AI PMs don't get better from watching. They get better from deciding.
LearnSignal is a training platform built entirely around one mechanic: put the PM inside the decision before revealing the answer. Every module is a real situation. Every lesson is earned through judgment, not absorbed through slides.
The knowledge gap is widening faster than anyone is closing it.
AI capabilities are compounding. The PM role has been rewritten. Yet most training built for AI PMs today is a patchwork — frameworks borrowed from general product management, lectures filmed before the current generation of models existed, and cohorts that move at the speed of a calendar rather than the speed of the field.
The PM in the meeting room knows how to ship. They do not know how to evaluate a model. They do not know when fine-tuning is the right call or a distraction. They are winging the conversations that shape what gets built.
This is not a personal failing. It is a structural one. The tools for developing this kind of judgment don't exist yet. LearnSignal is being built to create them.
Judgment is built through decisions, not definitions.
The scenario comes before the framework.Every module begins with a real situation — a hallucination reported the night before a board demo, a disagreement with an ML lead over fine-tuning vs RAG, an eval design with no ground truth. You inhabit the role first. You make the call. Then — and only then — does the technique reveal itself. That inversion is deliberate. That inversion is everything.
Pressure is part of the curriculum.Competence built in calm doesn't transfer to chaos. LearnSignal scenarios are written to recreate the actual conditions of the job — stakeholder pressure, incomplete information, tight timelines. The discomfort is the point.
Comparison is how you calibrate.After every decision, you see how your call compares to what experienced practitioners actually did — and why. Not to judge, but to build the internal reference point that separates a PM who guesses from a PM who knows.
What LearnSignal is not.
This is not a video course. Passive consumption builds familiarity, not fluency.
It is not a cohort. Scheduling dependency is a tax on learning that professionals cannot afford to keep paying.
It is not general product management education with an AI chapter bolted on. LearnSignal is 100% AI PM-specific. Every scenario, every tool, every concept exists because an AI PM needs it — not because it rounds out a curriculum.
It is not theoretical. The interactive tools embedded in the platform — prompt labs, cost calculators, eval builders — are the actual instruments of the job. You learn them by using them.
How we're building it.
Scenarios first. Everything else follows.The scenario engine is the platform. Before community features, before personalisation, before AI-generated content — we are writing the first 20 scenarios by hand, with practitioners, until they are genuinely hard to get right. The learning experience earns everything else.
Interactive tools embedded, not bolted on.A Prompt Lab where you experiment with system prompts in the context of a real product decision. An Eval Builder that walks you through designing measurement frameworks for outputs that have no ground truth. A Cost Calculator that makes token economics tangible. These aren't resources linked in a sidebar — they are woven into the learning flow.
Signals, not noise.The field moves too fast for a monthly digest. The Week's Signal surfaces one piece of research each week — stripped of academic framing, translated into the decisions it actually changes. We write for the PM who is too busy to read everything and too serious to accept a summary that doesn't tell them what to do differently.
Built in public, by a practitioner.LearnSignal is being built b
Join the waitlist
Founding access is opening in small waves. Add your email—we'll notify you when the next cohort unlocks.