Reconnaissance
Why Online Reputation Is Becoming the Critical Infrastructure for AI Visibility
About the Authors
Christian Scherg founded REVOLVERMÄNNER GmbH in 2007 — at a time when online reputation management was neither a recognised profession nor an established market. Since then, he has advised DAX corporations, senior executives, leading politicians and public figures. He serves as an expert consultant to the German Bundestag, conducts research in cooperation with the University of Oxford, and is described by the Westdeutsche Zeitung as "Germany's leading expert in online reputation management."
Luke Kotlin was Executive SEO Expert at REVOLVERMÄNNER. His focus: the algorithmic side of reputation. How search engines categorise brands, which signals actually matter — and why a company's own website often plays a smaller role than commonly assumed.
Recon Rise grew out of this shared work. Not as a response to a hype cycle, but because an insight that had been building for years was suddenly confirmed in a new way by the arrival of generative AI.
I. Reconnaissance
Every strategic decision should begin with reconnaissance.
Before you act, you should see. Before you position, you should understand the terrain. That sounds obvious. In practice, it rarely happens.
What we have observed again and again over nearly two decades of work on reputation: most organisations start acting long before they have genuinely understood how they are perceived. And what others say about them — in trade media, on platforms, in review portals — has more influence on their fate than anything they communicate themselves.
Reputation. That was the foundational insight of 2007. The fact that it is more relevant than ever today, in a fundamentally different technological context, is not some irony of history. It is, when you think about it, entirely logical.
II. What Eighteen Years of Reputation Management Taught Us
In 2007, Christian Scherg founded REVOLVERMÄNNER — for a field that barely anyone took seriously at the time.
The starting point was straightforward: the internet does not forget. Whatever was written about a person or a brand stayed — and often carried more weight in public perception than any communication the organisation itself put out. Press coverage, trade media mentions, reviews, articles: that was the material from which reputation was built — or destroyed. Not the company website, not the press release.
Earned Media. What was earned through credibility, consistency and presence in sources that others actually trust.
Over the years, REVOLVERMÄNNER advised most of Germany's DAX corporations, guided senior executives through crises, and supported politicians through difficult periods. What kept being confirmed: the signal structure underlying trust — algorithmic, editorial, structural — follows essentially the same patterns. The question of which sources a search engine trusts turns out to be less fundamentally different from the question of which sources a human trusts than one might initially assume.
And then something arrived that recalibrated those mechanisms entirely.
III. The New Terrain
What was distinctive about the development from late 2022 onwards was not its scale — it was its speed.
ChatGPT launched in November 2022. One million users in five days. One hundred million in two months. No digital product before it had spread that quickly.
Today, 900 million people use ChatGPT every week (OpenAI / Sam Altman, February 2026). Perplexity processes 100 million search queries daily (Perplexity CEO Aravind Srinivas, Bloomberg Tech Summit 2025). And Google has embedded AI Overviews deeply into its search — with tangible consequences for organic traffic.
The concrete implications can now be measured. The SEO tool SISTRIX analysed over 100 million German search queries. The result: AI Overviews reduce the organic click-through rate for the top Google position in Germany by 59 percent — from 27 percent to 11 percent. That translates to 265 million lost organic clicks per month, in Germany alone. (SISTRIX, February 2026)
This is not a forecast for 2027. This is the current situation.
There is also an effect that tends to get overlooked in the broader discussion: the clicks that do still arrive — via AI recommendations — convert at 4.4 times the rate of conventional organic traffic, according to Semrush. (Semrush, AI Visibility Index, 2025) What is being lost is therefore not evenly distributed. It is, on balance, the most valuable part of the traffic that disappears first.
The new terrain is not necessarily more hostile than the old one. But it is more selective. And in selective markets, positioning carries disproportionate weight.
IV. The First Selection
A scenario many will recognise: a procurement manager is looking for a management consultancy for a digital transformation project, an HR software provider, or an M&A law firm.
Three years ago: open Google, click ten links, visit five websites, send two enquiries.
