A few months ago, the founder of a mid-sized brand asked me a question that stopped our meeting cold: “We rank number one on Google. So why does ChatGPT keep recommending our competitor?”
The ground has already shifted
Consider the numbers. ChatGPT crossed 800 million weekly active users in October 2025 and has kept climbing since. Gartner projected that traditional search volume would fall 25% by 2026 as AI assistants absorb queries that once went to search engines. And Pew Research Center found that when Google shows an AI-generated summary, users click a traditional result only 8% of the time, versus 15% when no summary appears. Clicks on links inside those summaries? Just 1%.
Read that again. The click, the metric an entire industry was built on, is quietly disappearing. Your customers are still asking questions. They are just getting answers without ever visiting your website.
This is not a niche shift confined to tech-savvy early adopters either. It is showing up across categories that once felt immune, healthcare providers, B2B software vendors, local service businesses, even legal and financial firms whose clients used to start every decision with a Google search. If your category involves someone researching before they buy, AI assistants are already inserting themselves into that research.
Here is what worries me more: the answers are not always right. A study coordinated by the European Broadcasting Union and the BBC, spanning 18 countries and more than 3,000 responses, found that 45% of AI assistant answers contained at least one significant issue. If AI engines can misrepresent major news organisations, they can certainly misrepresent your brand, your pricing, or your product claims. And most brands would never know.
Fix your entity footprint
AI engines do not rank pages; they assemble an understanding of entities. If your company’s name, founding facts, services, and locations are inconsistent across your website, directories, LinkedIn, and knowledge databases like Wikidata, you are feeding the machines contradictions.
Clean, structured, consistent facts are the new technical SEO.
In practice, this means treating your entity footprint like your XML sitemap: audited on a schedule, not fixed once and forgotten. Start with a single source of truth for your company description, founding year, and service offerings, then check every external listing matches it word for word. Small inconsistencies, a mismatched founding date on Crunchbase, a service missing from your Google Business Profile, are exactly the noise that pushes an AI model toward a cleaner, better-documented competitor.
Earn citations, not just backlinks
Notice what AI answers actually cite: independent editorial coverage, review platforms, expert commentary, community discussions. A brand mentioned credibly in ten trusted third-party sources will outperform a brand with a beautiful website and no external validation.
Corroboration is the new currency.
This changes how PR and link building should be briefed. A placement’s value is no longer just the domain authority it passes; it is whether an AI model would recognise that outlet as a credible voice in your category. Trade publications, analyst reports, and active community threads increasingly carry more weight in AI answers than a generic guest post ever will.
Write for answers, not rankings
Pew’s research found that question-style searches trigger AI summaries about 60% of the time. Structure content the way people ask: clear questions, direct answers in the first two sentences, specifics an AI can quote.
If a machine cannot lift a clean answer from your page, it will lift one from someone else’s.
Measure what machines say about you
Traffic and rankings will not capture this shift. Track how often you are mentioned in AI answers, whether the information is accurate, and how you compare against competitors and review it monthly. Even a simple monthly log of AI responses beats flying blind.
Build this into your existing reporting cadence rather than treating it as a side project. A basic tracker, a shared spreadsheet with category-relevant prompts run monthly across ChatGPT, Perplexity, and Gemini, is enough to spot patterns: which competitors keep showing up, which claims get repeated incorrectly, and which content formats earn a citation. The tooling matters less than the discipline of checking regularly.
The window is open, for now
That founder I mentioned? Six months of unglamorous work later, entity clean-up, editorial placements, restructured content, their brand now appears in AI recommendations for their category. Nothing magical. Just early.
That is the real message. Generative Engine Optimization is not a dark art; it is disciplined brand hygiene for a machine-mediated world. The brands that treat AI visibility as seriously as they once treated Google rankings will own the recommendations everyone else is scrambling for in 2027. The audit costs you an afternoon. Being invisible costs considerably more.
(Views are personal)
