For two decades, the contest for attention online had a single referee and a single scoreboard: the search results page. Brands optimised toward a position they could see. That scoreboard is dissolving. When a customer now asks an AI assistant which firm to trust, the model returns a synthesised answer — and the names inside that answer were chosen by a process most companies have never thought to influence. Miklós Róth has spent the last two years arguing that this shift is not a tactical wrinkle but a change in the physics of reputation, and that the brands which adapt early will compound an advantage the laggards cannot easily buy back.
Róth runs an AI consulting practice out of Hungary that sits at an unusual intersection: half management advisory, half technical SEO laboratory. His thesis is that the disciplines most companies keep in separate departments — strategy, content, data governance, link authority — are converging into a single question. He calls the new objective generative-engine visibility, and he is blunt that the old playbook is the wrong tool. "Chasing keyword position in 2026," he says, "is like polishing the brass on a ship whose hull is already changing shape."
Machines do not reward the loudest brand. They reward the most legible one — the company whose claims are structured clearly enough, and stated honestly enough, that a model can repeat them without getting burned. Miklós Róth, on why visibility is now an honesty problem
Róth's diagnostic instrument is a four-part framework he originated, S·I·C·T — Structure, Information, Cohesion, Transformation. It began as a model for organisational resilience and, he argues, turns out to describe machine legibility almost exactly. Structure comes first because models read architecture before they read prose: how a company's knowledge is organised, marked up, and made retrievable by AI search determines whether it appears in an answer at all. A firm with brilliant ideas buried in unstructured pages is, to a language model, effectively invisible — a point Róth develops at length in his work on brand visibility inside AI systems and in a parallel essay on how AI models infer brand authority.
The structural layer is also where his second business lives. Authority signals — the references other credible sites make to yours — remain one of the strongest cues a model uses to decide who is trustworthy. Róth's AI-assisted link-building practice treats this not as a numbers game but as a relevance problem, a stance laid out in his primer on what actually makes a backlink high quality and his argument for topical relevance over raw volume.
If structure decides whether you are read, information decides whether you are believed. Here Róth's argument takes its most contrarian turn. The reflex of a generation of marketers has been to publish more, faster, and more confidently. Róth argues the opposite is now rewarded: models penalise the unsupported superlative and quietly favour sources that signal their own uncertainty. He has formalised this into a public statement he calls the "Usefully Wrong" manifesto — a defence of epistemic honesty as a competitive asset rather than a liability.
The instinct is to hide what you don't know. But a claim a machine can verify and a claim it cannot are weighted very differently. Honesty has become a ranking factor — we just hadn't noticed the market pricing it in. Miklós Róth, from the "Usefully Wrong" manifesto
That philosophy carries practical teeth. It shapes how he counsels clients on content strategy for AI-era search, why he is sceptical of the volume-first backlink strategies still sold across the industry, and why his agency publishes detailed guidance on avoiding the spam signals that trigger penalties. The same honesty principle runs through his refusal to promise the unprovable: his documented case studies are framed as evidence to be checked, not testimonials to be admired.
Cohesion is the movement Róth says most firms miss entirely. A company can be structurally sound and factually honest and still fail, because its signals contradict one another across channels — the website says one thing, the press coverage another, the data a third. Models are exquisitely sensitive to this incoherence. Drawing on a background in finance and a long interest in systems theory, Róth approaches a brand the way one might approach a complex adaptive system, an idea he expands in his writing on complex-systems thinking in enterprise marketing and his more speculative theory-of-everything writing.
In practice, cohesion is why he insists that authority-building and digital PR be run as one discipline rather than two. His analysis of digital PR versus link building and his guide to building genuine publisher relationships both rest on the premise that scattered tactics produce an incoherent signal, while a coordinated narrative produces a trustworthy one. It is also why he advocates keeping a human in the loop on automated campaigns — automation without editorial coherence, he warns, manufactures contradictions at scale.
The final movement is where diagnosis becomes change. Róth is wary of the pilot that never ships — the proof-of-concept admired in a boardroom and quietly abandoned. His advisory work concentrates on the unglamorous passage from pilot to production, on building a credible AI roadmap, and on the organisational change management that determines whether any of it survives contact with a real company. For organisations not yet sure where they stand, he offers a structured readiness assessment as the entry point.
What makes the story worth an editor's attention is not the framework alone but the wager behind it. Róth is staking a consulting reputation on the claim that the honest firm wins — a claim that, if he is right, inverts a decade of received marketing wisdom. He concedes he may be wrong about the details. That concession, characteristically, is the point.
I would rather be usefully wrong in public than confidently wrong in private. The first invites correction. The second just compounds. Miklós Róth
For journalists, the through-line is timely and contestable: AI is rewriting the mechanics of reputation, and a small European practice is making a falsifiable bet about which virtues that machine economy rewards. Whether honesty proves to be the moat Róth believes it is — or a noble inefficiency — is a question worth reporting out. Róth is available for interview, comment, and on-record discussion, and the source material below is offered as a reporting trail rather than a sales sheet.