Today: open ChatGPT or Perplexity. Ask a question. The platform responds — with an assessment, with recommendations, with names. Three, sometimes five.
In that moment, before the first email is written or the first call is made, the shortlist already exists.
94 percent of B2B buyers actively use LLMs in their purchasing process — according to Forrester, GenAI is already the most-used source for self-directed research, ahead of traditional search engines. (6sense, B2B Buyer Experience Report 2025; Forrester, State of Business Buying 2025)
The important context: 70 percent of the B2B buying process takes place in the so-called Dark Funnel — the phase before a potential buyer makes any contact with a vendor at all. (6sense, 2025) Buyers research, compare and prioritise largely on their own. Those who don't appear in this phase find it considerably harder afterwards.
The numbers are precise: 95 percent of B2B deals go to a vendor on the Day-One Shortlist — that list of providers a buyer assembles through their own pre-purchase research, before any vendor contact takes place. This pattern has been documented in B2B research for years. What has changed is how the shortlist forms: where it once emerged from hours of searching, network recommendations and trade media reading, today it can take shape in a single AI conversation lasting a few minutes. Eighty percent of deals are won by the vendor who was already the frontrunner before that conversation began. (6sense, B2B Buyer Experience Report 2025)
Those who are absent from that conversation lose — usually not because their product is inferior or their pricing is wrong, but simply because the AI gave a different answer.
V. Why Some Brands Get Chosen
That is the genuinely interesting question.
What we keep finding when we analyse AI visibility: the mechanisms are less unfamiliar than they initially appear. At their core, AI systems select based on similar signals to those that have always generated trust — consistency, presence in reliable sources, clarity about who you are and what you do. Just formalised, scaled, and considerably less tolerant of inconsistency.
A study by Ahrefs examining 75,000 brands reaches a clear conclusion: the strongest correlating factor for AI visibility is not Domain Authority, not backlink profile, and not SEO budget. It is mentions in editorial sources, industry publications and structured databases. (Ahrefs, 2025)
Earned Media — what others write about a brand.
McKinsey examined where AI systems draw their sources from: a brand's own website accounts for only five to ten percent of the sources AI uses. (McKinsey & Company, 2025) Edelman's findings are even more direct: ninety percent of brand citations in LLMs come from earned media — not from the brand's own domain. (Edelman, 2025)
That is the actual shift. A company can have a technically excellent website — well-structured, optimised, strong on content — and still barely register in ChatGPT. Because AI systems build their picture of a brand not primarily from the brand's own self-presentation, but from what third parties say: journalists, analysts, trade publications, review platforms, structured databases.
This is fundamentally the same logic that underpins good reputation management. Trust does not come from what you claim about yourself. It comes from what others confirm.
GEO is therefore not a replacement for SEO. It is more accurately the replacement of a particular premise — namely that a company's own website is the central lever for visibility.
The consequence of this selection logic is, with appropriate caution about overly rigid categories, sobering: only seven percent of brands achieve dominant status in AI answers. Seventy-two percent remain permanently in the long tail — cited in fewer than twenty percent of relevant queries. (Search Engine Land / Mike Sonders, 2026)
The middle ground is narrow.
One further dimension that is frequently overlooked: which sources AI systems favour is neither random nor consistent across platforms. Profound analysed 680 million AI citations. Wikipedia accounts for 7.8 percent of all ChatGPT citations and dominates among the top-10 sources. At Perplexity, Reddit leads with 6.6 percent. Across platforms, 57 percent of all brand citations go to social proof content — reviews, listicles, forums, case studies. Thought leadership content accounts for 5.4 percent. (Omniscient Digital, 23,387 sources, 2025)
VI. Reconnaissance Before Action
36 percent of German companies are now using AI — nearly double the figure from a year ago. And 51 percent say that companies not using AI have no future. (Bitkom Research, September 2025, 604 companies)
The urgency has landed. That is good. But there is often more distance than expected between recognising a problem and responding to it correctly.
What we observe in companies that are just beginning to engage with AI visibility: the most common impulse is to apply familiar tools to a new problem. Optimise the website for AI. Enrich content with specific phrasings. Find the GEO equivalent of keyword density. That is understandable — but in many cases, it misses the point.
