Blog · July 2026

How I Got COLAClear Cited by Name in Google’s AI Overview

A first-person account of the work behind a by-name AI citation — no budget, no team, no tricks.

A few months after launching COLAClear, I typed a question into Google that a winemaker might ask — roughly, “tool to check my label before submitting to the TTB.” And there it was, at the very top: Google’s AI Overview, the AI-written summary that now sits above the blue links, describing COLAClear by name. Not a result buried on page two. The machine had read my site, understood what the tool does, and recommended it in its own words. I’d built the whole thing solo, spent nothing on advertising, and launched only months earlier. Here’s exactly how that happened — because none of it was luck.

First, the context

COLAClear is a tool I built that checks U.S. wine, spirits, and beer labels against federal regulations before a producer sends them to the government for approval. It’s a narrow, unglamorous niche — which, it turns out, is part of why this worked.

The bet: my buyers had started asking AI

The instinct with a new product is to chase big keywords and hope Google ranks you. But two things were true. A brand-new site has no authority — you’re not out-ranking anyone on a broad term like “wine labels.” And, more importantly, my buyers had changed how they look for answers. A small producer with a labeling question in 2026 doesn’t always open Google and scroll — they ask ChatGPT, or read Google’s AI summary, and move on. If I wanted to be found, I had to be the answer those systems give, not just a link they might list.

What I actually did

I wrote for the questions people actually ask. Instead of targeting “wine labels,” I built pages around the exact things a stressed producer types: “how to fix a rejected COLA label,” “what does ‘Needs Correction’ mean.” Each got a direct, plain answer. Those question-and-answer pages are catnip for AI, because they map one-to-one onto how people ask.

I told the machines what COLAClear is, explicitly. I added structured data — invisible tags that label the site as a software tool, name me as the founder, and carry a real customer review. And I wrote an llms.txt file: a short brief that sits at the root of the site, written specifically for AI systems, spelling out what the tool covers and the exact regulations it checks. When the AI later described COLAClear, it used those very facts.

I made sure the AI crawlers were allowed in. This sounds obvious, but plenty of sites accidentally block the bots that feed AI answers. I checked that mine welcomed them — because if the crawler can’t read your page, you can never be quoted.

I backed it with real writing. I published industry analysis and placed bylined articles in the trade press my audience reads — Wine Business Monthly, Wine Industry Advisor, Craft Spirits Magazine. That earned the kind of authority algorithms respect, and those articles now rank on the same searches, reinforcing the signal.

Everything pointed at one consistent name and one story, everywhere COLAClear appeared.

The result

Within months, and with zero paid media, Google’s AI Overview cites COLAClear by name as the automated pre-screening tool — and even reproduces its pass / review / fail logic. It ranks page-one for the searches its buyers actually run, and the trade-press pieces rank alongside it. The AI didn’t just list COLAClear; it explained it, correctly, and recommended it.

Why it’s repeatable

The encouraging truth is that none of this needed a budget or a following. It needed a handful of specific things done deliberately: write for real questions, tell the machines plainly what you are, let them in, and earn authority the slow, honest way. In a niche, that’s often enough to become the answer — because in a niche, almost no one else is bothering.

Want the step-by-step version? It’s the Playbook. Want me to do it for your business? That’s Work with me.

Sources & further reading