Here’s a scenario. Someone types into ChatGPT: “Who’s a good web designer near me?”
The AI gives a few names. Yours is either in that list or it isn’t.
Welcome to GEO (Generative Engine Optimisation): making your business the kind of source AI tools want to cite, recommend and mention.
Why GEO Is Suddenly a Thing
A few numbers explain the urgency.
- Scale. OpenAI announced in late February 2026 that ChatGPT had reached 900 million weekly active users, roughly double the 400 million reported a year earlier.
- Local discovery. BrightLocal’s 2026 survey of 1,002 US consumers found that the use of ChatGPT and similar tools for local business recommendations rose from 6% to 45% in a single year, making AI the third most popular source of local recommendations behind Google and Facebook. 40% of consumers said they trust AI platforms for business recommendations (32% said they don’t), and 42% said they trust them as much as traditional reviews.
- Fragmentation. Similarweb data shows ChatGPT’s share of AI chatbot web traffic falling from 79.0% in May 2025 to 53.9% in May 2026, with Google’s Gemini at 27.9% and Claude at 9.2%. Optimising for a single tool no longer covers most of the audience.
- Growth. Average monthly web visits across generative AI platforms grew 70% year on year, according to Similarweb.
How Is GEO Different from SEO?
With SEO, you’re competing for a spot in a list of links. With GEO, you’re competing to be mentioned inside an AI-written answer. There’s often no list, no page two, and sometimes no link at all.
AI tools build answers from sources they judge to be credible, clear and current. Many pull live results from the web and cite them, a process Google describes as retrieval-augmented generation, where the AI grounds its answer in retrieved content to improve accuracy and freshness. Google’s own AI features can also run several related searches at once (query fan-out) to assemble a response.
The practical takeaway: AI tools lean heavily on content that already exists, is easy to find, and looks trustworthy. GEO is about being that content.
What GEO Focuses On
Original research
AI has plenty of recycled content to draw on. Something genuinely new, like your own survey, dataset or case study, stands out. Google’s guidance for generative AI features explicitly cautions against simply repeating what’s already out there. Even a small original data point (“we reviewed 50 local business sites and found…”) gives an AI something worth citing.
Expert opinions
Real people with real views. Generic advice is everywhere. A clear point of view from someone who does the work every day is rarer and more quotable. Put named authors, credentials and first-hand experience on your content.
Statistics
Specific, sourced numbers give AI concrete material to reference. Replace vague claims (“many businesses struggle with this”) with precise, attributed ones (“47% of consumers won’t use a business with fewer than 20 reviews, according to BrightLocal’s 2026 survey”).
Entity consistency
Your business name, services, location and description should say the same thing everywhere: your website, Google Business Profile, social profiles and directory listings. Inconsistency makes it harder for AI to work out who you are. (We cover this in depth in the AIO guide.)
Brand mentions
Being talked about across the web, not just on your own site: industry publications, podcasts, directories, partner sites, YouTube, reviews and community discussion. This is one of the strongest signals in the data we’ve seen, and one of the hardest to fake.
Fresh information
Up-to-date content signals that you’re still active and still worth listening to. Show publication and “last updated” dates, and actually update your key pages when things change.
Structured content
Clear headings, short sections, lists and tables make it easier for AI to pull out the right chunk. It’s the same discipline as AEO, applied to a broader set of tools.
What the Research Actually Says
The most-cited study in this space is GEO: Generative Engine Optimization, from researchers at Princeton, Georgia Tech, the Allen Institute for AI and IIT Delhi, presented at the KDD 2024 conference. The team tested a range of content changes against a benchmark of 10,000 queries.
Key findings, as summarised by several analysts of the paper:
- The best-performing tweaks boosted visibility in AI responses by up to about 40%, and the standouts were adding relevant statistics, adding credible quotations and citing reliable sources.
- Keyword stuffing didn’t help, and in some tests performed worse than doing nothing.
- Clearer, more fluent writing also helped, with gains in the 15 to 30% range in the paper’s tests.
- The effect was biggest for lower-ranked sites. In one analysis of the results, sites ranked fifth in the traditional results gained over 100% in visibility from citing sources, while the top-ranked site lost visibility from the same tactic. That’s encouraging if you’re the challenger.
A few honest caveats. The study tested specific AI systems at a specific point in time, and the “up to 40%” figure applies to the researchers’ own benchmark. It isn’t a guaranteed result you can copy and paste. Treat it as strong evidence about what direction to lean, not a formula.
The Brand Mention Effect
Ahrefs analysed 75,000 brands to see which factors lined up with visibility in Google’s AI Overviews. The strongest link was branded web mentions, with a correlation of 0.664, well ahead of backlinks (0.218). Brands in the top quartile for web mentions earned up to 10 times more AI Overview mentions than the next quartile down. A follow-up study extended the work to ChatGPT and Google AI Mode, and found YouTube mentions correlated even more strongly (about 0.74).
The important caveat: this is correlation, not proof that mentions cause AI visibility. Well-known brands attract mentions and AI attention. And you can’t shortcut it. Manufactured mentions lack the editorial context that makes real ones valuable. The takeaway isn’t “buy mentions”. It’s “build a business, and a body of work, worth talking about.”
What Doesn’t Work
- Keyword stuffing. The Princeton research found it didn’t help.
- Mass-produced content. Volume isn’t the signal. Ahrefs found content volume correlated weakly (0.194) with AI visibility.
- Fake reviews or manufactured mentions. They risk penalties and legal trouble, and consumers hate them. In BrightLocal’s survey, 97% of consumers think businesses should face consequences for fake reviews.
- Promises of guaranteed citations. Nobody controls these systems.
How You Know It’s Working
GEO measurement is still rough, but there are practical ways to track it.
- Prompt testing. Make a list of 10 to 20 questions your customers might ask an AI tool. Run them monthly across ChatGPT, Gemini, Perplexity, Claude and Google’s AI Overviews. Record whether you’re mentioned, cited, or missing.
- AI referral traffic. Check your analytics for visits from AI platforms. Keep in mind this undercounts: Similarweb notes that a lot of AI-driven traffic lands on homepages and may be filed as “direct” with no clear source.
- Brand search. Are more people searching your business name?
- Enquiry sources. Ask new customers how they found you. “ChatGPT told me” is now a real answer.
- Accuracy checks. When AI does mention you, is it right about what you do?
The Honest Bit
This is a new and fast-moving field. Nobody, and we mean nobody, has a guaranteed formula for getting into ChatGPT’s answers. The tools change constantly, and what works today may not next year.
What we do know: the businesses most likely to get mentioned are the ones that are genuinely credible, clearly described and consistently visible. That’s good marketing regardless of which AI is in fashion. And Google’s own guidance on generative AI features puts the emphasis on the same things it always has: useful, original, trustworthy content on a site that’s easy to crawl.