The Mediator Model

1994: the newsroom. 2004: the search engine. 2024: the model.

Someone looked you up this week. Not on Google. They asked an AI model, and the model decided what gets said about you before you could say a word.

There has always been an intermediary between a person and their audience. The newsroom decided who made the page and who did not. It had a phone number. You could call the editor, argue, charm, push. Then the search engine decided which page sat on top. It had rules you could study, and an entire industry learned to play them. Both gatekeepers had one thing in common: you could see them, and you could work them.

The new mediator is different in kind, not in degree. The model reads you before any human does.

I contrast two forms of mediation: a results list people choose from, and a written answer that has already synthesised sources. In practice they overlap. Modern search engines also provide direct answers and AI summaries; a link list is ranked rather than neutral.

Selection becomes judgement.

My image is a map for the index and a voice describing a route for the answer system. Both select. The mediator model is my framework for understanding that mediation, not a complete technical account of every search system.

The question therefore goes beyond ranking well: how does this mediator select information, and which sources make it into the answer?

The mediator decides who gets cited. Reach becomes the second question.

Who are the new mediators?

In briefAI models and answer systems such as ChatGPT, Perplexity, Google AI Overviews and Claude select and summarise information. In my model, they take on a mediating role between a person and their audience.

The new mediators are AI models and the systems that deliver them: ChatGPT, Perplexity, Google AI Overviews, Claude. They stand between a person and their audience, in the place the newsroom held first and the search engine after it. They read, weigh and summarise before any human decides.

What does the new mediator select for?

In briefI distinguish findability, consistency and credible corroboration by other sources. Repetition is not proof of truth, and the retrieval layer works differently from trained knowledge.

Impressions leave no evidence. Mentions do.

Two things matter: in how many places you are named, and how consistently. In my view, one mention in an established trade publication carries more weight in how models place you in context than a hundred impressions. This is an observation, not a measured weighting. My underlying assumption is that such mentions can build a more credible picture for the model.

I distinguish, in simplified terms, trained knowledge from information a system retrieves for a particular request. Whether it searches the web and which sources it uses depend on the product and query. Further training can change trained knowledge. One website may appear in multiple datasets; a broader presence guarantees neither inclusion nor citation.

My mediator model centres on three questions: can relevant material be found, does it form a consistent picture, and which credible sources support it? I use these questions to assess public visibility. They are not a complete account of how AI systems select information.

Loudness is not on the list. It still buys something: presence. And presence feeds the entity. What it never buys is authority. Loud and empty gets you noise, at worst a distorted picture.

Loudness gets you mentioned. Credibility keeps you there.

Which leaves the hardest question. Can you get cited for nonsense? A fluke in the retrieval layer is possible. False claims can also persist when they are repeated and corroborated. Corroboration rewards the repeatable, not automatically the true. Falsehoods, hype and campaigns get repeated widely too.

And yes, money and connections still count. The newsroom could be wined and dined, the ranking could be bought. Whoever knows the right people today buys traces the model reads as evidence. The morality of the system has not changed, the route has. You used to buy the mediator’s verdict directly. Now, at best, you buy the evidence. Bought evidence scales worse. Earned mentions last longer.

I have lived this myself. More on that in a moment.

Ultimately, I put my trust in credible corroboration. I expect journalists, experts and institutions to examine claims and challenge nonsense. On a contested, vetted topic, I therefore put my money on substance. In an unvetted niche, false pictures can settle too.

Reputation is the currency of AI. It pays out in mentions, not clicks.

What is an index echo?

In briefThis is my term for an old, well-documented picture of a person that continues to echo in AI answers after the person has changed. My example shows why a new direction also needs consistent, credible evidence.

Since 2010, I have worked in communications and marketing, building other people’s visibility. When I repositioned myself, the old picture refused to move. As late as May 2026, ChatGPT was still introducing me as a sustainability blogger. Four legacy URLs, one shop subdomain, four social profiles: the old trail owned the index. The model was not wrong. The old picture was simply better documented than the new one.

Well documented. Consistent across every surface. And outdated.

I call this the index echo.

An index echo is the old picture that continues to echo in the model long after the person has left it behind.

The index was never neutral. In the SEO era it decided which page ranked. Today the index helps decide who an AI system names whenever that system retrieves sources. The only new thing is that it now speaks into the answer instead of the results list. And sheer volume of evidence does not dissolve an echo and can even reinforce it. Only one thing helps: new, credible traces. Consistent. In vetted places. Until the new picture outweighs the old.

What does this mean for public visibility?

In briefA distinct position needs documentation that allows it to be found and placed in context. For me, credible corroboration becomes as necessary as self-presentation; reach alone does not address that task.

Whoever is not mediator-readable does not appear in the answer. If you do not exist in the mediator’s head as a distinct, documented source, you are not in the response. Not censored, simply never retrieved.

The dangerous part: in the age of AI, what decides is not your worth but your findability.

Public identity is built through documented validation, not through self-presentation alone.

You can buy visibility. Citations you have to earn.

Who reads you first: a person or AI?

Sources & image credits

Own foundation

• Related essays: Personal Brands between Attention and Findability (30 May 2026), on the gatekeeper’s return. The Human as an Entity (2 July 2026), on the node the echo hangs from.
• The comparison that one mention in an established trade publication outweighs a hundred impressions for thematic placement is my own assessment from observation, not a measured figure.

Terms

• Mediator model, mediator effect, mediator-readable and index echo are my terms in the sense described here. This essay is where I first use them in my writing. The mediator effect is the shift in one line: the mediator decides who gets cited, and reach becomes the second question. Index echo describes my observation of an older public image continuing to echo.

Systems mentioned

• Perplexity, ChatGPT with search (OpenAI), AI Overviews (Google), Claude (Anthropic), Gemini (Google). As of July 2026.

Technical context

• Google Search Central: Optimizing your website for generative AI features on Google Search. On the continuing role of SEO, the search index and ranking in Google’s AI answers.
• Anthropic: Claude can now search the web (20 March 2025, updated 27 May 2025). On retrieving current web information and citing sources in answers.

Image credits

• Header: concept and composition by Elfie Schürfeld-Todor. A stylised depiction of the mediator model, created with ChatGPT from an original brief; typography added afterwards. Rose and sage variant of the original Bauhaus composition.

Method

• This essay was written in dialogue between me and an AI model. The arguments, the stance and the editorial choices are mine. The model was sparring partner and co-author. It helped sharpen my arguments, challenged me, suggested sources and produced drafts for me to rework. Quotes and references are cross-checked. Augmentation, not automation.

Facts and sources reviewed and clarified on 11 October 2026. The conclusions, analogies and judgements in this essay are my interpretation. Source links were added and the framing of the mediator model and the statements about ranking and impressions were clarified on 14 October 2026. The original publication date is retained.