Factory pilot
Prove the Factory on real delivery.
One real value stream, one running Factory slice, and delivery evidence you can inspect. The pilot makes the readiness roadmap real before you scale it.
Scoped from the Factory Readiness Assessment
The output
A working reference, not a concept deck
The pilot is deliberately narrow enough to learn and real enough to matter.
Running Factory slice
Orchestration, repository integration, work isolation, quality gates, review, and human handoff working on a real value stream.
Delivery data
Baseline and pilot signals for throughput, AI use, change confidence, failure, and the client-specific outcome.
Operating constraints
The technical and organizational limits exposed by real use, written down before wider rollout.
Scale recommendation
What should expand, what must change first, who should own it, and how the rollout should proceed.
Selection
Choosing the pilot
The pilot decides what the result is worth. On the wrong project you learn little you can act on.
A real, active project
Work that is already on the roadmap and would happen without us. A sandbox mainly shows that the sandbox works.
Urgency and a management mandate
Real urgency behind the work, and management that has put its mandate behind the pilot. Without that backing, the pilot tends to lose priority to whatever escalates next.
Willing participants across experience levels
People who take part by choice, from senior to junior. Mixed experience keeps the result relevant to the whole team instead of to one group.
Access to codebase and CI
The team works in the real repository with the real pipeline. What gets built beside it is hard to carry back in.
What we rule out
Toy sandboxes, forced participation, all-junior staffing, and work under crisis pressure. Each of these makes the result hard to read and hard to transfer.
The pilot
One bounded learning loop
Scope, duration, and delivery shape come from the assessment rather than a fixed package.
-
01
Select the value stream
Choose a real product or engineering flow with enough value, evidence, and organizational support to learn from.
-
02
Build the slice
Connect intent, context, agents, repositories, CI, review, deployment, and human steering.
-
03
Run and observe
Use the Factory in real work. Capture delivery, quality, confidence, cost, failure, and handoff signals.
-
04
Codify and decide
Turn manual fixes into rules or tests, document constraints, and recommend the next rollout step.
Scope
What the Factory slice can include
We cut the slice to fit the chosen value stream. The pilot builds what that stream needs, not every component on this list.
Work tracking and state
Where a piece of work lives, what state it is in, and how an agent picks it up and hands it back.
Context and orchestration
The layered context an agent reads before it starts, and the rules that decide which agent works on what.
Repository and CI integration
Agents work in branches, open pull requests, and pass the same pipeline as everyone else. No side channel, no special path to main.
Tests and review
Executable checks and the review path a change passes before it merges. Agent review and human review sit on the same change.
Tracing and cost visibility
What a run did, what it cost, and where it failed. Without traces, the pilot ends in opinions about whether it worked.
Human and agent handoff
Where a person decides, where an agent continues alone, and how the handover works in both directions.
Pilot evidence
Measure changeability and outcomes, not generated-code volume
A Factory earns the name when it improves the delivery system and learns from every run.
- Lead time and throughput
- Flow
- Review and failure signals
- Quality
- Confidence and handoff
- Human
- Client-specific outcome
- Value
The boundary
Pilot before platform rollout
Big-bang Factory program
Infrastructure and governance are designed before the organization has a working reference.
- Assumptions harden into architecture.
- Teams wait for a central platform.
- Business outcomes stay abstract.
- Learning arrives late and expensively.
One real value stream
A bounded slice creates the evidence for wider investment.
- Architecture meets real constraints.
- The internal owner emerges through use.
- Quality and handoff are tested early.
- Rollout scope follows evidence.
Outcome
What the decision produces
The pilot ends with something running and a decision you can defend inside your own organization.
A running reference, built for your context
A working slice in your codebase, together with the artifacts we agree on at the start. It stays with you and your team can keep working on it.
Evidence against an agreed baseline
We agree on the baseline before the pilot starts and look at the same delivery and quality signals afterwards. The comparison is against that baseline, not against an outside number.
Documented constraints
What held the work back, technically and organizationally, written down rather than left as a hallway conversation.
Scale, adjust, or stop
A recommendation with its reasons, and a named person inside your company who carries whichever path you pick. Scaling without an internal owner does not hold.
Pilot pricing
Scope follows the evidence
The assessment defines the value stream, baseline, and measurement plan before we price the pilot.
Before the pilot starts
Can we buy a pilot without the readiness assessment?
The standard route scopes the pilot from the assessment. If equivalent evidence already exists, we can evaluate it rather than repeat work, but the pilot still needs a named outcome, baseline, constraints, and measurement plan.
Which value stream should we choose?
Choose real work with visible value, a mix of AI-curious and skeptical people, enough technical complexity to expose constraints, and leadership support. Avoid a toy sandbox.
Who owns the Factory after the pilot?
Internal ownership is part of the design. The pilot identifies the people, platform-team shape, operating model, and handover work needed for wider rollout.
Build one working reference before you scale
Use the readiness assessment to define the right value stream, outcome, evidence, and scope for your Factory pilot.