Getting Cited by the Machines

When I started All Angles Creatures, I made a decision that sounded slightly insane in 2024: I was going to treat AI answer engines as a real acquisition channel, on day one, with the same seriousness most people reserve for Meta or Google.

It was not insane. It was early — which is a different thing, and a much better one.

The problem, stated plainly

I was one person building a DTC subscription brand in a commoditized category. I could not out-spend anyone, and I could not out-staff anyone, because there was no one else. What I could do was get found — and "found" was quietly changing underneath everyone's feet.

A buyer with money in hand no longer types three keywords into Google and scans ten blue links. She asks an assistant a full question — what's the best option for my specific situation, and is it worth it? — and she gets back a written answer with two or three citation chips underneath. She never sees a list. She sees an answer, and the only brands in that answer are the two or three the machine decided to read out loud.

So the real question was not "how do I rank." It was: how do I become one of the sources the machine chooses to quote?

The loop I built

The instinct most people have here is to write more content. That is the artifact trap, and it does almost nothing. What I built instead was a loop.

First, content engineered to be cited, not just to rank — direct, quotable answers to the exact questions buyers ask, with clean structure, strong entity signals, and schema, so a machine could parse precisely what I was claiming and trust it enough to repeat it. Answer Engine Optimization, before the industry had settled on the name.

The model can only cite what it can read. It reads structure and specifics, and it has zero patience for the performance of expertise where it expected the substance of it.

Then — and this is the part almost nobody does — I built the measurement. There is no Search Console for AI search. So I built my own citation engine: a system that runs hundreds of real buyer questions across ChatGPT, Perplexity, and Google's AI surfaces, captures exactly which sources each one cites, and measures my citation share against competitors, query by query. It turned AI visibility from a guess into a number.

That closed the loop. Measure where I was being out-cited, open up the answers that beat me, reverse-engineer what made those sources more quotable, rebuild my content to win that exact citation, re-measure. A channel I could actually manage, with a metric and a feedback cycle, instead of hoping the machines noticed me.

The result

All Angles earns active, repeatable citations in ChatGPT and Perplexity in its category, alongside first-page Google rankings — and it got there at near-zero marginal cost, run by one person, while I was also building everything else. The same brand I scaled solo to $1.56M in annual recurring revenue in under a year.

The lesson I keep coming back to: AI search rewards being the clearest, most trustworthy answer to a specific question, and then measuring how often the machines actually choose you. That is a discipline you can build on purpose. Almost no one is doing it yet.

Which is exactly why it's sitting there waiting.