How does an Answersky scan work?
We ask the questions your buyers ask on 3 AI assistants, several times each, and count how often you're named.
Exactly how do you measure?
Here is the method, exactly as every report states it.
We ask each question 3 times on each assistant through their APIs with web search: ChatGPT (gpt-5.2), Gemini (gemini-3.5-flash), Perplexity (sonar). A one-line setup says the buyer is in the US and asks for companies by name. There's no device location, login or chat history, so the apps can answer differently. Your score is the share of answers that name you. 'Best X' questions count 3×, problem questions 2× and topic questions 1×. Questions that use your name are shown but not scored. Each call is tried up to 3 times. One that still fails is left out of the score, and the report says how many answers it used. Every report lists its questions, and they stay the same month to month so trends compare.
Which questions do you ask?
We ask the questions your buyers ask: "best X" questions, problem questions and topic questions. The questions are fixed for each business, so one month's result compares with the next.
A "best X" question asks for a recommendation, like "best moving company in Portland". A problem question describes what the buyer needs solved. A topic question asks about the subject in general.
Why ask each question 3 times?
Because the same question can get a different answer each time. One site scored 20.4 and 15.3 on the same day, Sep 23, 2026: 20.4 from a scan with 1 run per question and 15.3 from a scan with 3 runs. Asking 3 times and counting every answer gives a steadier number.
Do you keep the answers?
Yes. We keep every raw answer as evidence, and the report quotes them, so you can read what the assistants actually said about you and the businesses they named.
What does the score mean, and what doesn't it mean?
The score is the share of answers that name you, weighted as above, for these questions on these assistants on the date of the scan.
It doesn't measure your traffic or your search rankings, it isn't what every user sees, and it isn't a promise of future results. Scores move by about 5 points up or down between scans, so we count a change only when 2 scans agree.
How is the free Quick Scan different from the Full Report?
The free Quick Scan tells you whether you're named and who is named instead. The Full Report gives you a score for each assistant and the top 5 fixes.
| What you get | Free Quick Scan | Full Report |
|---|---|---|
| Assistants | ChatGPT and Perplexity | ChatGPT, Gemini and Perplexity |
| Runs per question | 1 | 3 |
| Result | Whether you were named | A 0 to 100 score for each assistant |
| Who is named instead | Yes | Yes, with the sites the answers rely on |
| Quotes from the answers | No | Yes |
| Check of your site | No | Yes |
| Fixes | No | The top 5 fixes |
Gemini's cited sites are left out of reports for now, until we resolve how its API returns sources.
Get a free Quick Scan
We ask ChatGPT and Perplexity the questions your buyers ask, and email you whether you were named and who was named instead. Report within 3 business days.
Got it. We'll check your request and email your Quick Scan to within 3 business days.