On June 2, 2026, communications agency Burson released The Credibility Paradox, a study run with AI platform Profound. The finding is blunt: being cited by ChatGPT, Perplexity or Gemini does not guarantee being believed. Across 85 companies and more than 55,000 believability forecasts, trust in AI-generated answers varies sharply depending on what they claim, and on who reads them.
In short: the first wave of GEO (Generative Engine Optimization) chased visibility, appearing in AI answers. The Burson x Profound study opens the second wave: credibility. An AI answer can mention you without the reader believing it. Factual, verifiable claims (products, innovation, workplace culture) are judged credible; subjective claims about leadership are not.
What the study found
Burson queried seven AI answer platforms and scored each brand response across eight “reputation levers”: innovation, creativity, workplace, products, financial performance, governance, citizenship and leadership. Every response got a believability score for three audiences, general population, opinion elites and business decision-makers, using Decipher, Burson's proprietary tool built with cognitive AI company Limbik. The total: more than 55,000 forecasts.
Three results stand out. First, business decision-makers rated AI answers 10% more believable than the general public: the audience most valuable to a B2B firm is also the one most inclined to trust what AI says about you. Second, content tied to the workplace is the most believable for the general public, because models lean on independently verifiable sources, employee reviews, labor reporting, earned media. Third, responses about leadership are the least believable across every industry studied.
“Visibility is not credibility,” says Burson CEO Corey duBrowa, who frames the agency's role as building “an evidence ecosystem so robust that the answers AI constructs are believable.” Steve Rubel, EVP Media Insights & Measurement, goes further: “GEO began as a visibility challenge” but has become “a test of whether the reputation a company has earned in the real world is legible, corroborated and believable.”
Why this is a turning point for GEO
For two years, the SEO industry has chased one question: how do I get cited by AI? We've measured impressions inside AI features in Search Console, tracked AI Overviews coverage on business queries, and optimized passages to be extractable. That race is still necessary, but it is no longer sufficient.
The study moves the target. An AI that cites you may still produce a sentence nobody believes, because it isn't corroborated anywhere else on the web. Conversely, a mundane claim that is cross-checked by verifiable third-party sources will be judged reliable. The credibility of an AI answer isn't decided on your “About” page: it's built in everything others say about you.
What it changes for businesses
The lesson is concrete: stop betting everything on self-promotion. Pages that shout “market leader” or “team of passionate experts” are exactly the kind of claims AI judges least credible. What makes an AI answer reliable is external evidence: precise product data, employee reviews on platforms like Glassdoor, trade press, quantified case studies, consistent mentions from one site to the next.
Three levers gain value:
- Build an evidence ecosystem. The more a claim is repeated and cross-checked by independent sources, the more AI treats it as credible. Earned media, partnerships, verifiable testimonials, open data.
- Favor facts over declarations. Replace “we're the best” with numbers, specs, and dated, sourceable results. That's what models can corroborate.
- Watch mention consistency. Contradictory information across sources, address, leadership, offerings, weakens credibility. It's the flip side of what Google's own AI search guidance has been signaling.
This logic extends what Google's GEO guidance has repeated for months: there is no trick to “hack” AI answers. You need a real reputation, legible to machines.
The limits of the study
A few useful caveats. The sample, 85 companies, is mostly large brands, not small or mid-sized businesses; transposing it takes judgment. The credibility scores are forecasts produced by a proprietary tool (Decipher), not real audience measurements. And the study covers perceived credibility, not conversions or revenue directly. It tells you how to make an AI answer believable, not how much that pays off. Still, the direction is clear: the next GEO battle will be fought over trust, not just presence.
Our take
At Cicéro, we've said it from the start: AI visibility without substance is a mirage. Being cited by a machine that contradicts itself, or that nobody believes, is worthless. The real question is no longer “is AI talking about me?” but “does what it says hold up, and is it corroborated elsewhere?” The next step in GEO isn't a new technical hack: it's earning, in the real world, the reputation AI reconstructs on your behalf.
Frequently asked questions
What is credibility in GEO?
It's how much a reader trusts an AI-generated answer about a brand. According to the Burson x Profound study (June 2026), a brand can be cited by AI without that answer being judged credible, especially when the claim isn't corroborated by any third-party source.
Which content does AI judge most credible?
Factual, verifiable content tied to products, innovation and the workplace. Subjective claims about leadership are judged the least credible across every industry studied.
How do you improve your brand's credibility in AI answers?
By building an ecosystem of external evidence: precise product data, employee reviews, trade press, quantified case studies and consistent mentions across sources. The more an item is cross-checked by independent sources, the more AI treats it as reliable.
Sources
- → Burson, The Credibility Paradox (June 2, 2026), study with Profound, 85 companies, 55,000+ believability forecasts.
- → MediaBrief, quotes from Corey duBrowa (CEO) and Steve Rubel (EVP), and the eight reputation levers.
- → MediaNews4u, data on the credibility gap between decision-makers and the general public.
Growth and SEO content strategist, I founded Cicéro to help businesses build lasting organic visibility, on Google and in AI-generated answers alike. Every piece of content we produce is designed to convert, not just to exist.
LinkedIn