Rosana Inacio – PM Insights

A few weeks ago, my company approved Claude for use across our teams.

I was skeptical at first. I had been using another AI tool for over a year. It worked. It was fast. I knew how to use it. Why change?

But I was curious, so I decided to explore.

What I found surprised me.

My Process

Here’s how I work. A product manager gives me a scope. I take that scope and feed it into an AI tool along with everything I know about the project, team size, sprint velocity, sprint length, timeline, constraints, dependencies, everything.

Then I ask for a complete project plan. I need a work breakdown structure, Jira tickets, an MS Project timeline, risk assessment, stakeholder management strategy, all of it.

The AI produces a draft. I review it with my team. They give feedback. I adjust. Then it goes into the official systems.

I do not create these from scratch myself. But I also do not take AI output and run with it. I validate everything with the people who will actually do the work.

What My Previous Tool Did

I provided the scope and all the context. It was executed immediately. No questions. Straight to the output.

It produced everything I asked for: project summary, WBS, Jira tickets, MS Project plan, risk register, stakeholder plan. But it gave me one piece at a time. Separate outputs that I had to compile, organize, and structure into something coherent.

Then came the rework. Review. Gap analysis. Refinement. Adjustments. More questions. More iterations.

It was still faster than doing everything manually, but it required significant effort before I had something solid enough to bring to the team.

What Claude Did

I gave Claude the exact same scope, context, and request.

But before producing anything, it asked questions.

Is the team experienced with this type of work? Is the sprint length fixed or flexible? Have you accounted for dependencies? What is your risk tolerance?

These were not obstacles. They were clarifications I should have thought through anyway. They highlighted gaps I might have missed.

Then it produced everything, not one item at a time, but one organized Excel file with multiple tabs. Project summary. WBS. Jira tickets with detailed descriptions. MS Project plan with predecessors and dependencies already built in. Risk assessment. Stakeholder strategy.

Everything is organized in one place.

I honestly did not expect it to be this complete on the first pass.

Why This Matters

Both tools helped me move faster than doing everything manually. That’s table stakes now.

The real difference was the rework. With my previous tool, I spent multiple rounds reviewing, restructuring, refining, and validating before I had something usable. With Claude, I started much closer to a solid final product.

That’s not just speed. That’s accuracy the first time.

But Here’s the Important Part

The tool I’m using today may not be the best one six months from now.

AI is moving fast. What leads today may be challenged tomorrow. New capabilities are emerging constantly.

I’m not saying stick with Claude forever. I’m saying stay curious.

Don’t assume the tool you already know is still the best option. Keep exploring. Try new things. Test them against real work. See what genuinely improves the way you deliver projects.

And honestly, that makes me excited for what comes next.

I’d also be curious to hear what AI tools other PMs are experimenting with right now.

Rosana Inacio — PM Insights