AI Engines · 9 min read · July 15, 2026
AI visibility for architects: Why clients need to find you in ChatGPT
More and more clients no longer ask Google, but ChatGPT: "Which architecture firm fits my single-family house in Regensburg?" The AI answers with concrete names. If you're not among them, you simply don't exist for this client. AI visibility increasingly determines which architects even make the shortlist, long before anyone opens your portfolio.
How clients really search today
A client planning a house used to google: "architect Passau single-family house". Today the same person types the question into ChatGPT, Gemini or Perplexity and phrases it far more concretely: "I have a sloping plot, want KfW-40 standard, budget around 600,000 euros. Which architecture firms in Eastern Bavaria specialize in sustainable building?" The AI delivers not ten blue links, but three to five names with a brief justification. That fundamentally changes your marketing.
The decisive difference: with Google you had ten results and a chance to still get into play via position four. With ChatGPT there's often only one answer with three recommendations. Whoever isn't named doesn't even come to the client's mind. There's no second page to browse further. Visibility has become binary: in or out.
For architects this is especially charged, because building decisions are long-term and expensive. A client often invests the largest sum of their life. They research for weeks, and the AI is their first advisor today. If that advisor doesn't know you, you lose the project before you even learn of its existence. This is exactly where AI visibility comes in.
What Generative Engine Optimization means for you
Generative Engine Optimization, GEO for short, is the art of appearing in the answers of generative AI systems. With classic SEO you optimize for rankings, for clicks and for the first page. With GEO you optimize so that a language model recognizes, understands and recommends your firm as a fitting answer. It's no longer about placement in a list, but about being mentioned in a sentence.
Language models draw their knowledge from what's said about you on the web: your website, professional portals like Houzz or Baunetz, chamber-of-architects directories, press articles, competition databases and reviews. The clearer, more consistent and more frequent your profile appears there, the more confidently the AI links your name with terms like "heritage conservation", "passive house" or "commercial construction in Munich". GEO means: building these links deliberately.
Important to understand: you can't bribe a language model and can't buy an ad that lifts you into the answer. You can only ensure that the factual basis is right. That's good news and a task at once. Whoever speaks cleanly and specifically about their work is rewarded. Whoever shows only beautiful pictures without context stays invisible to the machine.
Why architects have a particular problem
Architect websites are often a prime example of what AI can't read. They're visually brilliant, but textually mute. Large photo series, minimalist menus, project names like "House K" or "Villa by the Lake" without location, without building type, without square meters. To a human viewer that seems refined. To a language model it's a black box from which not a single useful fact can be drawn.
Added to that is the reticence of the profession. Many architects find active marketing disreputable and rely on referrals. That worked as long as referrals passed from person to person. Now the first referral runs via the AI, and it only knows what's documented. The quiet, excellent firm without a digital trace loses to the mediocre firm that describes its projects well.
A concrete example: two firms have both built an award-winning school building. Firm A writes "New primary school, timber hybrid construction, 2,400 square meters, DGNB Gold, Rosenheim district". Firm B writes "educational building, 2023". If someone asks the AI about sustainable school buildings in Upper Bavaria, Firm A is named. Firm B doesn't exist for this question, even though the building is of equal value.
The questions you want to be found for
The first practical step is banal and still rarely taken: sit down and write out the real questions of your ideal clients. "Who plans barrier-free conversions for seniors in Stuttgart?" "Which architect has experience with heritage-protected half-timbered houses?" "I want an office building made of wood, who do I ask in North Rhine-Westphalia?" Each of these questions is a stage on which you could stand.
Then test these questions yourself in ChatGPT, Perplexity and Gemini. Are you named? Is a competitor named? Is no one named, or is something false claimed? This snapshot is your starting position. It's often sobering, but it shows you exactly where the gap yawns between your real competence and your digital image.
From these questions you derive your topics. If you've done barrier-free building ten times but the word doesn't appear on your website, that's the first construction site. The AI can only reproduce what's stated somewhere. Your task is to systematically close the bridge between what you can do and what's documented.
Making your project pages readable for machines
Every project needs running text, not just pictures. For each reference, describe the hard facts: building type, location or region, year of construction, area, construction method, energy standard, special features and the task you solved. Write it the way a client would ask. Instead of "residential house, 2022", rather "new construction of a single-family house in Freising, 180 square meters, solid timber construction, Efficiency House 40, with a granny flat for the grandparents".
