The Loop Dispatch
Issue 006 · March 14, 2026
Lead essay · Quality Engineering
Most QA functions were built for a world where humans wrote all the code. That world is gone. Here's the operating model we've deployed across 30+ engagements to replace it.
Most QA functions were built for a world where humans wrote all the code. That world is gone. Here's the operating model we've deployed across 30+ engagements to replace it.
About the Author
Ben Fellows
The fundamental problem with traditional QA isn't headcount, tooling, or process. It's the operating model itself. QA was designed as a gate…
All Articles
Most QA functions were built for a world where humans wrote all the code. That world is gone. Here's the operating model we've deployed across 30+ engagements to replace it.
Ben Fellows
8 min read
Everyone is shipping AI agents with giant prompts. The teams getting real leverage are building purpose-built pipelines with roles, permissions, and project-specific rules instead.
Marcus Rivera
6 min read
We redesigned the TDD workflow for AI-assisted development. Tests define intent, AI accelerates implementation, and engineers own verification. The results changed how we think about quality ownership.
Sarah Chen
7 min read
Velocity metrics tell you how fast you're moving. Leverage metrics tell you whether the speed is sustainable. Here's the measurement framework we use with every client.
Ben Fellows
5 min read
Most maturity models are academic exercises. This one was built from 30+ real assessments and tells you exactly what to fix first based on where the leverage is highest.
Priya Sharma
6 min read
The SDET role is being redefined. Manual test execution is gone. What remains is automation engineering, infrastructure design, and the ability to operate AI-assisted workflows. Here's how we hire for it.
James Okafor
5 min read
How readers use the writing
The essays are most useful when they give you language and evidence for an argument you already wanted to make.
“I sent “Coverage Is a Vanity Metric” to my CTO. Two weeks later we redefined our QA OKRs.”
Series-B fintech · ~50 engineers
Used: Coverage Is a Vanity Metric
“The AI-generated tests piece reframed our vendor review. We rewrote the evaluation rubric the same week.”
Healthcare SaaS · 120 engineers
Used: Why Your AI-Generated Tests Are Worthless
“I forwarded the staffing-model essay to our head of engineering. It started a conversation we'd been avoiding.”
E-commerce platform · 200+ engineers
Used: The Staffing Model Is Broken
Names + companies anonymized at the speakers' request.
Watch · From the desk
Apr 30, 2026
Stop treating AI like factory work. Rigid, assembly-line workflows break down with complex codebases. Owning your agentic pipeline means customizing every step and refreshing context each time—leading to better accuracy, flexibility, and scalability where it actually matters. #agenticai #agentic #agenticcoding #coding
Apr 30, 2026
Why “agentic factories” don’t actually work Trying to force one system to handle every codebase is like expecting one factory to build every type of car - it just doesn’t scale. Instead, the future is agentic pipelines: flexible, tailored workflows built around your specific repo, while still reusing powerful components like agents, prompts, and shared memory. Smarter systems aren’t universal - they’re purpose-built. #agenticai #agentic #agenticcoding #coding #programming #code
Apr 30, 2026
Most people trying to do agentic development are building what I call “God agents” — one giant system that tries to do everything. In my experience, that approach breaks down quickly. It hits context limits, becomes hard to reason about, and fails in inconsistent ways. The result is usually more time spent debugging the agent than actually building software. This video is about a different approach. Instead of building one massive agent, I break down the five pipeline categories that I use across my codebases to make agentic development actually work at scale. These aren’t meant to be perfect or universal, but they’ve been a solid foundation for structuring real systems. The five categories I walk through are: surface area pipelines, change type pipelines, failure mode pipelines, integration pipelines, and confidence pipelines. Each one exists for a reason. As I’ve worked more with AI, one pattern has become clear. The failures are unpredictable at first, but over time they repeat. The same classes of mistakes show up again and again. These pipeline categories are designed to target those failure patterns directly, with loops and checks that are specific to the kinds of problems you’ve already seen. The goal isn’t to create a single “correct” pipeline or a magic factory. It’s to build a system of smaller, focused pipelines that are tailored to your codebase, your architecture, and the ways AI tends to drift in your environment. If you pair this approach with something like policy as code, you end up with both structure and control, which is where agentic development starts to become much more reliable. This video is meant as a starting point. These categories will evolve, and I’m still refining them as I see new patterns emerge. If you’re building with AI and trying to move beyond simple prompts into something more scalable, this should give you a useful framework to start from. Would love to hear how others are structuring their pipelines and what categories you’re seeing in your own systems.
Every article is backed by a published methodology paper. Download the frameworks, run the assessments, and see the full system.
90-Day QA Leverage Plan
Coming soonUse it on Monday · Editable doc + Notion template