AI REALITY INDEX
← BACK TO INDEX

Scoring Methodology

Our AI scores every claim against six weighted axes — from hard revenue proof to step-by-step execution detail. Claims with inflated numbers, missing costs, or platform-seller conflicts are automatically penalized. No human bias. No editorial opinion. Just signal.

The 6 Axes

Specific Numbers

WEIGHT: 20%

Raw revenue exports, dashboard screenshots, or verifiable data links. No vague claims.

Time Window

WEIGHT: 10%

Must show 90+ days of sustained unit economics. One-off spikes are penalized.

Cost Disclosure

WEIGHT: 15%

Ad spend, tool costs, team overhead, and hidden fees must be visible and itemized.

Customer Proof

WEIGHT: 20%

Third-party confirmation from clients or users. Self-reported metrics are discounted.

Execution Detail

WEIGHT: 20%

Step-by-step breakdown of what was done, what failed, and what was iterated on.

Replicable Steps

WEIGHT: 15%

Can someone with $500 and a laptop replicate this in a weekend? Scored for accessibility.

Penalty Rules

Obvious Promotion-25 pts

Claim is anchored to selling a course, tool, or service.

Unverifiable Hype-20 pts

Big numbers with no supporting evidence or timeline.

No Evidence-30 pts

Claim provides zero verifiable data points.

Confidence Model

Confidence reflects evidence density and signal consistency. It is separate from the score itself.

HIGH

Dense evidence across multiple axes. Consistent signals.

MEDIUM

Partial evidence. Some axes lack independent verification.

LOW

Sparse or ambiguous evidence. Treat conclusions with caution.

Limitations

  • Scores are based on publicly available information only.
  • We do not contact claim authors for private verification (yet).
  • Scores can change as new evidence surfaces.
  • This index does not constitute financial or business advice.

Want the top 3 replicable plays every Friday?

We filter noise, score real claims, and send the highest-signal opportunities.