SEO, AEO, and GEO are three disciplines with three target systems. This post sorts the terms, shows the solid data, and points out where the concepts have limits.
V. Murati, fdk.aiA few weeks ago, I sat down with a managing director in Schaffhausen. He had a quarterly report from his SEO agency on the table. Rankings stable, top-3 positions slightly expanded, technical audits green. Yet his inquiries were collapsing. He asked me whether the agency was cheating. They aren't. The world around him has shifted, but his agency's reporting no longer measures it.
Three acronyms are circulating in the market right now that overlap but don't mean the same thing: SEO, AEO, and GEO. Some providers throw them into one pot. Others build three separate business models around them. Both are wrong. They are three disciplines with three different target systems, and confusing them either burns budget or causes you to miss the shift that is happening right now.
This post sorts the terms, shows the data behind them, and points out where the concepts have limits. It is shorter than a long-read and deliberately not a sales pitch. What we do at fdk.ai gets exactly one mention, in half a sentence.
I, V. Murati, run fdk.ai. We sell exactly the services this text presents as solutions: SEO, AEO, and GEO for DACH SMEs. Read this with that knowledge in mind. I'm biased. I still try to show the data honestly, name the counter-arguments, and admit when something doesn't fit. Whether I succeed is your call.
SEO is the oldest of the three. Search engine optimization aims at a position in the classic results list, meaning a blue link that someone clicks. Success metric: rankings, clicks, organic traffic. Audience: Google and Bing in their classic form. The discipline has been established for over twenty years, and it still works because a large share of search queries still ends in a results list.
AEO, or Answer Engine Optimization, aims at something different. The goal is for content to appear as a direct answer, in featured snippets, in Google's position zero, in voice search responses, and increasingly in Google's AI Overviews. The person doesn't necessarily click; they read the answer and are done. Success metric: answer visibility, snippet frequency, mentions in AI Overviews.
GEO stands for Generative Engine Optimization. The term comes from a 2024 paper by Aggarwal et al. presented at the ACM KDD conference (lead author at IIT Delhi, co-authors at Princeton and independent). GEO is not the same as AEO, even though the terms are often used synonymously. GEO addresses generative engines like ChatGPT, Perplexity, Claude, and Gemini, meaning systems that synthesize answers from multiple sources. Success metric: brand mention and content citation within the generated answer itself, often without any click ever happening. The authors showed in a benchmark with 10,000 queries that targeted GEO measures can boost visibility in generative answers by up to 40 percent, depending on domain and strategy.
Practical rule of thumb: SEO gets you onto the list. AEO gets you into the answer box. GEO gets you into the answer text that an LLM generates for someone else. Three different mechanics, three different optimization levers. Anyone who rolls all three into one report is comparing apples to oranges to quinces.
A small reality check on the terminology confusion. HubSpot uses AEO as a broad umbrella that also includes generative engines. Other providers use GEO as a catch-all for everything related to AI visibility. Still others introduce LLMO, GAIO, or AIO; some are serious concepts, others are pure marketing labels. The industry isn't aligned, and that probably won't be settled in the next twelve months. I'll use the narrower reading from the Aggarwal paper: AEO for answer boxes and snippets, GEO for synthesized LLM answers. It is analytically cleaner and makes reporting definitions possible.
The data is messy but not contradictory if you look closely. Three studies are particularly solid.
Pew Research published an analysis in July 2025 based on 68,879 actual Google searches by 900 US adults who shared their browsing behavior in March 2025. Result: when an AI Overview appears, the click-through rate to a search result drops from 15 percent to 8 percent, nearly cutting in half. 26 percent of sessions end completely without the person clicking anywhere. AI Overviews appeared in 18 percent of all searches at the time. Later analyses by Semrush and Ahrefs show that this share fluctuated throughout 2025 (peaking around 25 percent in July, dropping to 16 percent in November) but trended upward overall.
Seer Interactive analyzed a larger sample between June 2024 and September 2025: 3,119 informational queries across 42 organizations, 25.1 million organic impressions. Finding: for queries with an AI Overview, organic CTR fell by 61 percent, paid CTR by 68 percent. But, and this is the interesting number, brands that were cited within an AI Overview had 35 percent more organic clicks and 91 percent more paid clicks than non-cited brands on comparable searches.
Ahrefs followed up with an analysis of 300,000 keywords using December 2025 data (published February 2026), comparing the click-through rate of position-1 content with active AI Overviews against the December 2023 baseline. Result: 58 percent decline. In April 2025, the same value was still 34.5 percent. The deterioration continues.
There is a counter-position. Semrush showed in a study of 200,000 keywords (January to October 2025) that the zero-click rate for keywords with AI Overview slightly fell from 33.75 percent to 31.53 percent. The reading: AI Overviews tend to appear on searches that rarely led to clicks anyway, and the effect is not uniformly causal. This difference is real and honest. It does not, however, change the underlying movement: search queries are shifting from results lists to answer formats, and the platforms are multiplying.
