thoughts
How AI changed my day as a Product Owner
When someone asks which AI tools I use as a Product Owner, they usually expect a list. A few names, maybe a recommendation, done. I understand the impulse, but the honest answer is less convenient: the list is the least interesting part. What matters is not which tools I have open. It is how they work together.
Because there is one thing I now believe with some confidence: AI does not replace Product Owners. It replaces Product Owners who don’t change how they work. That sounds harsh, and it is meant less dramatically than it reads. The real job of this profession was never writing tickets, running meetings, or maintaining documentation — those were always just tools. The real job is understanding problems, seeing how things connect, setting priorities, joining business and technology, and, in the end, making good decisions. That is exactly what AI does not do today. What it is good at is the repetitive process around it. So it does not change the core of my job. It changes my day: how I work, document, build up knowledge, communicate, develop, and prepare decisions.
And out of that comes the thing a Product Owner may value most — time. More time for discovery, for stakeholders, for strategy, for the hard decisions no one can delegate. So I don’t see AI as a replacement but as an amplifier: a bad Product Owner does not become good through it, and a good one can do considerably more. What follows is not a test and not a comparison. In this piece and the three that follow, I want to describe honestly what a whole product process looks like once you connect it, piece by piece — and why the seams between the tools matter more than the tools themselves. I’ll start with the one I open most often.
ChatGPT: my daily sparring partner
ChatGPT is one of my daily drivers. I use it practically every day, and not for the big, spectacular things — for the many small ones. To me it is a universal tool, the place where hundreds of tiny tasks land that together eat a surprisingly large part of the day.
A Product Owner makes hundreds of small decisions a day. How do I phrase this email so it doesn’t land the wrong way? Is this requirement actually unambiguous? How do I explain this technical dependency so that marketing understands it too? None of these questions is large on its own. In sum, they consume the day. And it is precisely there — in that layer of small, verbal, recurring tasks — that AI gives me the most.
In practice that means: I have emails rewritten, or I draft them together with ChatGPT from the start. I have requirements sharpened until they are genuinely unambiguous. I make texts clearer. I research things, ask about technical dependencies, and have it explain how a particular tool works. For recurring tasks I sometimes use my own GPTs, which I set up once with the right context and then reuse. And I reach for it surprisingly often when I set up new software — configuring Obsidian, say, or connecting two systems that would otherwise cost me half an hour of reading documentation.
It helps that ChatGPT has grown up over the years. Features like Projects let you bundle context, files and your own instructions instead of starting cold in every conversation. Deep Research handles multi-step investigations that used to be half an afternoon. Connectors let the model reach into the systems where the actual information lives. But I want to be honest about where the line runs: ChatGPT is not a developer to me, and not a product manager. It is an intelligent assistant. It makes no product decision, runs no discovery, replaces no product understanding. What it mostly does is help me get to the goal faster — and, by doing so, leave my head free for the things it can’t do.
That is the everyday layer — the many small tasks. The real jump starts somewhere else: when fleeting conversations and scattered notes turn into knowledge that lasts. That is the next part.