The AI Swimsuit Index For when the AI tide goes out.
Overview
AI Claim Finder
Company analysis
Insights
About

The Proof Gap: AI talk vs. disclosed proof

Share of companies that name a deployment in their filings, reach a quantified result (Proof Level 3), and a trackable result (Proof Level 4)

Names a deployment (filings) Quantified result · Level 3 Trackable result · Level 4

Disclosure from company filings, 10-K and 10-Q. Quantified impact and proof levels from earnings calls, 8-K releases and web research.

Talk vs. Proof

AI disclosure volume against disclosed, verified proof · bubble size = how little of its AI talk is substantiated (bigger = more talk, less proof). Select a sector:

Sources: financial filings, earnings calls, press releases and web search.

The Proof Funnel

AI impact disclosures by Proof Ladder level

Slide or play to watch the funnel evolve. Verified at the latest quarter; auto-extracted floor for history. Sources: filings, earnings calls, press and web search.

What kind of proof?

The kind of value AI delivers, and how hard the metric proving it is.

Select a company to see its AI Talk and AI Proof.

Explore what companies claim AI delivers

Build a question from the lenses below. Every figure is a deduplicated, sourced claim drawn from filings, earnings calls and 8-K releases. Pick a unit to see the spread of the numbers.

Insights

Essays and analysis behind the index — published on Substack.

The AI Swimsuit Index

How much large companies talk about AI vs. how much they can prove.

Why it exists

Every large company now says it's "powered by AI." Far fewer can point to a hard number that proves it pays. The index scores the distance between the two: how loudly a company talks about AI across its filings and earnings calls, against how much proof of real, disclosed value it can put on the table.

"Only when the tide goes out do you discover who's been swimming naked." The name nods to Warren Buffett: AI is the rising tide; this index is the gauge for when it goes out.

How it's scored

Talk - every AI mention in SEC filings (10-K/10-Q) and earnings-call transcripts, gated to exclude risk-factor boilerplate, themed and tagged, then normalised on a fixed scale so quarters stay comparable.

Proof - a 0–5 ladder for the observability and materiality of disclosed AI impact:

0
Silent: no material AI discussion
1
Aspiration: AI named as a priority, or a future target
2
Activity: a specific deployment, but only usage, adoption or spend to show for it
3
Result: a real, quantified value outcome ("$2bn saved")
4
Tracked metric: a defined economic metric (cost, revenue or margin), followable over time
5
Broken out: a standing, separately-reported line an outsider could audit each quarter

The strongest proof lives in earnings calls, 8-K releases and primary press, not 10-Ks. Coverage: S&P 500 (US, quarterly) and FTSE 100 (UK, semi-annual), each labelled by reporting period.

How to read it

Four tabs, each a different cut of the same data:

Overview - the index at a glance: the Talk × Proof scatter, the proof funnel, and the talk-vs-proof trend over time.

AI Claim Finder - the Insight Explorer: build a question from lenses (sector, value lever, mechanism, unit) to see the spread of disclosed AI figures and their sources.

Company analysis - search any company for its Talk and Proof scores, the disclosed evidence behind each, and a short AI read.

About the author

AM

Amin Mrini

Senior B2B operator · AI transformation · writer

A senior operator in B2B information services, working at the front line of large-scale AI transformation across events, subscriptions and media. Writes The AI Value Gap on where AI ambition meets operating reality, and has built the AI Swimsuit Index to turn that question on public markets: how far companies' AI talk runs ahead of their disclosed proof.