Two days after we launched AI True Score, we searched Google for “who is AI True Score.”
Good news first: our site was the #1 organic result. Indexing was fast, the rebrand redirects worked, the technical foundation held. Fine.
Then we asked the AI engines themselves who we are — the answer most people now stop at. And here’s the interesting part: they didn’t agree.
Perplexity got us right. Asked the same question, it identified us as a business rather than a person, named our actual products — the AI Readiness Glimpse and the AI True Score Full Report — described who we build them for, and cited our own pages as the source. It even offered to sort us out from any similarly-named person or app, unprompted.
Google’s AI Overview described other companies entirely. On July 16 it told us “AI True Score” was either an AI-powered resume and job-matching platform or a live sports scores app — two different businesses, neither of them us. By the next morning it had changed its mind again: there is no single entity by that name, it said, the term probably refers to the AI-detection scores sold by other vendors. Our site sat directly below that answer, in the #1 organic spot, uncited.
Same business. Same day. One engine knows us; another describes strangers. We build tools that measure exactly this split — and on day two, we were a case study in our own product.
Ranking and being known are two different things
Here’s what that screenshot teaches, and it’s the single most useful thing a business owner can learn about AI search right now.
Google ranked us because ranking is mechanical. We were a two-day-old brand name and Google had us at #1 almost immediately. That doesn’t happen by luck, and it isn’t magic either — it happens because the structure underneath was right: clean page structure AI and Google can read, one canonical address for every page, correctly executed redirects from our former name (so years of history carried over instead of evaporating), a sitemap that tells crawlers exactly what exists, and the same name used the same way on every page. Without that structure, this screenshot doesn’t exist — a renamed two-day-old site would have been invisible for weeks.
That’s the honest lesson of the first half: the mechanical layer works, and it’s the layer your website structure controls. It’s also exactly what the foundation checks in our reports measure. But ranking for your own name is where structure’s job ends — because then we read the AI Overview.
The AI answer didn’t include us because being known is not mechanical. An AI assembling an answer about “AI True Score” looked for corroboration — outside places that confirm the entity exists: who founded it, what it does, what it is NOT. We were two days old. The other companies with similar names had years of app-store listings and third-party pages. The AI did the reasonable thing with the evidence it had. We simply hadn’t given it enough evidence yet.
That’s the gap: your website can be winning at search while some AI engines still don’t know you exist — and different engines can hold completely different beliefs about you at the same time. Perplexity reads the open web fresh and found our pages; Google’s AI Overview leans on established entity records we were too new to have. Neither is broken — they just learn about you differently, at different speeds.
If that’s true for a company whose entire job is AI visibility, it’s absolutely true for a roofing company, a mortgage broker, or a med spa — where one engine recommends you, another names your competitor, and nobody sends you a memo about either.
What we’re doing about ours (the same things your report tells you to do)
We’re not special. We’re running the same playbook our own reports prescribe, in public, so you can watch whether it works:
- Say plainly, on our own site, what we are — and what we are not. Our About page now opens with a definitional sentence an AI can lift verbatim, including “not a resume, career, or interview service” — because that’s who AI kept confusing us with. If AI confuses you with someone, disambiguate on your own site first.
- Give the machines the boring facts. Founder, founding date, legal entity, former name — stated in structured data (the machine-readable layer of the site). Every fact real; nothing invented.
- Build third-party corroboration. Company profiles on real platforms, linked back and forth with the site, so the entity exists somewhere other than its own homepage. This is the slow part — and the part that actually moves AI answers.
- Tell the AI it’s wrong, politely, through the front door. Google’s feedback tools exist. We used them with accurate corrections, not marketing copy.
None of this pays off same-day. Entity recognition moves in weeks and months, not hours — anyone who promises to flip an AI answer this week is guessing. We said that about AI rankings on day one, and it applies to us too.
Why we’re publishing this
Because “we measure whether AI names you” lands differently when we show you our own scorecard, including the ugly part. At launch, ours reads: Perplexity knows us, Google’s AI Overview doesn’t yet. One for two — measured, not guessed.
We’ll update this post as the answer changes. If the playbook works, you’ll see it work. If it’s slower than we’d like, you’ll see that too — that’s the deal.
Meanwhile: if you’ve never checked what AI says when someone asks about your business — that’s the free Glimpse. It takes about a minute, and it’s the same first look we ran on ourselves.