Frequently asked questions about GEO and AI visibility

GEO is increasing the likelihood that AI assistants such as ChatGPT, Gemini, Perplexity and Google AI (AI Overviews and AI Mode) mention your brand and cite your website as a source. You will also come across the terms AEO (Answer Engine Optimization), LLM optimization and AI search optimization; the goal is the same.

It matters because customers increasingly ask AI a complete question ("what is a good accounting app for freelancers?") and immediately get a short selection of brands back. A brand that is not in that answer is skipped before the customer ever visits your website. According to McKinsey (October 2025), brands that do not optimize for AI search can lose 20-50% of their organic traffic.

FoundByAI is a Dutch GEO agency based in Rotterdam. We measure how often a brand is mentioned and cited in ChatGPT, Gemini, Perplexity, Google AI Overviews and Google AI Mode, and increase that visibility through targeted optimization of content and sources. We were founded in 2026 by Cas Brulot (data, generative AI and marketing) and Daan Hellinga (digital marketing and GEO).

The idea took shape during our traineeship, where we met, first on the football pitch and later mostly off it. We worked alongside experienced marketers with large budgets and the latest AI marketing tools, and saw that almost nobody had a clear picture of how ChatGPT and Gemini decide which brands appear in their answers. We wanted to make that measurable and practical. GEO is often presented as more complicated than it needs to be; many of its principles are easy to understand and apply, and we show how.

GEO suits any brand whose customers ask AI a question while exploring, comparing or buying. That happens in two ways.

For brands with high prompt volumes, such as FMCG brands, many consumers ask similar questions again and again ("what is the tastiest mayonnaise?", "which accounting app is good for freelancers?"). Every place in such an answer reaches a large audience. For brands with low prompt volumes but high lead value, such as a yacht builder, only a handful of buyers ask the question, but every inquiry is worth thousands or even millions of euros. One mention in response to the right question can then be worth more than thousands of mentions in a mass market.

Whether your category already receives such questions is something we test in a first conversation, using a few relevant prompts.

SEO improves your position in the results of Google and Bing. GEO improves whether and how your brand appears in the answer of an AI system. The two overlap: Google AI Overviews and AI Mode draw on the regular Google index, so a technically sound, high-quality website remains the foundation.

The difference lies in the sources and in what you measure. According to McKinsey, your own website determines only 5-10% of the sources AI uses; the rest comes from reviews, communities, video and trade publications. SEO measures rankings and clicks; GEO measures mention rate, citations, position within the answer and share of voice.

First citations often appear within days to weeks; stable visibility takes months. External measurements from GEO vendors show that range: Goodie saw new or refreshed content appear in AI citations within 2 to 4 days, AirOps reports first citations in ChatGPT and Perplexity within 2 to 4 weeks and in Google AI Overviews after 8 to 12 weeks, and Goodie expects 4 to 6 months for stable visibility across multiple platforms. How fast it goes depends on the platform, the authority of the source and the competition.

In our own first pilot, with a consumer brand measured across 22 prompts, the share of AI answers mentioning the brand doubled within three months and was more than 3.5 times as high after five. Results differ by brand, category and platform; that is why we measure at several points in time and keep the underlying answers.

GEO delivers two kinds of value: visibility you can measure, and visitors who convert better than average.

The visibility gain can be large. In our first pilot, the brand was mentioned 3.6 times as often in AI answers. That counts even when nobody clicks through: Semrush reports that users click a regular search result only 8% of the time when an AI summary is shown. The answer itself is increasingly the touchpoint where a brand builds recognition.

The traffic that does click through converts better. In June 2025, Semrush (500+ topics) found that visitors from AI search engines are on average 4.4 times as valuable as visitors from regular organic search results. Knotch measures a smaller difference, about 2 times, in the Conductor benchmark report covering 13,770 domains (via Mintec), and Seer Interactive saw ChatGPT visitors convert at 15.9% versus 1.76% for Google organic (via an overview of AI SEO statistics).

With every engagement we record a baseline measurement and re-measure the same prompts, so you can see the effect for your own brand.

Because search behavior is shifting from traditional search engines to AI. Users are in the middle of switching: according to Semrush, Google AI Overviews reaches 2 billion users a month, and Semrush expects AI search traffic could overtake traffic from traditional search engines in 2028. McKinsey (October 2025) estimates that consumers will spend around $750 billion through AI search engines in 2028, while only 16% of brands systematically measure whether they appear in AI answers.

The shift goes beyond a trend. According to Profound research (2025), 88% of users worldwide are expected to prefer AI search over traditional search by 2029. By then you do not want to be hearing about GEO for the first time; you want to have been in the answers for years.

The sources AI relies on, such as well-optimized pages, reviews, Reddit threads and trade publications, cannot be built in a week. They grow in authority and citation strength the longer they have been in place. Those who lay the groundwork now will have that foundation ready by the time the majority of users have switched, and will reap the rewards for years. Those who only start once competitors already hold a strong position in the answers have to close a gap while the choice is already being made by the AI. Make sure you are there before the shift happens, and prepare for the future of search.

No, and no agency can. AI models are probabilistic, not deterministic: the same model can give a different answer to the same question at a different moment, and answers differ by platform, wording and user. Nobody controls which brands a model mentions, so an agency that guarantees a mention is promising something it cannot back up.

What we can do: measure where you stand today, optimize in a targeted way on the sources AI uses, and then test the same prompts again. That increases the likelihood of a mention and shows whether it worked.

