I have a problem with rabbit holes. I can start investigating a small feature request at 9am and by lunch I have drafted a product brief for an entirely different initiative. I know this about myself. I wrote it into my own operating instructions so that my tools can push back on me.
AI makes this problem significantly worse.
Before AI, scope creep had natural friction. You wanted to explore a tangent? You had to open a spreadsheet, pull data manually, draft a doc from scratch. That friction was annoying but it was also protective. By the time you finished setting up the exploration, you often realized the tangent was not worth it. The cost of exploring killed bad ideas early.
Now I can ask an AI to draft a product brief in 30 seconds. I can generate competitive analysis, build a requirements table, sketch out a roadmap. All within minutes. The friction is gone. And with it, the natural filter that stopped me from turning every interesting thought into a project.
Last week I sat in a strategy meeting. One person mentioned a tangential idea. It was a good idea. Within an hour of the meeting ending, I had used AI to produce a detailed exploration of the concept, complete with market sizing and a phased rollout plan. I was ready to pitch it to leadership.
Then I stopped and asked myself: is this actually more important than the three things I already committed to delivering this quarter? The answer was no. But the AI made it feel ready. It made a 15-minute tangent look like a fully formed initiative. That is dangerous for a Product Manager.
Scope creep used to look like “we should also add this feature.” Now it looks like “here is a complete brief with success metrics and a timeline for a thing we mentioned once in a meeting.” The production quality of the creep has increased. The creep itself has not become more valuable.
The fix is not to stop using AI. It is to build constraints that replace the friction AI removed. I use a few:
Before generating anything with AI, I write one sentence by hand: “This is worth exploring because ___.” If I cannot finish that sentence with something tied to a current goal, I stop.
I keep a visible list of my current commitments. Not in a project tool buried behind three clicks. On my screen, in my face, all day. When AI produces something exciting and new, the list reminds me what I already owe.
I give myself a 24-hour rule for AI-generated ideas. Draft whatever you want. Come back tomorrow. If it still seems important after sleeping on it, maybe it is. Most of the time, it is not.
The uncomfortable truth is that AI amplifies whatever tendencies you already have. If you are disciplined about scope, AI helps you work faster within that scope. If you are prone to scope creep, AI gives you a production studio for your worst habit.
Know which one you are. Build the guardrails accordingly.
Leave a Reply