What score counts as real work?

Five out of five is real work. Three or four means fix the failing checks before scaling. Below three, the initiative is Optimization Theater and should be killed or rebuilt.

Why does a kill line matter?

An initiative with no defined kill line cannot be stopped by evidence, only by budget exhaustion. The kill line is what turns a bet into a decision.

Framework · Receipts check

PROOF

PROOF is a five-part scorecard for any AI initiative: Prove where the data came from, Run it again with the same quality, Outcome tied to money, Owner has a name, Failure mode is clear. It turns the Reversal Ledger discipline into a scoreboard a leader can run in a Monday meeting.

Share this framework

Posting to Instagram or TikTok? Copy the link, it carries the title, summary and share image.

What it is

Reported adoption and real work are different numbers. PROOF is the scoreboard that tells them apart before the CFO does.

Every check is a yes or no. An initiative that cannot answer all five is not ready to scale.

The five checks

Prove where the data came from

You know the source of the data, the model, and the inputs. Anonymous data plus a rented model plus an unknown prompt is an unauditable system.

Run it again

If it worked once, that was a demo. Repeatability is the difference between magic and machinery.

Outcome tied to money

The metric is a business outcome, not an activity count. Emails sent, meetings booked, and hours saved are inputs. Revenue, retention, gross margin, and cycle time are outcomes.

Owner has a name

A named human owns the outcome, not the tool. Tools do not miss quota. People do.

Failure mode is clear

There is a defined line at which you kill or replace it. An initiative nothing could kill is not real work.

The visuals

Figure

The PROOF report card

Five passes is real work. Three or four is fix these. Below three is kill it.

PProvenanceDo we know where the data, model, and inputs came from?
RRepeatableCan we run it again tomorrow and get the same quality?
OOutcomeIs the metric tied to money, not activity?
OOwnerDoes a named human own the outcome?
FFailureIs the kill line defined and dated?

Score: ___ / 5

Figure

Reported adoption against operational work

Reported adoption means a company says it uses AI. Operational means the work runs in production with a business result attached.

Reported, not operational 70.4%
Operational 7.6%
No reported adoption 22%

Reported adoption 78 percent, operational 7.6 percent. Momentum field analysis of 1,000+ calls across 150+ industries, January to July 2025, cited by Jonathan Kvarfordt on The AI Gap.

What you get out of it

Run PROOF and you get a scoreboard you can defend in front of the CFO and the board. Initiatives below the line get fixed or killed. Initiatives above it get scaled.

A defensible line between real AI work and Optimization Theater.
A quarterly scoreboard where every AI initiative has a score and a named owner.
A finance partner who trusts the AI budget, because every dollar maps to a scored initiative.

The mistakes it prevents

The demo that never scaled: It worked in a keynote and nobody reproduced it in production. Repeatability failed. Fix it or kill it.
The activity count: Ten times more emails is not an outcome. It is an activity, and it is where the adoption gap hides.
The tool as owner: A tool cannot miss a quota. A named human has to own the outcome.
The unkillable initiative: Large sunk cost, no defined kill line, years past usefulness.
The unauditable pipeline: Unknown data into a rented model with an unknown prompt. When it breaks, nobody can trace why.

How it is used here

PROOF is the standard scoreboard for every case, teardown, and post-mortem in the Report. Reversal Ledger entries and playbooks carry a score, and readers run the same scoreboard in their internal review cycle.

Don't worry about the AI metrics of who's adopting what. What matters is where's your business at and is the business metrics improving. Because then you know if AI is successful.
Jonathan Kvarfordt, RevOps Champions podcast

A lot of AI projects fail because they don't have good data, good clean data or structured data. 75 percent of the time is spent on preparing data.
Jonathan Kvarfordt, The AI Gap podcast

Terms used here are defined in the glossary and explained in plain language in the AI and Revenue Dictionary.

Apply it with a skill

Each skill turns this framework into a job you can finish. Copy the quick prompt for one task, or download the SKILL.md file, a reusable set of instructions for an AI assistant, for repeatable work.

Prove an AI Result Is RealA scorecard that separates a repeatable AI result from a good week, before you present it or fund more of it.

All skills and prompts →

Common questions

What does PROOF stand for?: Prove where the data came from, Run it again with the same quality, Outcome tied to money, Owner has a name, Failure mode is clear. Five checks scored out of five.
What score counts as real work?: Five out of five is real work. Three or four means fix the failing checks before scaling. Below three, the initiative is Optimization Theater and should be killed or rebuilt.
Why does a kill line matter?: An initiative with no defined kill line cannot be stopped by evidence, only by budget exhaustion. The kill line is what turns a bet into a decision.

Cite this framework

Kvarfordt, Jonathan. "PROOF." The Revenue AI Report. https://www.therevenueaireport.com/frameworks/proof

All frameworks →