Brand & Positioning · 9 min read · July 15, 2026

Timber construction, heritage protection, passive house: why specialisation is your AI signal

When a client asks ChatGPT "Who can plan me a passive house in timber construction?", it isn't the largest firm that wins, but the most unambiguous one. Language models reward specialisation because they need clarity: a firm that stands for timber construction, heritage protection or energy efficiency gives the AI a clean signal. That very signal decides whether you show up in the answer or not.

Why the AI overlooks generalists

Ask ChatGPT yourself: "Recommend me an architecture firm for renovating a heritage-listed timber-framed house in Franconia." You won't get ten names, but two or three – namely the ones whose profile clearly points to heritage protection. A firm whose website says "We plan homes, commercial and public buildings and advise on everything to do with building" gives the model nothing to hold on to. The AI reads breadth as arbitrariness.

Language models work with probabilities. They ask internally: which firm fits this exact query most unambiguously? A generalist is only weakly relevant to every question, a specialist extremely strong on their question. In practice that means: the firm with a clear timber-construction focus beats the three-times-larger all-round firm on a timber query. Not because it plans better, but because its signal is cleaner.

That's the central shift compared to Google. On Google you could rank a bit for everything with reach and lots of subpages. In generative search the AI decides who it even names – and it would rather name three unambiguous firms than thirty blurry ones. Whoever wants to be everything is made into nobody by the machine.

What Generative Engine Optimization means for architects

Generative Engine Optimization, GEO for short, is the art of showing up in the answers of ChatGPT, Perplexity, Gemini and Google AI Overviews. For architects this is no longer a sideshow. More and more clients, investors and even municipalities start their search for a planner not on Google but with an AI assistant. They describe their project in natural language and get firms suggested – often before they open a single website.

The difference from classic SEO is fundamental. With SEO you optimise for keywords and hope for clicks. With GEO you optimise so that the AI understands, categorises and quotes your firm. It's no longer about third place in a list, but about being part of the answer at all. And the AI preferentially quotes sources that assert one thing clearly and verifiably.

For you as an architect that means: your expertise, which you may have treated as a given so far, becomes the most important raw material. Passive-house certifications, references in timber construction, experience with heritage authorities – all of this is no longer nice-to-have extra info, but the signals from which the machine builds its recommendation.

Specialisation is not a sacrifice but an amplifier

Many firms shy away from specialising because they fear losing commissions. The thought: if I position myself as a timber firm, no one will call about the concrete office building any more. In practice the opposite is true. A clearly positioned firm gets asked more often and more precisely, because clients with matching projects find it in the first place – including via the AI.

Think of a real example: a southern German firm with five employees consistently specialised in multi-storey timber construction. Not single-family timber houses, but daycare, school and residential buildings in timber-hybrid construction. Exactly this sharpness ensures the AI suggests the firm for municipal queries about sustainable multi-storey building. The narrow focus is the reason for the reach, not its enemy.

Specialisation also doesn't mean you may only do one thing. You can absolutely have a main focus and two or three adjacent competencies – say timber construction plus passive house plus energy-efficient renovation. What matters is that these fields relate to one another and form a recognisable narrative. Three related focuses are a profile. Twelve unconnected ones are noise.

The three signals the AI reads with architects

For the AI to recognise you as a specialist, it needs three things. First: consistent language. If you do timber construction, the word timber construction has to appear on your homepage, in your project descriptions, in your blog and around your imprint – not as keyword stuffing, but because you genuinely write about it. The AI builds a stable link between your name and the topic from this repetition.

Second: verifiable references. A sentence like "We build passive houses" is weaker than "Certified passive house in Regensburg, 2023, 142 square metres, KfW-40 standard, planning and construction management by our firm". Concrete, verifiable projects with location, year, standard and figures are gold for the AI. It loves facts it can quote without embarrassing itself. Vague self-praise it ignores.

Third: external confirmation. If your firm is named in trade media, architecture portals, competition lists or municipal websites in connection with your focus, it reinforces your signal enormously. The AI trusts statements more when they don't come only from you. A heritage-protection award, a mention in a construction magazine or a talk at a timber-construction conference weighs more than any advertising copy.

Make your website machine-readable

The most beautiful architect's website is useless to the AI if it consists of large images and three words. Many firms present their projects as pure image galleries without text. To the client's eye that looks elegant, to the machine it's invisible. Every project needs a description in real sentences: what was the task, which construction method, which particularity, which standard, which result.

