The new mediator in the age of AI search
Verona Feldbusch was a brand before the word became normal for people. The FAZ called her “a brand in her own right”. I wrote my bachelor’s thesis on precisely this question: can the concept of a brand be applied to a human being? The answer was yes, on one condition. A person needs a mediator who makes them visible and presents them in a certain light. Back then, that mediator was the press.
One sentence stood in bold at the end of my thesis. Products are marketed passively. People actively participate in their own marketing. A second finding was less comfortable: image and media play the most important roles in building a personal brand. Performance is secondary. A person, I wrote at the time, can only position themselves through media coverage.
Ten years later, the same question has a new addressee.
Three waves. One gatekeeper
First wave: traditional media. Verona, too, had to stage her public persona; she was no passive figure. But recognition came through TV, radio and print, and that was where the gatekeepers sat. An editorial team decided whether you got on the programme. A casting decided whether you fitted as a presenter or an actress. Without that “yes”, no audience.
Second wave: social media. The “yes” disappears. Anyone can pick up a smartphone and film themselves exercising, cooking, losing weight or travelling. A quick edit in CapCut, and off you go. It sounds easier than it is. It takes practice and planning.
Third wave: the influencer. The hobby becomes serious. An editorial calendar, staged footage, every second calculated for reach and recognition. The same work as at a broadcaster, only without the broadcaster. Nobody casts you. You cast yourself. The tools were open to everyone. They still are today.
The influencer is not the opposite of the old media brand.
It is that brand without an editorial team.
The subtle distinction
Building an audience is work. And many influencers have substance: expertise, a community they have built, educational content, entrepreneurship. The point is not that there is nothing there. The point is the form in which it exists.
The currency of attention is not necessarily the currency of findability.
Self-presentation lives through the performance, the moment, the watching. Much of it exists in what is seen, not in a document. And what is not documented is hard for a model to pick up.
This is exactly where my old finding from 2016 sits. Visibility ran through image; performance was secondary. The influencer perfected this by removing the press from the equation.
And now comes AI
Not as a fourth wave of the same movement. As a reversal.
The press was the mediator. The influencer abolished it. AI builds a new one.
The gatekeeper returns. But it cannot watch. It reads. Today, people use AI to compare products and choose service providers. The model does not draw on their performance, but on what is tangible, documented and findable about them. It weighs two things: what you leave behind yourself, and who confirms you.
Anyone who spent a decade building pure image is suddenly exposed. The stage was open for ten years. The index is not. A performance cannot be crawled. The trace it leaves can. Where that trace is thin, there is little in the index. Or nothing.
Self-image equals public image
One condition stood above everything else in my thesis: a brand is strong when self-image and public image align. Brand philosophy and brand image. What you say about yourself, and what others think of you.
The self-image is the tangible, verifiable substance you make findable yourself. With a position clear enough for a model to pick up. The public image is how the web speaks about you. A backlink, a mention, a citation in a source the model trusts. The press is not dead. It has become a signal. The word of mouth from my thesis is the link graph.
You only get cited and found when both halves fit together. Substance you make findable, plus evidence others provide. And here is the uncomfortable qualification: a model does not judge quality. It weighs patterns, frequency, the available sources and connections. It can overlook someone with great substance if that substance is poorly documented. And it can elevate someone with little to show if enough has been written about them.
AI does not reward the greatest substance. It rewards the documented.
The model was right
In May I looked myself up in a model. ChatGPT listed me as a sustainability blogger.
That was not wrong. I spent years working freelance for sustainable and organic brands, writing recipes, blogging about the subject and getting involved. I still live those values today, with many lessons learned along the way. The model was not off the mark.
It was right. About an Elfie from yesterday.
My self-image had moved on. My public image in the index stood still. The old trail was crawlable, documented and maintained for years. The new position was young and thinly substantiated. Not an image problem. A timing problem. Self-image and public image had drifted apart, and the model believed the majority of the old data.
As an independent communications consultant, I developed and implemented communication strategies for my clients while neglecting my own.
Brand communication and brand strategy need to reposition themselves
The answer is not to erase the past. Nor to avoid the new medium.
Brand communication and brand strategy need to reposition themselves because traditional media no longer decide, and neither does the open stage alone. Anyone who wants to remain visible brings the new chapter up to date. Substance that can be crawled. Evidence provided by others.
The goal: bring self-image and public image back into alignment, in a medium that reads instead of watching.
With technology and science, not against them. The AI that put me in the wrong category is the same AI I use to correct it.
The human was never the passive product. That was true in 2016, it is true for the influencer, and it is true for GEO. What changes is the mediator and what it rewards.
The question is no longer just how people see you. It is which version of you lives in the index.
Sources, image credits and method
Sources
Own foundation
- Schürfeld-Todor, E. (2016). Der Mensch als Marke. Eine Untersuchung zur Übertragbarkeit des Markenbegriffs auf den Menschen (The human as brand name. A survey on the transferability of brand name to humans). Bachelor’s thesis.
Brand theory
- Herbst, D. (ed.) (2003). Der Mensch als Marke. Konzepte, Beispiele, Experteninterviews. 1st edition, Göttingen. Foundation for the identity-oriented reading and the thesis that image and media sustain a personal brand.
- Mellerowicz, K. (1963). Markenartikel. Die ökonomischen Gesetze ihrer Preisbildung und Preisbindung. 2nd edition, Munich/Berlin. A classic, feature-based definition of a brand that cannot be applied to people one-to-one.
- Hannemann, M. (faz.net, 2001). “Verona Pooth. Eine Marke für sich.”
Further reading
- Related insight: Der Funnel ist tot. Was übrig bleibt. On the shift from clicks to citations and the GEO logic behind it.
Image credits
- Header (three doors, “Wer erklärt die Welt?”): concept and composition: Elfie Schürfeld-Todor. Brand palette of violet and yellow on black. AI-assisted image generation based on my own briefing, with subsequent editing. The illustration is stylized, not photorealistic, and therefore does not fall under the labelling requirement of EU AI Act Art. 50.
- Portrait in the text: concept and composition: Elfie Schürfeld-Todor. One half of the face is a photograph; the other dissolves into lines of index data. AI-assisted editing of my own photograph. Photorealistic composition, therefore labelled. Image: AI-edited composite with ChatGPT, based on an own photo. (Labelled ahead of EU AI Act Art. 50.)
Method
- This insight was written in dialogue between me and an AI model. The starting point was my own bachelor’s thesis from 2016. The theses, the position and the selection are mine. The model was a sparring partner and co-author. It sharpened, contradicted, suggested sources and delivered drafts I reworked. Quotes and references are cross-checked. Augmentation, not automation.