What is an AI visibility audit?

I audited myself and scored 30 out of 100. That surprised me. I run more than 20 newsletters and host a podcast. There's plenty about me online.

The problem wasn't too little information. It was conflicting information.

What does the audit actually check?

One question: what do ChatGPT, Claude and Gemini think about you and your business right now? You ask the models directly, look at what they say, and then look at the sources behind it.

• Do the models know who you are and what you do?
• Do they get your location, services and numbers right?
• Do they recommend you when someone asks for what you sell?
• Where are they getting their information, and does it agree with itself?

Why did I score so low?

Because the internet couldn't agree on who I was:

• One page said I'm in Liberty Hill. Another said Austin. Another named a third town.
• One said 130,000 subscribers. Another said 160,000.
• Some said Larkin Jr. Some said Larkin.

"LLMs don't know if that's the same person."

My view of how these models recommend people: they pick whoever they feel most confident about. Confidence comes from lots of citations of your work, steady content, and all your information matching. Conflicting facts cut that confidence even if every individual page is accurate for the day it was written.

What do you fix first?

1. Pick one version of every fact. Name, city, founding year, headline numbers. Write them down.
2. Use one total at the top level. Across my network I now use 175,000+ subscribers on every national or company-level page. Each local brand uses its own local count on its own pages. Different scopes, but never two different totals for the same thing.
3. Update every profile and page. Your site, social bios, podcast pages, directories, old guest posts you can still edit.
4. Say the tricky stuff clearly. For a business with Texas in its name that also works with people elsewhere, I landed on "Texas first, not Texas only": keep the Texas city pages, and add a plain line that owners anywhere can work over Zoom. Put it where an out-of-state visitor makes the decision, in the FAQ, and in the summary the site gives AI.
5. Re-run the audit. Same questions, same models, see what changed.

Can it show you're doing well, too?

Yes. About a month after launching WilCo Guide, my local directory, I ran a deep research audit with Gemini. It told me the site was doing shockingly well for being brand new and that Google was already sourcing and citing it. Nobody else knew the site existed yet. That's the kind of signal you want to know about before you spend money promoting something.

Why does it work as a lead magnet?

Every business owner wants to know what AI says about them. A free audit answers a question they already have, shows them a specific gap, and starts a real conversation. That's why I think it's the perfect lead magnet for earned exposure.

One honest caveat: an audit is a snapshot. Models change, answers vary from one ask to the next, and nobody can promise a score turns into recommendations. What it reliably shows you is what to fix.