AI for Product Managers is not about knowing how to use ChatGPT but it’s about leveraging AI for strategic decision-making for your product.
AI is everywhere. So are AI training programs: it seems like every university, company, and parish now offers an AI program. It’s a hot topic, I get it.
But knowing AI is not only about understanding how it works and how to use the tools. It is also knowing how to leverage it in your specific job. I find this last point much more important. Learning how to use ChatGPT or Claude or n8n agents… is useful, but this is just the foundation.
The hard part is understanding how to leverage all these tools for your job. How to make sense of it, filter out the noise, and use the tools to accelerate, expand, or simplify your strategic (and tactical) decision-making?
Knowing that Claude Cowork can sift through all my files and highlight customer opportunities and pain-points is great – but how do I use this capability in my own job? In my own workflow? With my own team?
Using synthetic users to conduct customer research is great, but how do I know if the insights AI generated are useful or simply algorithm-averaged noise?
It’s the distinction between learning how to bake pizza and actually baking pizzas when I have guests over for a party. Or the distinction between learning how to code and actually building a useful app that people use. In one case, we simply learn how to use the tools. In the other, we learn how to master the tool and leverage it to use in our own job and do something useful with it.
Take synthetic users as an example. The tool is straightforward: give an AI a persona, ask it customer research questions, get back a response. But that output is only as good as the judgment you apply to it. Knowing whether an AI-generated insight reflects a real unmet need or simply mirrors the data you fed in requires you to know your users, your market, and your product deeply enough to spot the difference. The AI accelerates the research process. The PM still has to decide what to trust.
Here are the main points:
(1) Tool fluency is becoming table stakes
(2) What matters is workflow and strategic-decision making
(3) Inserting human judgment at the right moment is the actual skill – and this is where product teams need to focus their attention.
Working with teams and their leadership in adopting AI I find that everybody knows how to use ChatGPT/Claude/Gemini/[Your-favorite-AI-tool]. In fact, most of us use these tools every day for help in writing emails, creating powerpoint presentations, or writing proposals. Yes, some of us are more advanced users than others (Using Projects for context preservation, using Skills for repeated tasks, using Agents for automation) but knowing how to use the tools is becoming tablestakes. Like knowing how to use an electric drill.
What will distinguish product teams in the future is not how well they know the tools. It’s the use case. It’s the workflow. It’s the strategic decision-making that cannot be delegated to a machine, and knowing when human judgment needs to be inserted in the process.
Building an AI Agent is easy (open n8n or Make and build it). But understanding what to use the Agent for, and when not to use it, requires a higher level of thinking. Deciding when to insert human-judgement in the AI workflow, and when to automate it, may represent the difference between success and a compliance gap.
The uncomfortable part: inserting judgment into an AI-driven workflow requires slowing down. And the whole promise of AI is that it speeds things up. So the discipline of knowing when to pause, when to question the output, and when to override the recommendation runs against the grain of how most teams are adopting these tools. That is where most of the risk lives.
In practice, this looks like a PM who uses AI to synthesize customer interview transcripts, then reads those summaries critically rather than shipping the first opportunity it surfaces. Or a team that automates their competitive monitoring with an agent, but still holds a weekly conversation about what the signals actually mean for their strategy. The AI does the work it’s good at. The PM does the work that requires judgment.
We saw a similar pattern in the past when everyone (and I mean everyone) tried to learn Product Management techniques (“let me show you how to use RICE for your product backlog prioritization”, or “here is how you can leverage Jira to write user stories”). Learning how to use the tools led to the feature factory, until we took a step back and inserted judgment into the product practices, so that we started focusing on what delivered value rather than what we could build next.
So, let’s leverage AI to accelerate, expand, or simplify the PM job while inserting Product Thinking into the process to ensure we use the tools to build what makes sense and what drives value – not AI for the sake of AI.
AI TRAINING FOR PRODUCT TEAMS
I help product teams learn how to leverage AI to accelerate and expand Product Thinking, to identify the right use cases for AI, and where to insert strategic decision-making in agentic workflows.
Learn more: https://www.5dvision.com/training/ai-for-product-managers/
In-person training for product teams at your location or private online cohort. Train your entire team and create team workflows that leverage AI.