Start with the systems that matter.
An ARI pilot maps 10–20 repositories to business impact and the coding-agent tasks you want to enable. Leadership gets a baseline and a clear sequence for improvement.
Three decisions. Independently scoped.
Primary assessment for coding-agent adoption. Agree the target tasks, repositories, agent, and review constraints. Receive the static baseline, pilot priorities, and remediation plan; include approved command checks and task trials in the agreed scope.
Independent review of reproducibility, verification, documentation, and maintainability. Combine repository findings with an engineering walkthrough. Receive a health assessment and prioritised improvement plan. Broader architecture analysis is agreed separately with the system owners.
Independent review of verification, release gates, ownership, and operational controls. Agree access to CI, deployment, and rollback evidence where relevant. Receive a control-gap register and remediation priorities; production conclusions depend on the operational evidence reviewed.
Each review can be purchased on its own or alongside ARI, with its own scope and executive readout. The reviews combine tool evidence and expert assessment.
A representative slice of your systems.
Include Tier-1 production services, high-change areas, a legacy concern, a platform/library, and at least one lower-criticality comparison.
For each repository: system, owner, criticality, production status, sensitivity, and known dependencies. Paths stay inside your environment.
Provide the documented setup and test commands, CI policy, and deployment/rollback owners where available. For ARI, include the agent tasks you want to enable.
Name the person who can agree the scope, approve any task trials, and own remediation.
Diagnose, improve, prove.
- Week 0: agree inventory, target agent tasks, profiles, access, and what can be executed.
- Week 1: run private static scans; review evidence with senior engineers; run approved command plans.
- Week 1: deliver the ARI portfolio readout, risk register, and pilot priorities.
- Weeks 2–5: address the highest-priority controls; where approved, run bounded agent/verifier trials as separate evidence.
- Week 6: reassess with the same rubric and separate repository change from rubric change.
A decision you can act on.
At the end of an ARI pilot, leadership should know which systems need foundational work, which are candidates for supervised task trials, which evidence is still missing, and which improvements measurably moved the baseline. Independent reviews agree their own outcomes and timeline during scoping.
The template is a local JSON starting point, not a form submission. This site does not collect repository paths or source code.