one.

Now ask a second question:

What happens if this goes wrong?

Low stakes? Good. That's your starting task.

High stakes — a customer-facing email, a number that goes into a board deck, a decision that's hard to walk back? Park it. Not because the tool can't help eventually, but because your first rep with it shouldn't double as a crisis.

What this looks like in practice

For a PM, that first task is rarely glamorous. It's things like:

  • Turning a messy meeting transcript into clean, organized notes
  • Drafting the first pass of a status update from your raw bullet points
  • Summarizing a long thread before a stakeholder meeting
  • Restructuring a half-finished doc so it actually flows

None of these are "AI strategy." They're chores. Repetitive, low-risk, a little tedious — exactly the kind of work that's been sitting on your plate because it's never urgent enough to prioritize and never interesting enough to enjoy.

That's the point. You're not looking for the most impressive use case. You're looking for the lowest-friction one — something where being wrong costs you five minutes, not your week.

Don't skip ahead

Here's where most people go wrong: they read a case study about some team automating seven workflows into one beautiful dashboard, and they try to start there.

Don't. That dashboard took someone twenty failed attempts to get right, and you didn't see any of them.

Start with one task. Get it off your plate. Notice what worked and what didn't. Then — and only then — move to the next one and shift up a gear. Adoption isn't a launch, it's a habit you build one rep at a time.

You're still the one driving

It's tempting to think the tool does the thinking now. It doesn't — not the thinking that matters.

The capabilities are genuinely wild. But knowing what's worth doing is still entirely on you. Claude doesn't know which stakeholder is touchy about scope creep, which metric your VP actually cares about, or why that "simple" feature request isn't simple at all. You do.

So treat it less like a replacement and more like the most capable, fastest, least-territorial direct report you've ever had — one who still needs you to point it at the right problem.

Start small. Hand off the chore. Build the habit. The rest follows.