Microsoft AI Product Management

A coach that teaches by asking — not by lecturing

The Lighthouse Coach is an AI Socratic coach built on Cognosco Learning's Microsoft AI Product Management course, in the AI at Work category. It does not play videos at you. It holds a conversation: it asks you a question, listens to your answer, names the misconception it hears, affirms what you got right, and asks the next question — the way a great coach works in the room.

An AI feature makes a different promise. Every software feature used to promise: give it this input, and it returns that output, every time. An AI feature promises an output that is usually right, sometimes confidently wrong, and never identical twice. That single difference rewrites the product manager's job: you can no longer specify behaviour completely in advance, so you must specify how behaviour will be judged; you can no longer test your way to certainty before launch, so you must design launch itself as a controlled exposure with an exit; and you can no longer price a feature as if each use cost the same, because each use consumes compute, review time, and user trust in amounts that vary.

Module 1 teaches you to choose the bet. Run the worthiness test on any proposed AI feature and defend a no; assess what a model can and cannot do on cases drawn from your own workflow; state the outcome the feature must move and the guardrails it must never break; judge data readiness for feeding and for judging; write the evaluation that serves as the specification; decide build, buy, or partner layer by layer on a Microsoft-centred stack, with exit costs counted; design the human's place in the loop; and close the proposal as a bet memo that names its own falsifier.

Module 2 teaches you to ship it honestly. Sequence the roadmap by uncertainty rather than enthusiasm; prototype with the cheapest test that can kill the idea; run data operations as product work; design screens that show confidence honestly and let users recover; launch in rings behind kill criteria agreed in advance, with rollback rehearsed; price variable cost against variable value from the unit up; govern the feature without freezing it; measure whether it earned its place in user outcomes, in money, and in harm; and sign the whole under a personal standard.

It works in two movements. First, Module 1 — Choosing the Bet — work through Concepts 1–9, finishing with The Decision Pack; then, Module 2 — Shipping the Bet, Proving the Value — work through Concepts 10–18, finishing with The Launch-and-Value Plan.

Both capstones produce real artifacts — The Decision Pack, a complete go/no-go decision pack for a proposed AI feature that a leadership team could act on; and The Launch-and-Value Plan, a launch-and-value plan for an approved feature, defended in writing — each run on the running case and again on your own live product decision.

One company, judged honestly, the whole way through. Tessellate Systems, a fictional firm of about 240 people, sells Flowdesk, workflow software for mid-size service firms. Its proposed drafting assistant — drafting the structured summary, the category, and the first reply from a messy free-text request, for a human to approve — is examined, tested, priced, and judged across the whole course, on one honest set of facts.

Scope and honesty are stated up front. This course teaches the product-management discipline for AI products, and only that — general product-management craft is assumed, and model building and prompt craft are other books' ground. Platform facts are dated, product judgment is not: every Microsoft platform statement carries the book's framing in place — as of this writing, check the current Microsoft documentation before you rely on it — and no decision method in the course depends on any vendor's current catalogue.

The course is self-paced and always on — roughly 30 minutes per concept, about 9 hours in full, taken in as many sittings as you like. You can type your answers, or talk out loud: the Voice Coach speaks its questions and listens to your spoken replies. Learners who already know the material can test out concept by concept.

This is the two hundred and first course in The Lighthouse Signature Series — the online counterpart to the one-day, in-person learning Cognosco Learning has delivered since 2002.