Writing
Blog.
Essays on the systems, frameworks, and human dynamics that make — or break — product delivery at scale.
Get new essays in your inbox.
Free — no paid account required.
Out now on Amazon®
More Clarity, Less Noise
by Rick Pollick
Two volumes — writings on mindset, people, process, product, data, and AI.
Read or Buy on Amazon® →
2026
11 minEvals Are the New Acceptance Criteria: Rebuilding Definition of Done for AI Features
Acceptance criteria were built for software that behaves the same way every time. AI features do not. Here is how delivery leaders rebuild Definition of Done around evals, graded thresholds, and a quality stack of evals, guardrails, and observability so probabilistic features ship with confidence.
9 minDependency Density: A Better Risk Signal Than Critical Path for AI-Era Programs
Critical path was built for projects with clean linear chains. Modern programs do not look like that anymore. Dependency density is the metric that actually predicts which clusters will slip, and how to score, track, and act on it.
29 minMembersGive Your AI Agent a Brain: The Contextual Proxy Pattern for Targeted Context
Most AI agents fail not because the model is bad, but because the agent shows up every morning with no memory, no voice, no priorities, and no idea what your week looks like. A targeted "brain" fixes that. The contextual proxy pattern, the six components, the five-phase loop, the multi-repo and multi-target patterns, the four cadences that grow the "brain" over time, the layered guardrail model, a seven-step getting-started guide, ROI math, and the three failure modes that quietly kill the discipline.
9 minDORA Metrics in the Agentic Era: Why Deployment Frequency Stops Being a Signal
DORA metrics defined modern software delivery for a decade. With agentic AI shipping code on its own, deployment frequency and lead time can lie to you. Here is the new metric stack engineering leaders actually need.
11 minValuestream Episode 1: See It, Own It, Move It. Where Value Actually Flows in Modern Software Delivery
The launch episode of Valuestream. Nine senior people in a room, forty-five minutes blocked, one decision on the table, and nobody makes it. This is the quietest failure mode in modern delivery, and the three-word fix you can ship in a Google Sheet on Monday. Plus the operating model, platform engineering, and agentic AI layers that turn that one fix into real value flow. Full show notes, the framework, the data, and the Spotify embed.
9 minNot My Problem: A Decision Heatmap Talk from Product Camp Pittsburgh 2026
My Product Camp Pittsburgh 2026 talk on the three quiet failure modes of matrixed organizations, the Owner-Decision-Date framework that fixes them, and the one-page decision-aging heatmap you can implement in a Google Sheet on Monday. Full recap, the artifact, the impact chart, and the deck as PPTX and PDF.
13 minThe ADR Comeback: Anchoring Agentic Engineering Teams With Architecture Decision Records
Agentic coding tools are refactoring services that humans spent years getting right, and most of them have no memory of why the original design existed. Architecture Decision Records were a tidy idea when humans wrote all the code; they are now load-bearing infrastructure for any team where an AI agent touches production. A practical playbook for the ADR discipline, the new metrics that matter, and the three failure modes to avoid.
9 minMembersWhen to Burn the Plan: A Tactical Re-Baseline Playbook for Slipping Programs
Most project managers treat re-baselining as failure. The best ones treat it as the highest-value move available. A tactical playbook for when to call the reset, how to prepare the executive brief, the five-step conversation framework, and the new commitment statement template that holds. Includes three real-world cases and the post-meeting moves that turn a re-baseline into renewed credibility.