What Six Months of Testing Actually Found
Online Advantages released research on 9 August 2026 covering six months of work across Google AI Overviews, ChatGPT, Gemini, Claude, Perplexity and Copilot. The research repeatedly pointed toward a combination of technical SEO, entity clarity, case studies, original information, third-party authority, reviews, structured information and broader digital presence. The agency found insufficient evidence for keyword density, thin AI-generated content, relying on traditional rankings alone or a single technical AI SEO trick.
Importantly, the authors explicitly state that the observations should not be interpreted as scientifically proven ranking factors. Some conclusions come from client observations and internal testing, others from outside research, so this should be regarded as practitioner evidence rather than a controlled academic study.
That caveat is important. But the directional signal is consistent with everything else we observe: there is no shortcut.
GEO Is Ambiguity Reduction
The strongest practical observation from the research is probably this: make the business easy for machines to understand, verify and trust.
That is much closer to entity reputation engineering than conventional on-page optimisation. And it points toward a concept worth owning: GEO is ambiguity reduction.
Traditional SEO became extremely webpage-centric: keyword → content → backlink → ranking. AI recommendations have a broader verification problem.
If your website says you are the UK's leading provider, but reviews do not support it, LinkedIn does not establish expertise, independent media barely mentions you, case studies lack evidence, directories contain contradictory data and community discussion says something different — the AI has epistemic conflict.
The optimisation task becomes reducing ambiguity across the digital footprint. The stronger the agreement between independent signals, the easier it becomes for an AI system to confidently represent the entity.
The AI Verification Map
The practical tool for ambiguity reduction is an AI Verification Map: a systematic audit of how well the business is understood, verified and trusted across nine signal categories.
1. Website and schema — Is the website technically crawlable? Does it have accurate, complete schema markup? Are entity relationships clearly defined?
2. People and expertise — Are key individuals clearly identified with verifiable credentials? Do their LinkedIn profiles corroborate the company's claimed expertise?
3. Case studies — Do case studies contain specific, verifiable outcomes with named clients (where permitted), clear timelines and measurable results?
4. Media coverage — Is the company mentioned in credible independent media? Are the mentions accurate and consistent with the company's own claims?
5. Reviews — Do review platforms contain authentic, detailed reviews that corroborate the company's claimed capabilities?
6. Creator and community — Is the company discussed in relevant communities and by credible creators in ways that are consistent with its positioning?
7. Directories and structured data — Are company details consistent across directories, data providers and structured data sources?
8. Competitor comparison — How does the company appear when AI systems compare it against competitors? Are the comparisons accurate and favourable?
9. Evidence freshness — Are the key corroborating sources up to date? Stale case studies and outdated media mentions weaken the evidence graph.
New GEO Metrics
The research supports a shift in how GEO performance is measured. Traditional SEO metrics — keyword rankings, organic traffic, backlink count — do not capture AI search visibility.
New GEO metrics should include: entity consistency (how consistently the brand is described across independent sources), citation diversity (how many different source types contribute to AI mentions), source overlap (how many sources AI systems cite when recommending the brand), recommendation share (what percentage of AI recommendations in the category include the brand), and evidence freshness (how recently the key corroborating sources were updated).
These metrics require different measurement tools — AI search monitoring platforms like Profound, Otterly or manual query testing — and a different reporting cadence. Monthly keyword ranking reports do not tell you whether AI systems are confidently recommending your brand.
The Integrated.Social Perspective
GEO is not replacing SEO. It is forcing SEO to finally deal with reality outside the website.
The brands that will win in AI search are not those that produce the most AI-optimised content. They are those that build the most coherent, verifiable, consistently corroborated evidence base across the full digital footprint.
That is a harder problem than keyword optimisation. It is also a more defensible one. Ambiguity reduction compounds over time: each new case study, each media mention, each authentic community discussion, each employee thought leadership piece adds another signal to the evidence graph. The brands that start building now will have a compounding advantage over those that wait for the magic hack that the research says does not exist.
