The overarching idea
The circle: what you say, what you do, what others say, and what AI learns from it.
The model
The trueroot model
You remain the source of information about you. Others respond to you.
Four steps
AI is not the cause. It is the result of what people say about you.
The occasion is "becoming more visible in AI". The cause is making sure people talk positively about you and your services or products. AI machines only summarise what people say.
So it is broader than copywriting. It is not only about what you say, but about what you do, then what others say about you, and what AI takes from that. That is a circle, and a circle does not finish: an ongoing process, which is why the execution is a retainer. Trueroot's data analyses make that process possible.
The circle starts with telling what you do. Many companies already do a lot but do not tell it: they use hollow marketing phrases, without proof. That comes first. Changing what you do is a fine philosophy, but in practice you do not change that just like that.
The root
You remain the source.
That is why the model is simply called trueroot. You remain the source of information about you, and others respond to you. That even goes for negative reviews on external sites: they respond to what you said or did.
So the work does not start at the AI machines, and not at the external sites either. It starts with you telling what you do, with proof.
The old insights
PR has been saying this for a long time.
PR is different from trueroot, but the way of working is comparable. The first three steps of the circle are close to how the Chartered Institute of Public Relations describes the field:
Public Relations is about reputation - the result of what you do, what you say and what others say about you.
That it is not only about what you say goes back further. One of the Page Principles, based on the writings and speeches of Arthur W. Page, is "Prove it with action". Its wording:
Public perception of an enterprise is determined 90 percent by what it does and 10 percent by what it says.
Those percentages are rhetoric, not a measurement. The point is the order of importance.
That it does not finish is not new either. Mary Jo Hatch and Majken Schultz describe organisational identity as "an ongoing process that behaves like a conversation" between the organisation's self and the images held by key stakeholders. And a PR textbook on John Marston's RACE model (research, action planning, communication, evaluation) says: "Evaluation and measurement should not take place only at the end of your efforts."
The new insights
What is new is the fourth step.
AI machines summarise what people say. A study from the University of Toronto (Chen and others, 2025, a preprint that has not been peer reviewed) reports "a systematic and overwhelming bias towards Earned media" in AI search. In its figures for cars in the United States, a web-enabled GPT drew 81.9% of its sources from earned media, against 45.1% for Google.
And it keeps moving. The web is constantly moving, and language models keep citing other sources. So we have to repeat.
In practice
trueroot is about two ideas.
The tactical idea
One root, and proof for every claim: a True Root that says who you are, and an evidence architecture around it.
What we do not know
Each source covers one link.
We found no research that shows the whole chain, from what you do to what AI says. The circle as a whole, with AI as its fourth step, is our own model, building on PR. The Toronto figures cover a web-enabled GPT, three sectors, and the United States and Canada.