Our Expertise Services Industries Case Studies Insights Connect
AI Search

Most of what wins in AI answers is work you already know how to do.

A lot of firms claim they have cracked these new channels. Nobody has. Here is the work we see moving both surfaces.

Talk to us about your site Try Forge →
The Quick Answer

What SEO, AEO, and GEO each mean.

SEO
Search engine optimization. Getting a site to rank in the normal Google results, where nearly all organic traffic still comes from.
AEO
Answer engine optimization. Getting cited, and being the source the answer is built from, inside ChatGPT, Perplexity, and Google’s AI Overviews.
GEO
Generative engine optimization. Being described accurately when a model writes an answer about your category.

All three pull from the same website, which is the whole argument of this page. Definitions follow Digiday’s GEO and AEO explainer. Digiday

Where The Traffic Is

How big AI search is.

Where the traffic is today
87.63%

of search referral traffic came from Google in May 2026, versus 0.29% for all AI chatbots combined.

Source: Cloudflare Radar, via Technology Checker
~91%

of the global search engine market is still held by Google.

Source: Statcounter
Where it’s heading
42.8%

year-over-year growth in AI search visits in Q1 2026, while Google grew 2.4% over the same window.

Source: Wix Studio AI Search Lab
37%

of 500 people who already use AI tools said, in a November 2025 survey, that they now start their searches with AI rather than a traditional search engine.

Source: Search Engine Land
900M

weekly active users on ChatGPT, per OpenAI, February 2026.

Source: Search Engine Land
~2B

people reached monthly by Google AI Overviews.

Source: Higoodie
693%

year-over-year growth in AI referral traffic to US retail in the 2025 holiday season, and it converted 31% better than non-AI traffic.

Source: Adobe Digital Insights, via Omnibound

Both are true at once, and the growth column is the part people skip. So the channel is small, growing fast, and converting better than average, and we tell partners to build for it without pulling money out of the one still paying the bills.

The Foundation

What carries over from SEO.

Folks early in a discovery process want to know whether AI search means doing something completely different. Mostly it doesn’t. If there is technical debt on your site, that is already bad for Googlebot, and just as bad for the ChatGPT and Anthropic crawlers reading the same pages. Internal linking carries over, and so do page speed, backlinks, and schema, which is the bit of code that spells out for a machine what a page is about.

Google’s own documentation says it more plainly than we could. There are no additional requirements to appear in AI Overviews or AI Mode, and no special optimizations, so there is no AI text file and no AI-specific code to chase. The page has to be indexed and eligible to show in Search with a snippet, and that is the whole gate. Google Search Central

Speed is one of the few things anybody has measured from the outside. Across 216,524 pages on 129,000 domains, the pages that put something on screen in under 0.4s averaged 6.7 ChatGPT citations, against 2.1 for pages over 1.1s. That is roughly a 3x penalty for being slow. We have run refreshes for partners strictly on speed, and it is some of the least glamorous work we do. SE Ranking

Good fundamental SEO is good GEO. We don’t think you can outrun bad fundamentals and get rewarded in AI answers.

The Overlap

Six fundamentals, and what each one does on both surfaces.

ItemOn GoogleIn the answer engines
Crawlability
On GoogleIf Googlebot can’t fetch the page and index it, nothing else here matters.
In the answer enginesGPTBot and the rest hit the same URLs on the same servers, so a page that blocks one blocks all of them.
Structured content
On GoogleClean headings and a sane folder structure are how Google works out what a page is about.
In the answer enginesThose same headings are what a model lifts and attributes.
E-E-A-T
On GoogleExperience, expertise, authoritativeness and trust, the four things Google’s own quality raters are told to look for.
In the answer enginesNamed authors, real citations and first-hand detail are what make a page safe to repeat.
Schema
On GoogleCode that labels your prices, locations and hours for a machine, and earns you rich results.
In the answer enginesGoogle says there is no AI-specific markup to add, so schema helps for the same reason it helps Search.
Speed
On GooglePage experience has been an input for years, and slow pages lose on mobile first.
In the answer enginesSame fix, sharper penalty. The citation gap above is one of the few AI-side effects anybody has measured.
Authority
On GoogleBacklinks still move rankings, and that part of the job has not changed.
In the answer enginesAuthority is the one input we have watched move AI citations on its own, with nobody touching the site.

Draw a Venn diagram of SEO, GEO and AEO and the differences sit out at the edges. The work that decides whether a crawler can find your content, read it and trust it is the same in all three.

What To Check

Four things you should be able to see for yourself.

  1. 01

    Where the numbers come from

    Ask to see the actual Search Console and Analytics accounts behind the reporting, then ask which parts nobody can measure yet. Be careful with any report that won’t tell you which half is which.

  2. 02

    Whether a crawler can read the site

    Schema, load times, internal linking, and a structure that makes sense to a person and to a machine at once. Most of what gets sold as AI-search work happens right here, and all of it is checkable line by line.

  3. 03

    Whether every article holds the same line

    Pull three published pages at random and see whether they carry the same author block, the same citations, and the same house style. Ours run through Forge, our content system, so they do.

  4. 04

    Whether the library improves after publish

    Published content either compounds or ages out, and the difference is whether anybody goes back to it. Ask what gets refreshed, how often, and what was added the second time.

Most of what gets added on that second pass is more audiences. Somebody asking ChatGPT packs far more into one question than they would type into Google, so content can’t stay focused on one keyword. We go back and ask who else would benefit from the page, and then we say it outright, in the headings or in the FAQ.

On an auto transport client, that looks concrete. The New York to Florida route page runs the general service information first, then a block for snowbirds, who all ship in the same few weeks and need their own pricing and reviews. A page for a university town gets a block for out-of-state students. “I’m a student in Rochester and I need to ship my car to California, and I don’t have a lot of time” is a real question.