We measure the same prompts on five platforms: ChatGPT, Gemini, Perplexity, Google AI Overviews and Google AI Mode, and we report on each platform separately. Each platform chooses different sources: a brand can be mentioned often in Google AI Mode and not at all in ChatGPT. A single average hides where you need to step in. You can read what works per platform on our platform pages.

Together with you, we put together a fixed set of prompts: the questions customers ask an AI assistant while exploring, comparing and buying. We run them on five platforms in clean sessions without personalization and record every full answer. From that come five measures: mention rate (the percentage of answers in which your brand appears), position within the answer, share of voice relative to competitors, citations per domain and sentiment.

You can trace every score in the dashboard back to the exact prompt, platform and answer. You can read the full method in How do you measure your brand's AI visibility?.

We combine measuring with doing. Every engagement starts with a baseline measurement: which prompts you appear on, which competitors are mentioned more often and which sources AI uses for your category. Based on that we draw up an opportunity plan, prioritizing prompts and sources by impact and effort. We then carry out the optimizations ourselves, re-measure the same prompts and discuss in a performance review what has changed. After an engagement you can choose a monthly retainer with ongoing monitoring and new optimizations.

Which optimizations we use depends on what is decisive in your answers. Two directions come up often. These are examples, not a fixed list:

• Your own website and domain: setting up the technical foundation so AI crawlers can read and understand your site well (crawlability), and writing content that answers a question directly: FAQs, blogs, product pages (PDPs), listicles, metadata and schema markup.
• External, organic third-party content: the places AI draws its answers from and that you do not control yourself, such as Reddit, Quora, publishers, review sites, bloggers and app stores (App Store and Google Play).

For each brand we choose the mix with the greatest effect. For one client the gain lies in rewriting product pages; for another, in a presence on review sites or in community threads AI already cites.

In a sprint of 8 weeks, followed by monthly maintenance:

• Week 1, focus: sharpen target audiences, key questions and competitors.
• Weeks 1-2, baseline measurement: record visibility, sentiment and the AI sources used.
• Weeks 2-3, opportunity plan: choose the prompts and sources with the most to gain.
• Weeks 3-7, optimize: improve your website, content and external sources in a targeted way.
• Week 8, measure and adjust: re-measure the same prompts and determine next steps.

After that we maintain your lead with monthly optimizations and reporting.

Yes. In every measurement you see which competitors appear in the same AI answers, in which position and with which sources. The dashboard shows your share of voice: how often you and your competitors are mentioned. That shows you where competitors are winning, which sources help them and on which prompts your opportunity is greatest. In the retainer we track your competitors' performance every month.

Then we put together a tailored scope. We work with four packages (Lite, Core, Pro and a monthly retainer, see the pricing page); the variables are the number of prompts, the number of optimizations, the channels (own site, Reddit, YouTube, Medium, app stores, external authority), the market and the language. Describe your situation in a first conversation and we will tell you honestly whether one of the packages fits or whether custom work is needed.

Yes, the first audit is free and non-binding. In a first conversation we test a number of relevant prompts right away and show how your brand and your competitors appear in AI answers. Book a conversation.

Four things:

• Software and execution in one: many GEO providers offer either software that shows visibility or consultancy. We measure with our own software and carry out the optimizations ourselves.
• Every score is traceable: for each prompt you see the platform, the full AI answer, the position, the sources, the competitors, the sentiment and the date.
• Measurement with a control group: we compare optimized prompts with prompts we leave untouched. That makes an increase demonstrably attributable to the work we delivered, and not to chance or a changing market. It means you can see the value of our service directly in the numbers.
• Experience at one of the largest FMCG companies: we started out as marketers at Unilever, in an environment with experienced marketers, large budgets and the latest marketing tools. We apply that knowledge and way of working for brands of every size.

There are now dozens of agencies in the Netherlands offering GEO, and their approaches differ widely. These are some of the parties you will come across, described as they position themselves:

• FoundByAI (Rotterdam): measures with its own software on five AI platforms and carries out the optimizations itself, with a baseline measurement, a control group and a follow-up measurement, so every result is traceable to the work delivered.
• Citeerbaar (Diemen): GEO specialist for SMEs, hospitality, events and webshops, focused on schema markup, semantic structure and access for AI crawlers.
• Clarius AI ('s-Hertogenbosch and Nijmegen): GEO and AI visibility agency with a six-step approach, from baseline measurement and technical setup (entities and structured data) to content creation, authority building and monitoring through its own dashboard.
• Searchlab (Amsterdam): SEO, Google Ads and GEO agency, with AI visibility as a separate line of delivery.
• Seeders (Zwolle): large international SEO and GEO agency with its own GEO audit and tracking tools, active in more than 50 countries.
• SmartRanking (Groningen and Mijdrecht): SEO agency that launched GEO as a service in January 2025.

The best choice depends on what you need. When comparing, look at five points: does the agency first run a baseline measurement and repeat it? Can you trace every score back to the underlying AI answer? Does it use a control group to demonstrate effect? Does it report per platform instead of a single average? And does it carry out the optimizations too, or only provide software or advice? FoundByAI combines these points and is a young agency (2026); ask every agency, us included, for demonstrable results.

Yes. We measure in the market and language of your target audience. Our first pilot was measured in the US market, and the site is also available in English. You can read which sources AI cites in the Netherlands on our ChatGPT page.