Princeton University conducted a broad study in 2024 examining which content strategies actually improve AI visibility. The result: well-structured, source-based content increases AI visibility by up to forty percent. Keyword optimisation — the core tool of classical SEO — tends to have a negative effect in generative engines. (Aggarwal et al., Princeton University / IIT Delhi / Georgia Tech, KDD 2024)
The tools and the underlying logic do not transfer straightforwardly.
Reconnaissance before action therefore means, first and foremost: understanding where you stand. How present is your brand in AI answers, and for which queries? On which platforms — ChatGPT, Perplexity, Gemini, Claude — and with what sentiment? What do LLMs say about the relevant competitors? Which sources are being cited, and which are missing?
Those are the questions that should be answered before any measures are taken.
VII. Architecture, Not Activism
In practice, we see two typical response patterns to this shift.
One: act quickly, produce content, add FAQ blocks, do something that feels like GEO. The equivalent of activism.
The other: understand how AI systems actually evaluate brands, build the relevant signals deliberately, and create a structure that holds up over time and across platforms.
AI visibility emerges — as far as the available evidence currently shows — primarily at the intersection of three factors:
Brand Clarity — Does an AI clearly understand who this brand is, what it does, and for whom? A brand that remains unclear to AI systems will not necessarily be described incorrectly. It will often simply not be mentioned at all.
Information Consistency — Are the facts about this brand consistent and free of contradictions across all sources, platforms and databases? LLMs aggregate from hundreds of sources. Inconsistencies create uncertainty — and that uncertainty frequently results in a brand being absent from the answer.
Reputation & Presence — Is this brand anchored in the sources that AI systems regard as trustworthy? In trade media, structured databases, editorial contexts. Presence here is not the same as volume. It is about being embedded in the right sources.
This is not conventional SEO work, and it is not a PR campaign. It is infrastructure — with the corresponding time horizon and the corresponding commitment.
A campaign creates an impulse. Infrastructure creates a durable condition. Recon Rise does not build one-off optimisation packages. It builds a system: one that continuously measures, analyses and constructs and maintains the signal architecture that keeps a brand present in the relevant AI answers over time.
VIII. Recon Rise
Recon Rise was not founded because a market gap looked compelling on paper.
It was founded because eighteen years of work on reputation questions led to a conclusion that the era of generative AI has confirmed in a new way. REVOLVERMÄNNER is the market leader in strategic reputation management in Germany. What has consistently been at the centre of that work: not what a brand says about itself, but what others say about it — in which sources, with what consistency, with what credibility.
The machine that evaluates those signals has changed. First it was Google. Today it is LLMs. The underlying principle has stayed the same.
What we spent two decades building for search engines — earned media structures, entity strength, signal consistency — is at its core exactly what AI systems require today in order to trust a brand. That is not a nostalgic parallel. It is the concrete reason why reputation expertise is a relevant foundation for GEO.
What does a brand need to be visible in the era of generative AI? Earned media. Clear entities. Consistent signals. Trustworthy sources.
That is not a new insight. It is an old truth in a new system — with considerably larger consequences for those who ignore it.
Most organisations will not miss this shift. They will notice it late. Because the old metrics will still work for a while: website traffic is still there, rankings are still visible, lead generation is still running. The moment when it becomes clear that the actual selection happened earlier — and elsewhere — tends to arrive without a clear warning signal.
Reconnaissance means: not waiting for that moment.
The relevant question today is no longer whether AI search will become dominant. For a growing share of purchase-relevant research, it already is. The question is whether a brand is part of the answer — or not.
Recon Rise measures where a brand stands today. Recon Rise analyses why AI systems do not mention it — or mention it incorrectly. Recon Rise builds the architecture that changes that.
Intelligence Brief — No. 01 Recon Rise GmbH · Düsseldorf · reconrise.ai © 2026 Recon Rise. All rights reserved.
Recon Rise, Düsseldorf · reconrise.ai
© 2026 Recon Rise. Alle Rechte vorbehalten.