Explainer · 2:18
The AI value gap, explained
Organisations can meter AI to the token, but most cannot connect that spend to business outcomes. This short film sets out the gap, the discipline that tries to close it, and where to start on this site: the question you need to answer, the system, a working tool and the toolkit.
Short version · 0:49
The argument in under a minute
The paradox, the gap and the calculator, cut for sharing. Captions are part of the picture, so it reads with the sound off. Every line in it comes from the transcript below.
Transcript
0:00The paradox
AI may be the most precisely metered cost in the enterprise: every token, GPU-hour and seat. But ask what the organisation actually gained, and the answer is far less clear.
0:14The missing connection
Value has to travel a long chain: from investment, to AI activity, to outputs, to successful outcomes, then business value. Each link can leak: output that needs rework, tools nobody adopts, time saved but never redeployed. The Hub calls this the AI Value Gap: spend that cannot be connected to a measured outcome.
0:38AI Value Management
AI Value Management is an emerging discipline for closing that gap. It connects the full cost of AI, including human review and rework, to evidence of what changed, with named owners. Lower cost is not the same as better value. The aim is a defensible decision: scale, hold, redesign or stop.
1:01Start from your question
AI Economics Hub is independent and vendor-neutral. Start from the question in front of you: what an AI use case really costs, whether it is creating value, or how to prepare an investment decision you can explain.
1:16The AI Economics System
The AI Economics System proposes four phases: measure, prove, decide and operate.
1:24Cost per successful outcome
The browser tools make trade-offs concrete. In this illustrative case, the small model is cheapest per attempt. But a person must fix more than a fifth of its work. Count that, and a routed setup costs about a third as much per successful outcome. Change the assumptions and the answer can change, so compare per workload.
1:47The toolkit
The toolkit adds working templates, from a value ledger to a portfolio review pack, free to adapt.
1:55Better decisions
Finance, technology, business and governance teams can then work from the same evidence, and move from ‘what are we spending?’ to ‘what should we fund next, and why?’
2:07Closing
AI Economics Hub. Make AI investment economically legible. Start at aieconomicshub.com.
About this film
- The meter readings in the opening and the flow through the value chain are illustrative, not data. The leak points come from The AI Value Gap.
- The calculator figures are the illustrative defaults of the cost per successful outcome calculator, which reproduce the worked example in The Economics of Model Selection, computed with the calculator's own formula.
- The AI Value Gap, the AI Economics System and the templates are this site's proposals, not industry standards, and the film labels them that way.
- The narration is a synthetic voice generated with ElevenLabs. The music is original and was composed in code for this film. The screens show the site as it was on 27 October 2026.
Where to start
AI Economics in 5 Minutes
The whole argument in one read, with its sources
The AI Value Gap
Why AI value is hard to prove, and how to close the gap
The AI Economics System
Measure, prove, decide and operate, with templates for each
Cost per successful outcome
Compare model setups on the cost of finished work
The toolkit
Templates and browser tools with worked examples
Start Here
Find your path by task or by role