Framework
The Absorption
Gap
The distance between the software production capacity an organisation now has and the capacity it can actually absorb.
A capacity problem,
not a tooling problem.
AI agents increase the rate at which intent becomes code. Very little of the surrounding system—verification, architecture, governance, operations, decision-making—increases at the same rate.
The gap between those two rates is where transformations stall. Output grows, and the organisation feels slower.
How the gap shows itself
Output rises while validated business change does not
Review and verification queues lengthen
Architecture decisions lag behind the code being produced
Governance becomes a bottleneck rather than a guardrail
Teams absorb rework instead of releasing value
Variables
What to measureProduction capacity
How much intent can be turned into code and change.
Absorption capacity
How much of that change the surrounding system can verify, integrate and operate.
Constraint location
Where the system is currently limited—and where removal will move it.
Validated business change
The only output that counts.
Where the
constraint moves.
Removing one constraint exposes the next. That pattern held through CI/CD, quality engineering, cloud and platform engineering, APIs and data. AI agents accelerate it rather than end it.
The practical work is to redesign the production system around the new capacity—which is the subject of the AI Engineering Production System.
Read the essay