Add the adoption data. Eurostat reported that for 2025, 47 percent of the Swiss population aged 16 to 74 used a generative AI tool at least once in the three months prior to the survey. In Germany the figure is 32 percent, EU average 33 percent. ChatGPT reports around 900 million weekly active users in early 2026 (DemandSage, March 2026). Anyone who assumes their customer base is over 40 and concludes that none of this applies to them should look at the Eurostat numbers for the 35 to 54 age groups. The gap to the younger cohort is closing faster than expected.
One last observation that is most often underestimated in practice: results are volatile. An Authoritas analysis shows that around 70 percent of pages cited in AI Overviews change within a two-to-three-month window, and these changes don't correlate with classic Google rankings. A position 1 in Google is therefore no guarantee for citations in AI answers, and a citation today doesn't mean it will still be there next month. Anyone expecting the same reporting model as for SEO ("optimize once, then stable") will be disappointed.
Three pragmatic consequences, in the order I most often see them surface in client work.
First, SEO is not dead. Anyone who writes that is either selling GEO as a miracle cure or hasn't read the data. Classic search queries still account for the bulk of volume, especially for local, transactional, and brand searches. Without an SEO foundation, your content rarely gets cited in AI Overviews or LLM answers either, because the models prefer indexed, technically clean sources. SEO is the foundation on which AEO and GEO operate.
Second, AEO and GEO are not optional when the target audience actively uses AI tools. In most of the engagements we run at fdk.ai, AI visibility across all four major engines (ChatGPT, Gemini, Perplexity, Google AI Overviews) starts at zero. Not near zero, actual zero. Which means: when a prospect asks ChatGPT "Who does X in my region?", the company in question doesn't appear, even if it ranks position 2 on Google for the same terms. That is a different league of invisibility than a lost ranking, because the person never makes the click.
Third, the lever is not "more content." Aggarwal et al. show in the same study that adding statistics increases visibility in generative answers by up to 41 percent, adding citations by up to 28 percent. Keyword stuffing, by contrast, decreases it, possibly even more aggressively than on Google. The discipline calls for substance, clear structures, citable statements, and machine-readable data (Schema.org markup for articles, FAQs, local business data, persons, products). Exactly the things good content should already have.
A technical point that many underestimate: robots.txt decides whether AI crawlers like GPTBot, PerplexityBot, or Google-Extended can access content at all. We regularly see in client engagements that websites accidentally block these bots, often through an older Cloudflare default setting. If you aren't crawled, you can't be cited. The check takes ten minutes; the effect is binary.
When this isn't worth it: when the audience is local, older, and not digitally inclined (example: a local hairdresser in a rural town). When the business is high-frequency and transactional (example: takeaway). When the margin can't sustain two- or three-figure monthly investments. In these cases, solid local SEO and a well-maintained Google Business Profile are enough.
If you take only one thing from this text: tomorrow, search for your own company in ChatGPT, Gemini, Perplexity, and Google AI Overviews. Type the questions your customers would type, not your company name. If you aren't mentioned, you don't necessarily have an SEO problem. You have an AEO and GEO problem, and that is a different construction site. Anyone who has understood this knows why three terms are three disciplines, and not one marketing label with three spellings.
SEO optimizes content for classic results lists in Google and Bing. Success metric is ranking and click. AEO (Answer Engine Optimization) optimizes for direct answer formats like featured snippets, voice search, and Google AI Overviews. Success metric is the mention in the answer box. GEO (Generative Engine Optimization) optimizes for synthesized answers in ChatGPT, Perplexity, Claude, and Gemini. Success metric is brand citation in the generated text, often without a click.
Yes, and it remains the foundation. Classic search queries still account for the bulk of volume, especially for local, transactional, and brand searches. LLMs like ChatGPT also prefer indexed, technically clean sources. Without an SEO foundation, your content rarely gets cited in AI Overviews or LLM answers either.
The simplest method: monthly, type 10 to 20 typical customer questions into ChatGPT, Perplexity, Gemini, and Google AI Overviews and log whether your brand is mentioned. Specialized tools like Profound, Otterly, or Peec.ai automate this. Classic SEO tools don't show this kind of visibility.
When the audience is local, older, and not digitally inclined. When the business is high-frequency and transactional (example: takeaway). When the margin can't sustain two- or three-figure monthly investments. In these cases, solid local SEO and a well-maintained Google Business Profile are enough.
Three criteria are decisive in practice. First, does the provider measure citations in LLM answers, or only classic rankings? Anyone reporting only Google rankings hasn't understood the shift. Second, can they show concrete before-and-after data from engagements, ideally with methodology? Third, is the optimization methodology documented or a black box? Specialized providers in the DACH region are still rare in 2026; many agencies are only just starting to build the capability.
The central data points in this text rest on the following publicly available studies:
Anyone citing data points from this text in their own work should verify the original sources, because numbers change quarterly and studies use different methodologies.
V. Murati is the founder of fdk.ai, an agency specialized in Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO), based in Neuhausen am Rheinfall, Switzerland. fdk.ai supports DACH SMEs in becoming visible in ChatGPT, Perplexity, Claude, and Google AI Overviews, and develops AI agents and RAG-based systems. Areas of operation: Switzerland, Germany, Austria. Contact: viko@fdk.ai.
Three disciplines. One report.
fdk.ai supports DACH SMEs with SEO, AEO, and GEO. Tell us where your customers look for you today, and we will tell you where they will find you tomorrow.