Who is behind these numbers
One person, one measurement, and a published methodology. Who is behind the numbers on this site and why they are reported the way they are.
Standing is very small
Said plainly, because you would find out anyway and because it bears on how much to trust the method.
There is no large team behind the “we” you will read elsewhere on this site. It is the company's voice, not a headcount claim. The same hands write the measurement code, run the scans and answer the emails.
That is worth knowing before you buy anything. It means you will talk to somebody who wrote the method rather than to an account manager, and it means the delivery windows on the services page are small-team windows. They are published, and they are the ones we hold to.
Why the methodology is the product
Every tool in this category prints a score. The argument is about what a score is allowed to mean.
Ask an AI engine the same buying question twice and the answer changes. That is not a bug in the engines, it is what they are: non-deterministic systems sampling from a distribution. Every tool I looked at asks once, prints a number out of 100, and moves on. When that number moves from 34 to 38, nobody can tell the customer whether it was their work or the weather.
So Standing asks each question five times on each of five engines and reports the score with the interval it is accurate to. A change only counts as a change when it is larger than the amount that domain's own number drifts between runs. The question basket is fixed and published, so anyone can re-run any number on this site and get the same answer.
That is the whole product, and it is deliberately unflattering. It means a re-scan sometimes reports that work a customer paid for has not landed yet, which is a bad sales technique and the only honest thing to do.
What we will not do
The commitments that cost money, listed because they are the ones worth checking.
Guarantee a citation, a mention or a ranking
Nobody controls what these engines say. Services are priced on the work delivered, never on an outcome we do not control, and anyone promising a guaranteed AI mention is selling something they cannot deliver.
Publish a score without its interval
A number with no error bar cannot be acted on. If a scan degrades far enough that no interval can be computed, the result is withheld rather than published bare.
Claim a training-crawler block stops you being cited
Blocking GPTBot or ClaudeBot keeps your content out of future models. It does not remove you from an answer. Most tools in this category report the two as the same thing, and that error is why several of these pages exist.
Sell reviews, or gate them
Review work here means making it easy for real customers to leave real reviews. No incentivised reviews, no filtering by sentiment before asking.
Score something we could not establish
A refused fetch is our request being turned away, not a finding about your site. Absence of evidence gets reported as unknown.
Published work
Everything here is live and checkable, which is the only kind of credential this field has.
- AI Crawler Access Index
74,157 domains swept for AI crawler access, published in full.
- Y Combinator AI readiness census
Every company in the YC directory, checked and published with the method.
- The measurement methodology
The question basket, the sampling depth and the interval maths.
- Five free tools
Crawler access, page readability, structured data, llms.txt and bot verification.
How Standing describes itself
The same two sentences used in our schema, our listings and every profile.
Standing measures whether ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews recommend a business when buyers ask. Every score is published with the confidence interval it is accurate to, so a real change can be told apart from the noise these engines produce on their own.
Check it yourself
The fastest way to judge any of this is to run the free check on your own domain and see whether the report tells you something you did not know, including where it says it cannot tell you anything.