Structures

The question this section answers: how can a demand stay understandable while its interpretation changes, different people and systems work on it, and decisions about it are made at different moments?

The model rests on a hypothesis: AI could change the cost of learning, producing, coordinating and deciding. The structures described here do not cause that change. They are what makes it usable. Faster learning or execution is worth little if every new participant has to reconstruct the intention, the criteria and the commitment from scratch.

How the pieces fit

• A Delivery Unit keeps the reference to the demand and its context across executors.
• A Delivery Stage bounds an advance that can be evaluated.
• A Stage Contract makes explicit the conditions for carrying out that advance and judging it.
• The Delivery State supports continuity and resuming work, by showing what has happened and what allows or prevents moving on.
• Artifacts are the products of the work.
• Evidence is the information that supports decisions and offers about them.

Read together, they preserve three distinctions that the rest of the model depends on: an output that has been produced, an output that has been accepted, and continuation that has been authorized.

Where to start

If you are new to the model, start with the Delivery Unit, then follow the pages in order. If you want to see the structures at work first, the example Delivery Unit applies all of them to one demand.

Relation to the other sections

This section defines the units, conditions and information that represent delivery. The resources, policies, pools, coordination and evaluation that carry the work out belong to the Software Delivery Engine. Estimates, commitments and economic decisions are covered by Progressive Estimation, Progressive Commitment and the Commercial Gate.