Does AI recommend your brand?

What happens when you press check

1. 01

   We write the questions, not you

   Spegla generates ten buying questions for your category, in your language. Your brand name never appears in any of them. The result only means something if the assistants bring you up on their own.

2. 02

   Four assistants answer all ten

   Every question goes to ChatGPT, Claude, Gemini and Perplexity through their official APIs. That is forty answers, collected the same way each time, which is what makes two checks comparable.

3. 03

   Every answer is read and scored

   We record whether you were named, how early you appeared, which brands were recommended instead, and which sources the assistant cited. The whole run takes about a minute.

Your category decides how useful the result is

The category is the one thing you choose, and it shapes all ten questions. Write it the way a buyer would describe what they need, not the way your industry describes itself.

A narrow category produces the questions your buyers actually ask. A broad one produces questions nobody asks, and a score that describes a market you are not in.

What you get back

Visibility score

One number from 0 to 100, averaged across the four assistants. It combines how often you are named with how early you appear, because being mentioned last is not the same as being the recommendation.

Mention rate

The share of the ten questions that named you at all. A high mention rate next to a low score means you are present everywhere and first nowhere.

Named instead of you

The brands the assistants recommend in your place, counted across all forty answers. This is usually the most uncomfortable part of the report and the most useful one.

Sources the AI cites

The pages an assistant drew on, where it reports them. Models repeat what their sources say, so these are the pages worth being on.