Articles
Notes from shipping software with AI
For engineering leaders who want AI-assisted delivery they can check. Written from our own work, not from the sidelines.
Topics
- AI agents (13)
- AI Coding Agents (1)
- AI delivery (88)
- AI Delivery Programme (12)
- AI Workflow Audit (2)
- Architecture Review (2)
- Laravel (1)
- MVP to Production (1)
Start here
Review capacity
The review queue is the bottleneck AI coding created
Why more code did not mean more shipped, and where the queue actually forms.
Instruction drift
Agent instruction drift in enterprise AI teams
How every developer ends up with a different AI, and what a shared rule set changes.
What was tested
AI coding checkpoint contract: make checks reproducible
What “tested” should mean before a pull request is allowed to say it.
Latest
-
How do you stop new developers asking AI the same codebase questions every week?
Treat repeated questions as missing project memory; turn verified answers into maintained module notes, commands, decisions, or troubleshooting steps.
-
Why is approving an AI tool different from approving its development workflow?
Tool approval covers a product and its configured access; workflow approval covers how people use it, what data crosses boundaries, and which actions require…
-
How do you know whether an AI pilot only worked because one senior developer babysat it?
Give the same workflow to a second group and remove the original champion from execution; compare delivery, review, and recovery evidence.
-
What should you do when CLAUDE.md, GEMINI.md, and AGENTS.md disagree?
Choose one repository-owned source of truth, document the precedence rule, and generate or check provider files from that shared policy.
-
Which AI coding rules should become executable checks?
Automate rules with an objective signal and a safe failure message; leave architectural judgment with named reviewers.
-
How do you keep different AI coding agents from following different rules?
Keep one shared policy for project rules, then use thin provider adapters only where tools genuinely differ.