What Working Niche Scoring Looks Like (AI SaaS)

An AI SaaS founder considers five niches.

The Candidates and Scores

• AI for code review at startups: R6 F7 S5 D4 B7 = 29
• AI for legal contract review: R7 F7 S9 D6 B5 = 34
• AI for customer support triage: R8 F8 S9 D5 B6 = 36
• AI for sales call analysis: R8 F7 S9 D5 B7 = 36
• AI for medical scribing: R5 F8 S8 D7 B3 = 31

What the Numbers Reveal

Two niches tied at 36.

Customer support triage and sales call analysis. Both reachable, both painful, both with established existing spend.

The technically interesting niches (code review, medical scribing) scored lower. Code review on reachability, medical on builder-fit because the founder had no medical domain access.

Tiebreaker

When top two are within three points, pick the one with higher existing spend. Both at 9 here, so the founder picked sales call analysis on builder-fit (8 vs 6 — they had a sales background).

What This Caught

The exciting code review niche would have failed on reachability. The medical niche would have failed on builder-fit.

Neither would have surfaced from gut. Both surfaced from honest scoring.