This is the build log: a running account of what we are doing, what the pipeline produces, and where it falls over. Posts start as the dogfood month runs. No invented numbers, no polished retrospectives before the work is done.
We are running our own publisher sites through Quarrybank this month. That means the agent pipeline finds keyword angles, scores their winnability, drafts the content, runs it through an independent quality gate, and queues it for human approval. We do not publish anything the gate rejects, and we do not fabricate results.
The log will track: how many pieces the pipeline produces per week, how many clear the quality gate on the first pass, which ones need a second attempt and why, and what early citation signals look like in Google AI Overviews and similar systems. When numbers arrive, they will be real ones, with the caveats attached.
We will also write about what breaks. The agent does not always get the angle right. The quality gate catches things the agent missed, and sometimes flags pieces that are fine. The pipeline itself has rough edges we are fixing in real time. That is the point of building in the open: the messy middle is visible, not just the tidy summary at the end.
There are no posts here yet. The dogfood run started this month. The first write-up will arrive once there is something honest to report: a real batch of articles through the pipeline, with actual pass rates and actual outputs. Until then, the Learn section below covers the strategic context that frames what we are building and why.
AEO is the discipline of structuring content so that AI systems, not just search engines, cite your page as the authoritative source. The criteria are different from classic SEO, and the gap is widening.
Getting cited in AI Overviews is not the same as ranking on page one. This guide covers what the research shows, which signals appear to matter, and what actually changes when you put them into practice.
A content moat is the compound advantage you build when topical depth, real data, and a consistent editorial voice make your site progressively harder to displace. Here is how it forms and why it matters for AI citation.
The simplest way is to join the waitlist. We send a write-up when the first real dogfood numbers land: pass rates, citation signals, what the pipeline produced, and what we changed as a result. No marketing cadence, no drip sequence. One email when there is something worth saying.
You can also see how the pipeline works and what the agent flow looks like in practice, both on the main site.
Join the waitlist and we will send one email when the dogfood results arrive. No invented figures before then.
Join the waitlist