Readings of what AI says about your company

Brand and reputation reading

Know what AI says about your company before your customers, investors and journalists do.

Why you need it

AI answers are becoming the first thing people read about a company. A model that holds an old fact or a past controversy repeats it to everyone who asks, and nobody tells you. This reading does.

What it tells you

What the models say about you and your rivals, which facts are incorrect, which objections the models raise without being asked, and which pages the incorrect facts come from. Every figure is counted, with the range across repeated runs.

What you do with it

Correct the facts at their source and fix the pages that carry them. Lead with the strengths the models already see.

The difference

We trace every incorrect fact to the pages the model drew on. A list of errors is a worry. A list of pages is a plan your team can start on Monday.

Head-to-head preference

When a person asks "this company or that one", the answer is usually "it depends". We read which way it leans, and what it leans on.

Why you need it

People ask AI to compare, and the models rarely pick a winner outright. "It depends" is the normal answer. That matters, because "it depends on what you need" is where the model says what each company is best for. A brand wants two things at once: to be preferred overall, and to be known for the right thing when the answer hedges.

What it tells you

For every comparison question, three readings. Overall preference: how often the answer favours you, hedges, or favours the rival. Point-of-difference preference: when it hedges, which company it prefers for which need. And what that point of difference is, in the models' own words, so you can see whether you are known for the thing you want to be known for.

What you do with it

Decide which territories to lead in and win, and make the case for them in the places the models read. Then watch "it depends" turn into a preference over the following readings.

The difference

Most tools count whether you were named. We read the verdict of every answer and the reason behind it, so you know not just whether you are winning but what you are winning for.

Cited but not named

Your page is doing the work and someone else is getting the credit. We find every answer where that happens, and who is named instead.

Why you need it

An AI answer can cite a company's own page as its source and still not name the company. The page supplied the facts, and the answer either names nobody or names a rival. This is common, and it is most common for B2B companies, whose pages explain a subject well and whose names the models have less reason to repeat.

What it tells you

How often your pages are cited against how often you are named, for you and every rival, on one chart. Then, for the answers where you were cited and not named, who was named instead, what they were named for, and the gap between their page and yours on that subject.

What you do with it

The challenge is to be mentioned. The gap tells your content team what the named rival's page says that yours does not, which is usually a clear claim about who the company is and what it is for, placed where the model reads it.

The difference

We join citations to mentions answer by answer, so the finding is a list of real answers and the pages behind them rather than a ratio. That is the list a content team can act on.

AI against people

See where AI's view of your brand is ahead of what people think and where it is behind.

Why you need it

AI models and your customers can hold different views of the same brand, and where they differ one of them will move the other. A brand tracker tells you what people think. This tells you what AI thinks, on the same measures, so the two can be compared.

What it tells you

Your brand scored on the same measures by the AI models and by people, side by side. The people side is the survey you already run or a public one. For UK banks that is the CMA service quality survey.

What you do with it

Put effort where AI is behind the people and protect where it is ahead. Report both beside your tracker at each reading.

The difference

The brand questions are yours to set, from your tracker, a public survey or the claims you most want checked. One rule holds: every company in the category is asked the same words.

Find out more about our AI survey framework →

Campaign study

See whether a campaign reached the AI answer and how long it took.

Why you need it

A message that lands in the press and online may never reach an AI answer, and when it does it arrives weeks later. Nobody measures that today. This study does, week by week.

What it tells you

Which messages reached AI answers, how long they took, which never arrived, and what else changed in the same weeks.

What you do with it

Know which channels feed the models and plan the next campaign to reach them.

The difference

The campaign is followed through press, online conversation and AI answers in one record, so the lag between them is measured rather than assumed.