Readiness doesn't guarantee AI visibility: what 12 real audits show
If you want technical AI-search readiness to be the whole answer, the data doesn't back that up. Across the 12 real third-party audits in our corpus, readiness and AI-answer visibility move together, but only weakly. Readiness looks necessary. It is nowhere close to sufficient.
This is an early, honestly small sample. Twelve businesses is not a lot of data, and we're saying that up front rather than burying it: a correlation computed on 12 points has a wide confidence interval, and you should discount the number below accordingly. We'd rather publish it small and honest than wait and round the corners later.
What we measured
Every business in this sample got the same two scores, from the same audit method we run for clients:
- AI-answer visibility (0–100): whether the business was named, and how, when we asked real buyer questions to ChatGPT, Claude, Gemini, and Perplexity.
- AI-readiness (0–100): a technical score covering structured data (schema markup), question-answering content, and entity clarity, independent of what the AI engines actually returned.
The 12 businesses span retail, automotive, dental, B2B software, home services, consumer goods, and financial services. No business, agency, or domain is named anywhere in this piece; the numbers below are aggregates only, computed once across the full sample.
Does higher readiness mean higher visibility?
Weakly, yes, not reliably. Readiness and visibility scores across the 12 audited businesses are positively correlated at Pearson r ≈ 0.60. That's a real relationship, and it's also a loose one: on a sample this size, the honest read is "readiness moves in the same direction as visibility more often than not," not "readiness predicts visibility."
| Metric | Mean | Median | Range |
|---|---|---|---|
| Visibility score (0–100) | 31.7 | 28.5 | 0–68 |
| Readiness score (0–100) | 48.6 | 47.5 | 25–72 |
n=12. We're not rounding that into "our data shows" language stronger than 12 points support.
Readiness without visibility is a real pattern, not the norm
We've hit this in our own audit of ncited.co: our /aeo page went from 83 to 100 out of 100 on readiness, and visibility across the four engines stayed at 0. One business in our third-party sample shows the same shape, readiness in the high range with visibility at zero. It's a real pattern and worth watching for. It is also the outlier in this sample, not the modal case; most of the 12 businesses show some positive relationship between the two scores. "Readiness doesn't matter" is a stronger claim than the data supports. "Readiness alone doesn't guarantee visibility" is the claim that survives.
Most businesses are absent from their own buyer questions
We logged 36 query/answer pairs across the sample (real buyer questions, run against all four engines). The outcome split:
| Outcome | Count | Share |
|---|---|---|
| Absent | 19 | 53% |
| Mentioned (named, no source link to the business) | 9 | 25% |
| Cited (named, with a source link to the business's own page) | 8 | 22% |
| Present as the lead answer | 0 | 0% |
Nobody in this sample owns an unbranded buyer question outright. The best outcome any business achieved was being cited alongside competitors, not named as the clear answer.
What this sample doesn't tell us yet
We track which third-party sources (review sites, forums, industry publications) mention each business, but that field isn't standardized across audits yet, some records log a single named source and others log several under one string, so it can't be cleanly cross-tabulated today. That's a data-quality problem to fix before we publish anything about which authority sources correlate with visibility, not a conclusion we're ready to draw from this sample.
n=12 will grow as more audits run. We'll republish this with a larger sample rather than treat today's numbers as settled.
What counts as a business being "visible" in an AI answer?
We track three outcomes: mentioned (named, no link to the business's own site), cited (named, with a source link to the business's own page), and present as the lead answer. In this sample, 22% of query/answer pairs were cited, 25% were mentioned, and none reached lead-answer status.
How is the readiness score different from the visibility score?
Readiness is a technical check, structured data, question-answering content, entity clarity, that doesn't depend on what any AI engine returns. Visibility is the observed outcome: whether ChatGPT, Claude, Gemini, or Perplexity actually named the business on a real buyer question. This dataset exists because those two numbers turned out not to move together as tightly as you'd expect.
Is this dataset going to grow?
Yes. n=12 is the count as of this audit corpus today. We'll update these numbers as more third-party audits accumulate, and we'll flag it clearly if the correlation or the absence rate moves once the sample is bigger.
Want your own numbers instead of the aggregate? Run a free AEO audit and see your readiness and visibility scores side by side. If you want to understand how the audit itself works, see our AEO audit.