The Coroin Books Rebuild · Part 2 of 2

Now to get it found

Being found in the AI era isn't magic, and it isn't a guru tax. It's structure you can build yourself.

Published 2026-07-06 · 2 min read

You can build a beautiful website and still be invisible to the tools people now use to find things. Here is how I made an independent book imprint findable by search and AI assistants in one afternoon. Real identifiers. Clean schema. Crawl rules you can set up yourself.

Cover tile for the findability case study, with before-and-after scores rising across SEO 74 to 88, AEO 72 to 90, GEO 63 to 85, and a unified score 69 to 87.

Last time, I told you the site was built to be found. Schema, crawl files, all wired in from the first commit.

That was true. It also wasn't enough.

Wiring in the plumbing is not the same as being findable. A machine could see the shape of a book on the page. It learned nothing about which book.

So two days later, I filled it in.

"Getting found" used to mean ranking on a page of blue links. Not anymore. Now people ask an assistant. "Who published this memoir?" They read one answer. They don't click ten.

If the assistant can't tell who published what, you are not in the answer. You don't exist.

And somewhere right now, someone is paying an expert $18,000 to hear they won't get found on AI without some special framework.

The framework is structured data. You can do it yourself.

Three letters at a time. SEO is being found by search. AEO is being the source an assistant quotes. GEO is being cited inside the AI's answer.

None of it is magic. It is structure. Real identifiers. Crawl rules. Showing up. Boring, repeatable work. Which is exactly why it is yours, and not a guru's.

Here is the whole move. It took an afternoon and an evening.

I made one data file the single source of truth. Now the schema a crawler reads can't drift from the page.

I backfilled real identifiers. ISBN and LCCN for all five books. ISNI for the imprint and each author. "Trust me" is not an identifier.

I linked every author to the places they already exist. Their own sites. LinkedIn. The library and bookseller records.

That last part mattered most. We went from zero outside links to a real web of them. It was the biggest single jump in my before-and-after audit.

Authority tile: socials and a self-written site do not prove authorship, but a registry-backed ISNI does. The imprint plus Richard, Erik, and Maddie each now carry one, and GEO rises 63 to 85.

I also killed a render-blocker. The largest paint dropped from 5.3 seconds to 1.1. A page the machine can't load fast, it quietly buries.

Speed tile: a render-blocking font pushed largest contentful paint to 5.3 seconds. The same font, loaded differently, brings it to 1.1 seconds, and Lighthouse performance from 0.74 to a perfect 1.0.

So here is your move today. Give the machines real facts. Tell them where you already exist.

Not a framework. Your actual identifiers and an hour.

The ilities still matter. Findability is one of them.

There's a reason a publishing imprint cares this much about being found. That's the next story.

Closing tile: findability is not won once. Search engines and AI reward fresh work, so the imprint keeps creating, with translations coming soon.

Website by Claude Design + Claude Code. Carousel by Claude Design.

coroinbooks.com · @erikcto

Erik Larson is a 7x CTO and private-equity advisor who taught himself to code and now directs AI agents like a dev team. He writes at erikcto.com and runs the publishing imprint Coroin Books.

Also on LinkedIn · X.

Now to get it found

Permission to Reprint