Frameworks
The frameworks we test AI claims with
Each framework here is used in the reporting, not invented for a slide. Definitions are stable, sources are named, and every framework carries the decision it is supposed to produce.
Ask the frameworks
Rather read the right framework by asking a question? Skip ahead to the question-based advisor. Type what you are trying to decide, and it maps your question to the primary framework, a paired framework, the first move, and the failure mode to avoid.
Canonical definitions
The four frameworks this publication is cited for most
Stable reference definitions with formulas, thresholds, and the decision each one produces. Built to be cited.
The Proof Gap
The measurable distance between an enterprise's AI spend and its AI-attributable revenue in a given quarter, calculated as total AI spend divided by verified AI-influenced pipeline.
Read the definition →The Reversal Ledger
A running, named record of AI tools that B2B revenue teams have shut off, with reason codes and dollar cost, maintained by The Revenue AI Report.
Read the definition →The Eight-Seat Read
A fixed eight-metric AI operating scorecard cut by revenue function: Sales, Marketing, RevOps, Enablement, Customer Success, Partnerships, Executive, and Revenue Finance.
Read the definition →Optimization Theater
The pattern of AI programs producing activity metrics, such as pilots, dashboards, and time savings, without producing revenue metrics.
Read the definition →Compare
Which framework answers which question
Filter by the stage of work, then select a framework to see what it gives you, what it does not cover, the framework it pairs with, and the published research to read alongside it.
Whole-org read
The Eight Seats
Who owns which part of this AI deployment, and what does each one decide?
- What it gives you
- Eight named owners, eight decisions, eight failure modes, kill criteria per seat.
- Not the right tool for
- A single-team experiment with one owner and no cross-function dependency.
- Symptoms it fits
- Nobody can say who owns the tool after the pilot
- The win one team reports is disputed by another
- A deployment was reversed and no seat had kill criteria
- Works together with
. Set the ambition level first, then assign the seats that ambition actually requires.
. Each seat brings its own receipts instead of one shared slide.
- Evidence to read with it
The Eight Seats baseline. Seat-by-seat evidence on which functions are losing definition, with sample sizes stated.
What separates the deployments that work. Named ownership is one of the separators in the published data.
Ask the frameworks
Describe the problem, get the framework and the evidence
Type the situation you are in. The answer is drawn only from the published frameworks and research on this site, with the source named. It does not invent numbers, and it will say when the archive does not cover your question.
Start from a common one
Next
From framework to tooling
Once you know which framework answers your question, the AI Tech Landscape shows what is actually being sold against that question. Some of it holds up, some of it is deck copy.
Open the AI Tech Landscape →Shorter definitions live in the glossary and the AI and Revenue Dictionary.