Prompt Analytics

Sample data — design review

How well people are writing prompts — and how much we improve them

Prompts validated62,140

+8.1%Validate Prompt runs

Prompts tested48,320

+9.4%Side-by-side comparisons

Prompts improved29,870

+12.6%Accepted a rewrite

Avg quality score72 / 100

+3Up from 69 last month

Avg improvement+28 pts

+246 before → 74 after

Weak prompts flagged31%

−4.2%Scored below 50

Prompt quality distribution

Validator score of every prompt written in the last 30 days

2,140

0–20

6,980

21–40

14,210

41–60

18,460

61–80

6,530

81–100

Most learners sit in the 61–80 band. Moving the 41–60 group up one band is the single biggest quality win available.

What the validator flags most often

Average validator score before and after each AI feature is used

FeatureRunsScore before → afterLift

Validate Prompt58,940

+22

Improve Prompt49,760

+35

Test Prompt38,180

+22

Practice24,420

+22

Challenge13,020

+26