Prove
What changed because AI was used, and how strong is the evidence?
Cost and usage are recorded automatically; value usually is not. Prove ties a workflow to an owner, a baseline and a measured outcome, then labels the claim with how strongly it is known. On this site's proposal, no monetised claim stands below evidence level E2, and released time is not treated as realised value until it shows up in an accepted record.
Do this
Templates, guides and methods for this phase, in a sensible order.
Browser tool, v0.5
Attribution coverage calculator
Calculate what share of AI spend is connected to a measured outcome. Runs in your browser; nothing is sent anywhere.
Hub proposal, v0.5
The AI Value Ledger
One record per material workflow: owner, baseline, cost, evidence level and a decision. With a template.
Excel template, v0.5
AI Benefits Register
Track each benefit by type, owner and booking record, so only value that has reached an accepted record counts as realised.
Tool
AI Economics KPIs
The KPI library across seven governance domains, led by attribution coverage.
Guide
How to Measure AI Value: From Observed Change to a Defensible Investment Decision
The steps between 'it saved time' and ROI: a baseline, a quality floor, netting out rework, converting with care, then deciding. With three worked examples.
Guide
A Proof of Concept That Proves the Technology Has Proved Almost Nothing
Technical feasibility is not evidence of business value. An AI pilot must prove cost, workflow change, adoption, risk and value capture.
Guide
What CFOs Should Ask of AI ROI Claims
Most AI ROI cases are built to survive scrutiny, not invite it. What a credible AI business case looks like, and why CFOs should insist on the difference.
Related pages
Practical pages that sit alongside the templates.
The thinking behind it
The models and evidence that explain why this phase works as it does, in the Understand section.
Hub proposal, v0.9
The AI Value Gap
Why AI cost is visible and AI value is not, and how attribution coverage measures the difference.
Hub proposal, v0.9
AI Value Attribution
Connecting each unit of AI work to a baseline, an owner and evidence of change.
Reference explainer
AI ROI Models
Return measured across several dimensions, built to survive board scrutiny.
Adapted concept
The productivity J-curve
Why AI returns lag investment by design, and how to tell a dip from a failure.
Reference explainer
How to read an AI metric
The distinction ladder from activity to strategic impact, then the set: attribution coverage, cost per successful outcome, value dimensions and capture status.
Evidence
What the Engineering Evidence Says About AI Coding Value
Studies of AI coding assistants point different ways. Task speed, delivery performance and perceived productivity differ, so check which one a claim rests on.
Essay
The Token Is the Meter, Not the Value
Tokens make AI consumption measurable and priceable. They do not tell an organisation whether anything valuable happened.
Essay
The Behavioural P&L of AI
AI adoption changes trust, review, learning, workarounds and decision behaviour. Those effects create real economic assets and liabilities.
Next phase
Each phase feeds the one after it.