Choosing a problem for AI
Judge whether a problem suits AI: frequent and valuable, tolerant of some mistakes or easy to check, backed by data, and better solved by AI than by a simpler fix.
Beginner 3 · Data and toolsabout 2 hours5 sections
Quick answer
A good AI problem is frequent and valuable, tolerates some mistakes or is easy to check, has data behind it, and is not solved more simply by a rule or form.
Why it matters
Many failed AI features were the wrong problem, not the wrong model. A PM who can say "not this one" with reasons, and pick the problem where AI clearly wins, saves months and builds trust with leadership.
Sign up free to read the whole lesson
The rest of this lesson: 5 parts, video lessons, the best course, practice and a quick check. It is free, takes a few seconds and needs no card.
Accounts are for learners aged 18 and over. Under 18? The quick answer and the first part of every lesson stay open: browse lessons by topic.
Lessons on AI Product Manager0 of 19 done
- 1What AI, machine learning and a language model are
- 2How AI products work
- 3Tokens and the context window
- 4Prompting, RAG or fine-tuning
- 5Choosing a problem for AI
- 6Feasibility and build versus buy
- 7UX for uncertain output
- 8Designing human in the loop
- 9Test sets and evals
- 10The KPI that matters
- 11Data strategy for AI products
- 12Bias and fairness
- 13Use AI safely at work
- 14AI risk frameworks
- 15Writing an AI policy
- 16Cost control
- 17Pricing AI features
- 18Change management and training
- 19AI PM interviews
Related
On the roadmap: AI Product Manager
ProductBeginner 3: Data and toolsVideo lessonsKPIHuman in the loopDiscoveryEvals