Where AI Search Differs

The AI-specific moves.

None of these replace the fundamentals, and we would hold off until the site underneath is clean and fast. After that, these are the moves we have seen sharpen how often a page gets cited.

  • Front-load the answer. 44.2% of the citations AI tools hand out come from the first 30% of a page, so put the answer near the top (Position Digital).
  • Add expert quotes, statistics, and inline citations. Those were the three top-performing changes in the Princeton GEO study, which lifted visibility 30–40% on the paper’s position-adjusted word count metric, with quotations best at 41%. It is the structure Forge makes our writers use (Aggarwal et al., KDD ’24).
  • Expect platform fragmentation. AI Overviews and AI Mode land on the same URLs only 13.7% of the time (Ahrefs), and only 11% of domains get cited by both ChatGPT and Perplexity (The Digital Bloom), so coverage is earned on each surface separately.
  • Build distribution beyond your own site. Stories placed through earned media earned a median 239% more AI citations than brand-owned content alone, and that is where authority compounds too (Stacker).
  • Treat citation as a ranking signal. Being cited in an AI Overview delivered 120% more organic clicks per impression than not being cited, across 2.43B impressions, so a citation pays you back in both channels (Seer Interactive).
From the Front Lines

What happened when one client picked up a run of backlinks.

8X
Leads from ChatGPT, in one week

The most dramatic live example we have seen was a partner who picked up a windfall of quality backlinks, with nobody touching the content on their own site. We saw that client’s leads from ChatGPT go up 8X in a week, off a meaningful base, and the results have sustained for over a month.

We are not saying that works for everybody. What we take from it is that the groundwork we had already done, the content and the speed work, meant the new authority had something to land on.

First-party account, from a live client engagement
What Nobody Can Measure Yet

What the reporting can and can’t tell you.

For a growth partner like us, the whole point is to anchor performance in data. Google Search Console and Google Analytics are fantastic for that, because the reporting they give you is real and you can go and look at it yourself. We love plenty of the other data services too. The AI-search side is not there yet. There are services that run their own test queries and estimate results, but none of them shows you what people are actually asking.

So until you can see that reporting coming from the services themselves, we think you need to be wary of anyone claiming to have all the answers on AI search, and that includes folks like us. We do see sometimes incredible case studies of success on these channels, and we are still fundamentally skeptical, because new models keep getting released, and the models are black boxes, and the reporting is a black box.

That is the strategy we bring to clients. Get the fundamentals right, build a content engine that compounds, and keep refreshing it so it gets smarter instead of aging out. We see real results with it, in Google and in the answer engines both.

References

Where these numbers come from

  1. Digiday. “WTF are GEO and AEO, and how they differ from SEO.” digiday.com
  2. Cloudflare Radar, via Technology Checker. “Search Engine Market Share.” Referral share, May 1–31, 2026. technologychecker.io
  3. Wix Studio AI Search Lab. Crystal Carter, “AI search vs. Google,” April 30, 2026. Similarweb data, worldwide, Q1 2026 against Q1 2025. wix.com
  4. Search Engine Land. Danny Goodwin, “37% of consumers start searches with AI instead of Google: Study,” January 7, 2026. Reporting Eight Oh Two’s 2026 AI and Search Behavior Study, a survey of 500 active AI users fielded November 2025. searchengineland.com
  5. Statcounter GlobalStats. “Search Engine Market Share Worldwide,” July 2026. gs.statcounter.com
  6. Search Engine Land. “OpenAI: ChatGPT now has 900 million weekly active users,” February 2026. searchengineland.com
  7. Higoodie. Mostafa ElBermawy, “AI Search Market Share Report, Wave 2,” May 21, 2026. Cited for the AI Overviews monthly reach figure, which Higoodie attributes to Alphabet, Q1 2026. higoodie.com
  8. Adobe Digital Insights, via Omnibound. “AI Search Statistics.” US retail, 2025 holiday season, reported January 2026. omnibound.ai
  9. Google Search Central. “AI features and your website.” Updated December 10, 2025. developers.google.com
  10. SE Ranking. Yulia Deda, “Ranking Factors for ChatGPT,” November 24, 2025. 216,524 pages across 129,000 domains in 20 niches. seranking.com
  11. Position Digital. “AI SEO Statistics.” Relaying Growth Memo, February 2026, for the citation-position figure. position.digital
  12. Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan and Deshpande. “GEO: Generative Engine Optimization.” KDD ’24, ACM SIGKDD, 2024. Measured on GEO-bench, 10,000 queries. arxiv.org/abs/2311.09735
  13. Ahrefs. Despina Gavoyannis and Xibeijia Guan, “AI Overviews vs AI Mode,” December 15, 2025. 540,000 US query pairs, September 2025 data. ahrefs.com
  14. The Digital Bloom. Vlad Kuriatnyk, “2025 AI Citation and LLM Visibility Report,” December 14, 2025. thedigitalbloom.com
  15. Stacker. “Latest Research on Expanding Brand Visibility Across LLMs,” March 16, 2026. 87 stories across 30 brands, 2,600+ prompts, 8 AI platforms, 30 days. stacker.com
  16. Seer Interactive. Tracy McDonald, Hannah Cooley and Marketa Williams, “AIO Impact on Google CTR: 2026 Update,” April 24, 2026. 53 brands, 5.47M queries, 2.43B impressions, January 2025 to February 2026. seerinteractive.com
What’s Next

We’ll walk your site with you, on Google and in AI search.

Tell us where you’re stuck and we’ll tell you which of the six fundamentals is costing you the most, and which ones are already fine and don’t need your money.

Talk to us about your site Try Forge