Know what AI says about your brand and change it
We ask the main AI models the questions people ask about brands and keep every answer with its date. You see what they say and what moves it and every figure shows how sure we are.
What we do
AI answers are becoming the first thing people read about a company and nobody tells the company what those answers say. We measure it and we write the plan that puts it right and we put AI to work in the teams that deliver it.
Discovery in search and AI answers
Know where your brand stands at every step people take from the first question to the purchase and what to do about it. We work with marketing teams and with brand and reputation teams and with the analysts who research companies.
How it works →AI brand content strategy
What to publish and fix and in what order so that AI names you and describes you correctly. The plan is built from the pages the models already read and every change is measured.
How it works →AI consulting and transformation
Put AI to work in your team in thirty days and keep what works. We audit the team's week task by task and choose two routines with the team and count what is still in use at day 30.
How it works →AI agents for business
Agents that each do one job and give the team its hours back. Each one is handed over with written instructions so that your team can run it and change it without us.
How it works →Model Research IQ
Our platform. It brings AI answers and search and social and your own surveys together in one record you own and answers questions about your data in a sentence and runs the agents. Access is by invitation.
The platform →Why we do it
Your brand data lives in six places and none of them agrees with the others. We bring it into one record so that a trend can be read across every touchpoint.
What AI models say about you and your site analytics and your search and social data and your brand survey arrive from different tools with different definitions. One record holds them for the same brands and the same measures with one set of definitions so that a change in one shows beside the others and the way your touchpoints interact can be read rather than guessed.
The record is yours. It is a stored database with a layer of meaning over it so that the dashboard reads from it and the AI chat answers from it and an agent can be built on it. The same figure appears everywhere because it is only counted once.
Data you can trust
AI never does the counting. The models answer and the platform counts and every figure shows how many answers sit behind it and how sure we are.
Everything is stored and cleaned and versioned so that a reading can be checked and repeated. How we keep the data honest →
For the people who build or buy or protect a company
Three kinds of people use a reading and each gets something the others do not need.
Build
Brand and communications and insight leaders get a reading that says what AI tells their customers and which facts are incorrect and which pages the incorrect facts come from. That last part is the difference because a list of errors is a worry while a list of pages is a plan.
Buy
Investors and their advisers get an AI exposure reading inside a diligence timetable at a fixed price. It shows whether AI recommends the target or its rivals and what AI holds against the target and whether the answers are stable or moving because that is the question a buyer actually has.
Protect
Reputation advisers get what AI says about a client and its public leaders with every claim traced to its sources and then a dated record that shows whether a fix held. A correction that holds for a while is not the same as one that holds for good. Only a record can tell them apart.
Up and running in minutes
Setting up takes a few clicks and the first readings land the same day. The platform then runs itself on the schedule you choose.
- Name your brand and its rivals and the market you sell in.
- Connect Search Console and drop in the files you already have. A file is recognised from its columns and lands in the right place.
- The first AI readings land within minutes and the record starts from that day.
- Readings repeat on the schedule you choose and the record keeps itself. Nothing needs maintaining by hand.
Questions people ask us
What does AI visibility actually mean?
The share of AI answers that name your company when someone asks about your category. It is the first measure but it is not the useful one on its own. The finding that moves a decision is being named for the wrong thing or being named and then not recommended. Our readings go on to those.
How is a Model Research reading different from an SEO tool that tracks AI?
A tracking tool counts names. We ask the questions people ask on the way to a decision and read every answer for what it says and why. We keep the answers with their date and put the result beside search and social and what people tell surveys. The record and the method are the product and not the dashboard.
Does Model Research only read large companies?
No. A reading only needs a company that people ask AI about and that is most companies with customers. The B2B readings are often the most useful because a B2B company is frequently cited in an answer without being named.
Who sees a company's reading?
You and anyone you choose to share it with. We never publish a client's reading and we do not publish sector findings. The record from our own panel of public companies is available to clients and partner firms.
A sample reading shows one company and its rivals with the answers it was built on. To see one our details are on the contact page.