How AskAiRank works

Everything in AskAiRank comes from one loop that runs every day. Understanding it explains why numbers move, why they sometimes do not, and what you can actually control.

1. Your prompts go out

A prompt is a question a buyer would type, not a keyword. “Best AI writing tools for marketing teams” is a prompt. “ai writing” is a keyword, and no one asks an assistant that way.

Once a day, every active prompt is sent to every AI engine you track. One prompt on four engines is four runs. Prompts are sent as written, with no hints about your brand: the moment we told an engine what we wanted to hear, the answer would stop being evidence.

2. Answers come back and get read

Each answer is stored in full and then parsed for four things:

• Was your brand mentioned? Name variants and common misspellings count.
• Where in the answer? Position 1 is the first brand named, which is what “Top 3” and “Average Position” are built on.
• Which other brands were named? These become your competitor list; you never have to type it yourself.
• Which sources were cited? Links the engine leaned on, which become your Citations.

Sentiment is measured for your own brand, on the sentence that mentions you. It is not measured per competitor, because the engines do not give us anything reliable to read there.

3. Numbers are recomputed

Runs are aggregated per day, per engine and per prompt. The dashboard reads those aggregates, which is why a fresh account shows nothing until the first poll has finished, and why one bad day barely moves a 30 day average.

4. You act on it

The rest of the product is about the “so what”:

• Competitors shows who is being recommended instead of you.
• Citations shows which sites the engines trust in your category, which is the list of places worth being mentioned on.
• Site Scan checks whether AI crawlers can read your site at all.
• The Playbook turns all of that into an ordered checklist.
• Blog Calendar and Listings are where the work actually happens: your own content, and your presence on the sites the engines quote.

What you control, and what you do not

You control the prompts, the engines, the markets, and the work you do off the back of the data. You do not control the engines themselves: models are updated, answers vary between identical runs, and a brand can drop out of an answer for a week and come back with no change on your side. That variance is exactly why the product measures shares over many runs instead of celebrating a single good answer.

Related: Prompts, AI engines and models, What each metric means.