Use structured data wherever you can. Clean schema markup for your company with location, service area and projects helps the AI categorise you correctly. Equally important are clear headings that contain the technical terms of your niche. Instead of "Our work", write "Heritage-appropriate renovation of historic building fabric". The heading is a strong ordering signal for the machine.

And think of the questions clients really ask. Build text sections that answer concrete questions from your niche: What does a passive house cost per square metre? How long does a heritage-protection permit take? Can you energetically renovate a timber-framed house without destroying its character? If your page answers these questions cleanly, it becomes the source the AI draws from.

Three niches, three signals: timber, heritage, passive house

Take timber construction. A firm strong here should write about fire protection in timber construction, about timber-hybrid structures, about the embodied energy of building materials and about the permitting practice for multi-storey timber buildings. These topics are unambiguous markers for the AI. Whoever occupies them becomes the obvious recommendation for "sustainable building with timber" – even against big names.

With heritage protection, proximity to the authority and to the building fabric counts. Write about dealing with the local heritage authority, about reversible interventions, about the balancing act between requirements and modern use, about typical funding. Owners of a listed building have very specific fears – if your content addresses exactly these fears, the AI recognises you as the point of contact for this delicate field.

With passive houses, the numbers rule. Heating demand, airtightness, freedom from thermal bridges, certification routes via the Passive House Institute. The more precisely you document these figures in your references, the more robust your signal becomes. A firm that shows its completed projects with real energy figures is infinitely more credible to the AI than one that just throws "energy-efficient" into the room.

The most common mistake: hiding your positioning

The most expensive mistake isn't missing specialisation, but hidden specialisation. Many firms are in fact specialists – they do eighty per cent timber construction – but communicate like generalists, out of fear of excluding someone. The website then reads a neutral "Architecture for people and space", and the actual strength hides in project number seven. The AI doesn't find this strength because it doesn't come to the surface.

A second contradiction: firms that are strong on substance but silent to the outside. They give talks, sit on expert juries, know every standard – but none of that is online. For the AI, only what is findable and textually graspable exists. Your head knowledge is worthless to the machine as long as it isn't in sentences on an indexable page.

Check yourself honestly: if you paste your own firm text into ChatGPT and ask "What is this firm specialised in?", what does the AI answer? If the answer is vague or reads "in many areas", you've found your problem. The machine mirrors back to you exactly how unclear your signal is.

Five steps to a clear AI signal

First: decide. Choose one main focus and at most two adjacent fields. Write this decision in the most visible place on your website – in the first sentence of the homepage. Second: rework your project pages so that every reference project is described with real text, location, year, construction method and figures. Images alone aren't enough.

Third: build a knowledge section that answers the concrete questions of your niche – honestly, expertly, without marketing chatter. Fourth: ensure external mentions. A trade article, an interview in a construction magazine, an entry in an architect directory with the correct focus. Every independent mention of your specialisation measurably reinforces your signal.

Fifth: check regularly. Ask ChatGPT and Perplexity every few months for firms in your focus and region. Do you show up? Are you categorised correctly? This self-check costs ten minutes and shows you in black and white whether your specialisation lands as a signal. GEO isn't a one-off project but a habit – and for specialised architects the most effective lever for being found in the AI world.

Common questions

Do I lose commissions if I publicly commit as an architect to timber construction or heritage protection?

As a rule, no – on the contrary. A clear focus ensures that clients with matching projects find you in the first place, especially via AI assistants. You may still take on other projects. All that matters is that your communication has a recognisable core competence. A blurry all-round profile is more likely to lead to the AI not naming you for any query.

Is it enough if my project images are good, or does the AI need text?

The AI needs text. Pure image galleries, as many architect sites use, are almost invisible to language models. Every reference project should have a description in real sentences: task, construction method, standard, location, year and particularity. Only this text makes your skill readable and quotable for the machine. The images convince the human, the text convinces the AI.

How do I check whether my firm even shows up in ChatGPT or Perplexity?

Ask the AI assistants the questions your desired clients would ask – for example firms for passive houses in your region. Check whether you get named and categorised correctly. Additionally paste in your own website text and ask the AI what your firm is specialised in. If the answer is vague, your signal is too unclear and you should sharpen your positioning.

Read on

Content & Answer Pages