§ 01The claim
Make AI employable
for real work.
Your team is already using AI. Contractual tells you what work AI can actually own.
0%up to40%Arms I–III · Prompting aloneArm IV · Under Contractual
Loading
Your team is already using AI. Contractual tells you what work AI can actually own.
The share of a work queue AI completes end to end, confirmed in your system.
Synthetic benchmark · six models · 100 cases · four arms. Full methods in the report.
The same models score zero completed cases under prompting alone. They can decide. They cannot finish.
Replay historical work
Run completed cases back through the rules.
Measure what AI can finish today
You get one number: how much of the queue AI finishes on its own.
Draw the delegation boundary
Green, yellow, red — written down and versioned.
Expand autonomy as evidence compounds
AI takes on more only where the results back it up.
Six frontier models · 100 cases · four arms · one question
Fig. 2
The Delta Index.
| Prompting alone | Under Contractual | ||||
|---|---|---|---|---|---|
| Model | A1Prompting | A2 | B | Rate | |
| gpt-5.6-terra | 0% | 0% | 0% | 38% | |
| gpt-5.6-luna | 0% | 0% | 0% | 40% | |
| claude-opus-5 | 0% | 0% | 0% | 40% | |
| claude-sonnet-5 | 0% | 0% | 0% | 40% | |
| claude-haiku-4-5 | 0% | 0% | |||
Two weeks later you get a decision, measured against work your team already did.
| 0% |
| 40% |
| claude-fable-5 | 0% | 0% | 0% | 40% |
|---|
Source: Delta Index, note 1.