AI calibration

Buyers now ask ChatGPT, Google, Gemini, Perplexity, Claude and other AI engines which supplier to use, which car to buy, which part fits and which garage to trust. The answers sound confident whether they are right or wrong. Some are built from a live search of the web. Others come from what the model absorbed in training, sometimes months or years earlier. From the outside you can’t tell which.

The risk is that an old directory listing, a discontinued product page or another market’s specification sheet can reach a buyer as fact. It may even carry a reference link that makes it look checked.

We don’t sell AI placement

Nobody can guarantee your business will appear in an AI answer. The engines disagree with each other, and the same engine can give a different answer to the same question asked twice. In one peer-reviewed study of 11,500 queries, Gemini’s sources for an identical query overlapped by less than half between two runs.¹ Anyone promising a ranking, or a percentage uplift in “AI visibility,” is selling something they can’t control.

What you can control is the information these systems draw on. That is where we work.

What we do

We find out what AI is saying about you. We put the questions your buyers actually ask, not just the ones that flatter you, to the AI services they use. You see exactly what comes back.

We work out why. Every wrong answer starts somewhere. We find where, whether that is your own website, a third-party source or something the AI has carried over from the past.

We fix what can be fixed. Any instrument drifts over time, and AI answers drift faster than most. A good result today proves little about next month, so we recalibrate regularly and report honestly, including what hasn’t moved.

What you get

Each check produces a Calibration Certificate. It is a dated record of what AI says about you and your competitors, which errors we found, where they came from and what has been done about them. Over time it shows whether things are getting better, measured by enquiries that suit your business rather than raw mention counts.

Start with a Calibration Check

We’ll test what the main AI services say about you and your competitors today, show you every answer and tell you which errors we can fix. 
 

^{1 * Grossman et al. (2026), How Generative AI Disrupts Search, SIGIR 2026.}