A note on why this exists

This is the Agentic Engineers journal. Notes from inside the practice, written for operators. If you are looking for thought leadership, you are in the wrong place; if you are looking for what we are actually seeing on engagements this quarter, keep reading.

We started writing publicly because the gap between what the AI-forward category claims and what it actually ships has become impossible to ignore. The category is loud. The operating models behind the loudness are mostly invisible. We are going to be specific about ours, in writing, in public, so that the next founder evaluating an engineering partner has something concrete to read instead of another deck.

What this is

The posts that follow are working notes. They cover what we are seeing on engagements, how the agentic layer is changing the operational shape of an engineering org, what compounds and what does not, what we are tuning month over month, and what we get wrong often enough to learn from. They are written in the voice we use internally, because the voice we use internally is the voice we use with the founders we work with.

A few things shape what gets written here.

1. Operator-grade, not buyer-grade. The audience is the person inside the founder's room making real decisions about how engineering runs. They have read the marketing. They are looking for what is true underneath it.
2. Specific over general. Where we can describe a workflow, a cadence, a failure mode, or a measurable outcome, we do. Where we have to abstract, we say so. We will not write the kind of essay that could have been written by any agency.
3. Honest about limits. We turn down engagements often, including the kind that pay well. We sell systems, not engineers, and the system has a real shape. The journal will be honest about what that shape excludes.
4. Written by humans, drafted with agents. The writing here goes through the same operating system we sell. The first pass is drafted with our internal agentic stack. The final pass is written, edited, and signed off by an operator. The voice is ours. The leverage is real.

Who it is for

If you are a founder or operator trying to move a SaaS product from 1 to N, the writing here is for you. The 1-to-N phase is where the operating system that built the product becomes the thing slowing it down, and the patterns that work at three engineers stop working at eight. That break is the subject we know best, and most of the journal will return to it from different angles.

If you are a buyer comparing AI-forward engineering partners, the writing here is also for you. We would rather you read what we actually do before booking a call than discover it in the third meeting. Skipping a misfit early is better for both sides.

If you are an operator inside a peer firm, you are welcome here too. We have learned a lot from people writing about their own practice in public, and the journal is partly a return of the favour.

What we will write about

Some of the topics already queued, in no particular order: how the monthly review keeps an engagement honest, why we price capacity instead of hours, what the timezone seam between Australia and the US actually unlocks, how code review becomes the highest-leverage agentic workflow on an embedded team, what we look for before taking on embedded technical leadership and why we turn most inquiries down, what "agentic engineering" actually means as a category we are defining instead of a slogan we are using, and the operating-model break that hits most SaaS companies between three and five million ARR.

Each post will try to be specific enough that you could test the claim against your own engagement, and short enough that you would actually read it on a Tuesday morning before the standup.

Where to start

If you have read this far, you probably want to know how to engage. Start a conversation when you want to see the system run against your stack. Plan the next phase when you are further out and want to talk through fit first.

We will write more soon.