Phase 1: Recovery

What we found:

An application running in production without meaningful monitoring. When something went wrong, it was reported by users, not detected by the system. Deployment was manual — changes applied directly to production without a staging environment. No rollback procedure existed.

What we fixed:

• Monitoring deployed: application error tracking, uptime monitoring, performance dashboards

• Staging environment established: no change goes to production without staging validation

• Deployment process documented and automated: changes deployed via CI/CD pipeline with rollback procedure

• Critical incidents responded to with formal post-mortems — documented, root cause identified, recurrence prevention implemented

Law 1 in practice: Every incident was previously reported by users. After monitoring was deployed, incidents were detected internally before users were affected. The distance between problem occurrence and problem awareness collapsed from hours to minutes.

Phase 2: Maturity Uplift

With stability restored, we systematically assessed the application against a maturity framework covering: code quality, testing coverage, deployment practices, monitoring completeness, documentation, and security posture.

Halcrow's maturity assessment framework rates each domain from "Poses Risk" through Bronze, Silver, and Gold. The assessment gave Attention Nest's leadership a clear picture of where the product was strong, where it was vulnerable, and what the improvement path looked like.

Over three months, we moved the highest-risk areas from "Poses Risk" to at minimum Bronze standard — establishing baseline practices that a new team could maintain without specialist knowledge.

Phase 3: Hand-Back

A hand-back without preparation creates a new version of the original problem. We structured the hand-back as a formal knowledge transfer:

• Service runbook: complete operational documentation covering architecture, monitoring, incident response, deployment, and escalation paths

• Handover Slack channel: outgoing team and incoming team in the same channel for four weeks, with the incoming team escalating questions before the outgoing team disengaged

• Codebase walkthrough: recorded session covering architecture decisions, technical debt known and documented, and the reasoning behind key implementation choices

• First 30 days post-handover: Halcrow on reduced availability (2-4 hours/week) to answer questions the incoming team couldn't resolve independently

The incoming team's first unassisted incident — a deployment issue detected by monitoring two weeks after handover — was resolved by the team using the runbook, without Halcrow involvement.

Law 8: Activity is not progress. The maturity work wasn't about adding features or lines of code. It was about building the operational infrastructure that makes a product sustainable to run.