AI Product
Discovery
We don't bolt on AI.
We find problems that need it.
Research that finds recurring problems and tests whether AI is actually a better solution.
THE QUESTION
What problem should Newon's
next operated AI product solve?
- STATUS
- RESEARCH
- STAGE
- RESEARCH
- CATEGORY
- AI
CORE PRINCIPLE
WE DON'T START WITH AI.
WE START WITH THE PROBLEM.
DISCOVERY PIPELINE
- 01PROBLEM
Find recurring user problems
- 02SIGNAL
Confirm demand & behavior signals
- 03AI FIT
Judge if AI beats existing approaches
- 04PROTOTYPE
Minimum viable experiment
- 05VALIDATE
Validate with real users
- 06PRODUCT
Graduate to a Newon product
EXPERIMENT BOARD
RESEARCH CONCEPTRESEARCH CONCEPT
- PROBLEM
- Prioritizing App Store review themes
- TARGET USER
- Product teams / solo makers
- FREQUENCY
- HIGH
- AI FIT
- HIGH
- BUSINESS POTENTIAL
- MED
- STATUS
- RESEARCH
RESEARCH CONCEPT
- PROBLEM
- Drafting / routing support inquiries
- TARGET USER
- Small ops teams
- FREQUENCY
- MED
- AI FIT
- MED
- BUSINESS POTENTIAL
- MED
- STATUS
- RESEARCH
RESEARCH CONCEPT
- PROBLEM
- Turning lab notes into product hypotheses
- TARGET USER
- Newon internal
- FREQUENCY
- LOW
- AI FIT
- MED
- BUSINESS POTENTIAL
- LOW
- STATUS
- RESEARCH
IDEA TEST
PROTOTYPE HEURISTIC — not a live AI APILAB NOTE
What we found
Vague problems always score high on AI fit. Frequency and alternatives are the real filters.
Next to verify
Keep or cut one non–review-intelligence candidate on the board.