AI visibility

What it measures

People increasingly ask an assistant rather than searching. AI visibility asks the major assistants the questions your customers would ask — in your market and your language — and records whether you are mentioned, how, and who is mentioned instead.

Two scores:

• Visibility — how often you come up at all.
• Reputation — how you are characterised when you do.

They move independently, and the gap is informative: being mentioned constantly but unfavourably is a different problem from not being mentioned.

The questions

Questions are generated per industry you select, phrased naturally in the language of the market being scanned — a Swedish customer does not ask in translated English, and asking in translated English would measure the wrong thing. You can add your own questions if there is something specific you need to know.

Why it asks before running

Each question is asked of each assistant, so the cost of a scan is the product of questions and engines. The app shows the number of calls a configuration will make before you run it, and will not start a scan that exceeds your plan's cap. Widening the industry list or adding questions raises that number quickly — the readout updates as you change the configuration so it is never a surprise.

Reading the results

Every answer is stored in full. When a score moves, you can read the actual text that produced it rather than trusting a number. Competitor mentions are extracted from the same answers, so "who do assistants suggest instead of us" is directly answerable.

Caveats worth holding onto

• Assistants are non-deterministic. The same question can be answered differently. Treat single-scan movements as weak evidence and trends as strong.
• Coverage differs by engine. Not being mentioned by one assistant while being prominent in another is common and does not indicate an error.
• Models change underneath you. A shift across every question at once usually means the model changed, not that your position did.