Skip to content
askvisibly

How askvisibly measures AI visibility

Four steps turn a domain into a score you can track. The formula is public, and so are its limits.

  1. Generate prompts

    We read your site’s title, description and headings and ask a language model to write the questions your buyers would type into an AI assistant. The prompts fall into five intents: category (“best tool for…”), comparison (“A vs B”), alternative (“alternatives to A”), how-to, and brand (“is A good”, “A pricing”).

    Category, alternative and how-to prompts never contain your brand name. That is the test: does the engine bring you up on its own? A free scan uses 10 prompts: 4 category, 2 alternative, 2 how-to and 2 brand. Starter uses 25 and Pro 40, with comparison prompts added when we know your competitors. Paid plans can edit the set.

  2. Ask the engines

    Every prompt goes to every engine in your plan: Perplexity and Claude on Free, plus ChatGPT, Gemini and Google AI Overviews on Starter and Pro. Engines run with web search switched on, so an answer can include live sources and a list of citations. Google AI Overviews are read from Google’s results through a data provider.

    We keep the full answer text and every cited URL for each prompt and engine pair. When Google shows no AI Overview for a prompt, that result counts as not mentioned.

  3. Analyze the answers

    For each answer we record five things.

    • Mentioned: your brand name, one of its aliases or your domain appears in the answer text.
    • Cited: at least one cited source is a page on your domain.
    • Position: your rank among the brand-like names in the answer (bold names, list items and known competitors), starting at 1. Empty when you are not mentioned.
    • Snippet: the sentence around your first mention.
    • Competitors: the other brand names the answer contains.

    Mention and citation are separate on purpose. An engine can name you without linking to you, or link to you without naming you.

    For small teams comparing invoicing tools, the usual shortlist is Brand B, Your Brand and Brand C.

    Illustration, not a real answer. Your Brand is mentioned and ranks second, so its position score is 0.8.

  4. Compute the score

    The Visibility Score is one number from 0 to 100. It weighs how often you are mentioned, how often you are cited and how high you rank when you are mentioned.

    score = round(100 × (0.5 × mention rate
               + 0.3 × citation rate
               + 0.2 × position score))
    • Mention rate: the share of answers that mention you.
    • Citation rate: the share of answers that cite your domain.
    • Position score: the average, over the answers that mention you, of 1 for first place, 0.8 for second place, 0.6 for third place, 0.4 for fourth place, 0.2 for fifth place and 0 from sixth place on.

    Each engine gets its own score from the same formula on its own answers. Share of voice is your mentions divided by your mentions plus your competitors’ mentions.

    Worked example (illustrative numbers, not a real brand)
    Answers
    10 prompts on 2 engines20
    Mention rate
    mentioned in 10 of 200.50
    Citation rate
    cited in 4 of 200.20
    Position score
    average over the answers that mention you0.60

    Visibility Score43

What the score does not tell you

It measures a sample of answers, taken the same way each time.

API answers approximate the consumer apps
We call engines through their APIs. The apps add personalization, location, chat history and their own instructions, so what a person sees can differ from what we record.
Answers vary between runs
The same prompt can produce a different answer an hour later. One scan is a sample, not a ranking. Look at the trend over several scans.
Engines change
Models, search providers and answer formats change without notice. A score can move because an engine changed, not because your site did.
Prompts are our best guess
Generated prompts approximate buyer questions. They are not search volume data. Free scans use the generated set; paid plans can edit it.
Position is an estimate
We rank the brand-like names we can find in the answer text: bold names, list items and known competitors. A list that names tools in no particular order still produces ranks, and a name we miss can shift them.
Mentions are not endorsements
A mention counts even when the answer is critical. We keep the sentence so you can read it, but we do not score tone.

See your own numbers.

A free scan takes about a minute and needs no account. The report is public, so you can share it.

Scan my brand