Frequently asked questions
Straight answers about what we measure, how we measure it, and what you get.
What exactly does Aivosity measure?
Whether, how often and how favourably AI assistants (ChatGPT, Gemini, Claude and Perplexity) mention your business when a potential customer asks for recommendations in your industry and location. We also record which competitors they recommend, what they get wrong about you, which sources they cite, and how ready your website is to be read by AI systems.
Where do the 30 questions come from?
We generate them from your business type, industry and location, then read through them ourselves to check they sound like something a person would actually type. They cover the whole buying journey and never include your business name. You can see the set and change it before we run anything.
Which AI models are tested?
ChatGPT (OpenAI), Gemini (Google), Claude (Anthropic) and Perplexity, through their official APIs. Where a model offers web-grounded answers, we test the grounded mode because that is what buyers see. Additional models are available on agency plans.
AI answers change every time. How is a single report meaningful?
That is exactly why we ask each question three times per model and report the total: 360 answers per report. Mention rate is the share of all those runs that name you, not the outcome of one lucky query. Monitor plans keep the same question set every month so the comparison actually means something.
What counts as a factual error?
Any claim a model makes about your business that does not match the facts you give us or what your own website states: hours, prices, services, locations, staff, certifications, awards. Each error is quoted verbatim with the model and question that produced it, and grouped by severity.
What is the website readiness score?
A crawl-based assessment of how easily AI systems can read and trust your site. It looks at whether your pages clearly state what you do, for whom and where; whether structured data is present and correct; whether key facts are trapped in images or scripts; crawlability; and consistency with your third-party listings.
Can you make ChatGPT recommend us?
Nobody can guarantee what a model will say, and we would not trust anyone who claimed otherwise. What we can do is show you why you are or are not being recommended today, and which sources and pages are behind it. Changes usually show up over one to three monthly runs once those sources are corrected, but that depends on your market.
How long does a report take?
Most Snapshot reports are delivered within five working days of receiving your business details. Monitoring reports arrive on the same day each month.
Do you work outside Spain?
Yes. Nothing in the method is tied to a country — we work with businesses in Spain and across Europe, and the location can be a city, a region, a country or simply "online". Question sets come in English or Spanish, and we can do other languages on request.
Do you need access to our website or accounts?
No. Everything comes from public information plus the facts you give us on the order form. We never ask for logins, and our crawler only reads pages that are already open to anyone, respecting robots directives.
Is my data shared with the AI companies?
The questions we send to the models contain your industry and location, never your business name or the facts you gave us — those stay on our side and are only used to check what came back. Our privacy policy sets out exactly what we hold and for how long.
Are you affiliated with OpenAI, Google, Anthropic or Perplexity?
No. Aivosity is an independent company with no commercial relationship to any of them. We pay for API access like any other customer, and the model names are trademarks of their respective owners.
Still have a question? Contact us and we will answer within one working day.
Get your first report.
One business, one location, four models, 30 questions.