These texts help twice over. They feed the AI with the facts it needs to recommend you, and they help real clients recognize themselves in your projects. A prospect with a sloping plot is looking for someone who has already built on a slope. If the word "sloping location" appears with you, they find you and the machine links you with the topic. If not, neither happens.
Add structured data and a clear page structure. A well-maintained services page per focus area, a meaningful firm profile with location and specialization, consistent details across all channels. It's not about keyword-stuffing, but about clarity. Write honestly and precisely, then the machine reads you correctly and the recommendation falls in your favor.
Mentions beyond your own website
Language models don't trust only your own page; they often weight independent sources more highly. That's why what others write about you counts. A complete profile in the chamber-of-architects database, entries on Baunetz, Competitionline or Houzz, a report in the local paper about your new daycare project, a specialist article on a construction blog. Each of these sources is a thread that links your name with your competence.
Competitions and awards are worth gold here, because they land in databases that AI systems read out. When you win a competition or receive a commendation, make sure it's documented and findable. An award that only you know about in the drawer doesn't help your visibility. An award that appears on three portals makes you the obvious recommendation for the machine.
Pay attention to consistency. If your firm is sometimes "Müller Architects", sometimes "Architecture firm Müller + Partner" and sometimes "M. Architecture", your profile splinters. The AI can't be sure it's the same company and distributes your reputation across three weak identities. A uniform name, a uniform address and a uniform description across all platforms bundle your visibility.
Show expertise instead of merely presenting projects
Clients have questions long before they commission an architect. What does an extension roughly cost? Is a passive house worth it? How does an approval procedure for a plot in an outlying area work? If you answer these questions in honest, helpful texts, you become the source the AI draws from. Sharing expertise publicly is perhaps the most effective GEO strategy for architects at all.
This need not be a novel. A well-written guide on your focus area, a blog post about a detail of your latest renovation, an honest explanation of why timber construction is the better choice on certain plots. Such content positions you as an expert on a topic and gives the AI exactly the context sentences it needs to name you for a fitting question.
The side effect is trust. A client who comes across your specialist article via the AI and finds it helpful enters the first conversation with a head start of trust. You didn't tout yourself, you helped. In a profession that views advertising skeptically, that's the most elegant path to visibility: showing competence instead of claiming competence.
Staying on it and measuring the effect
AI visibility is not a one-off project, but a habit. New answers arise continuously, models are updated, competitors catch up. Resolve to document every new project properly right away, and check every few months whether the AI names you for your core questions. That way you notice whether your work is having an effect and can readjust before a gap becomes a problem.
Measure the effect with simple means. Ask new clients how they came to you. If "ChatGPT recommended you" or "I asked an AI" comes up more and more often, you're on the right track. Regularly test the same questions and note whether and how you're mentioned. This small routine replaces expensive tools and gives you an honest feel for your position.
The competition for AI visibility has only just started, and most architecture firms haven't reacted yet. That's precisely your chance. Whoever sharpens their digital profile now, describes their projects clearly and shares their knowledge secures a lead that later stragglers can only close with difficulty. In five years AI visibility will be a matter of course. Today it's an advantage.
Common questions
I've lived off referrals for 20 years. Do I even need AI visibility?
Referrals remain important, but they increasingly run via the AI. Even someone who hears your name from an acquaintance often checks it today in ChatGPT or Google. If the AI finds nothing substantial there about your work, that weakens the referral. Conversely, a clear digital profile reinforces every personal recommendation. AI visibility doesn't replace referrals, it supports and extends them.
Do I have to clutter my beautiful, minimalist website with text?
No. Your visual language can stay as it is. It's about adding readable facts behind or beneath the images: project descriptions, locations, building types, focus areas. Much of this can be placed elegantly without disturbing the aesthetic, for instance in project captions, a well-written about page or a guide section. Clarity and design aren't mutually exclusive, they complement each other.
How long does it take until I appear in AI answers?
That can't be said to the day, because language models update their knowledge in waves. Roughly: new content on your website and in professional portals takes a few weeks to months to show up in AI answers. Competition entries and press articles often work faster, because they count as trustworthy sources. What matters is the early start and consistency, not a single perfect text.
Read on