Burson, the global communications agency focused on creating value through reputation, has released The Credibility Paradox, a new report revealing significant variations in how audiences perceive and believe AI-generated answers about brands and companies. The research advances the conversation around Generative Engine Optimization (GEO), shifting the focus from visibility and source citations to the strategic challenge of building credibility and believability.
Based on more than 55,000 believability scores across 85 companies, the study found that the credibility of AI-generated responses varies significantly by audience, with business decision-makers rating answers 10% more believable on average than the general population.
“In today’s zero-click world, LLMs have become the new gatekeepers of reputation – how brands are discovered and evaluated. But visibility is not credibility,” said Corey duBrowa, CEO, Burson. “AI synthesises, summarises and delivers information directly to audiences. Appearing in these LLMs is necessary, but it is no longer sufficient. Our role is no longer just to make clients visible, but to build an evidence ecosystem so robust that the answers AI constructs are believable to the audiences that matter most. This research is our playbook for turning the credibility paradox into a competitive advantage.”
Burson partnered with Profound, a leading AI marketing platform, to field thousands of reputation-related responses across seven major AI answer platforms. The study assessed 85 companies against the eight pillars of Burson’s Reputation Capital framework: Innovation, Creativity, Workplace, Products, Financial Performance, Governance, Citizenship and Leadership.
Responses were assigned believability scores for three audience groups – General Population, Opinion Elites and Business Decision Makers – using Burson’s proprietary Decipher tool, developed in partnership with cognitive AI company Limbik. The research generated more than 55,000 believability forecasts.
Key Findings
* AI rewards proof, not positioning. Fact-based claims linked to innovation, products and workplace culture consistently outperformed those associated with more subjective attributes such as leadership, governance and citizenship. This highlights the importance of a balanced mix of earned, owned and social content in GEO strategies, as AI systems place greater weight on independent validation from media coverage, reviews and online conversations.
* Workplace is an underutilised credibility driver. As highlighted in Burson’s Global Reputation Economy research, workplace reputation remains one of the most underleveraged drivers of reputation capital. Workplace-related responses were found to be the most believable among the general population, reflecting LLMs’ reliance on independently verifiable sources such as employee reviews, labour reporting and earned media coverage.
* Leadership is AI’s toughest credibility test. Responses to leadership-related prompts consistently ranked among the least believable across industries. The sectors that performed better, including Aerospace and Technology, shared a common characteristic: credibility was built through governance practices, business performance and third-party validation rather than executive messaging alone.
* Believability varies by audience. A narrative that appears credible in an AI-generated response may not resonate equally with customers, investors, employees or regulators. Business decision-makers rated AI-generated answers 10% more believable on average than the general population, while specialised audiences were generally more receptive to innovation-focused narratives and broader business context. The findings underscore the need for audience-specific GEO strategies.
The research has informed a framework developed by Burson to help clients build and protect reputation across AI platforms. Rather than treating earned media, owned content and social engagement as separate disciplines, the framework takes a holistic approach, cultivating an ecosystem of independent and credible voices whose coverage and commentary reinforce key narratives over time.
Burson also incorporates language and market-specific nuances to help organisations navigate reputation challenges across regions and cultures.
“Across APAC, much of the conversation around AI has centred on whether brands appear in AI-generated answers, while far less attention has been given to whether those answers are accurate, credible and believable. That is the gap our report addresses,” said Red Surtida, APAC Head of Intelligence & Transformation at Burson.
“As AI becomes an increasingly influential layer between companies and their stakeholders, it is shaping not only how brands are discovered, but also how they are understood and evaluated. The real opportunity for organisations is not simply to secure a share of AI-generated answers, but to ensure those answers are grounded in evidence, backed by credible sources and believable to the audiences that matter most.”
Steve Rubel, EVP, Media Insights & Measurement at Burson, added: “GEO may have started as a visibility challenge measured through audit reports, but it has evolved into something far more consequential. The data from this study makes clear that it is now a test of whether a company’s real-world reputation is legible, corroborated and believable within the AI-mediated environments where audiences increasingly form opinions. Our framework provides communicators with a practical path forward and establishes GEO as a new domain in reputation management.”
