A residential extension. A listed building conversion. A sustainable new build. Architecture is one of the most researched purchases a person will make. That research now starts with AI. When a potential client asks ChatGPT to recommend an architect, the answer comes from structured data. If your practice has no schema markup, you are not part of that conversation.
Architecture is research-intensive by nature. Clients spend weeks or months exploring design possibilities, checking credentials, and comparing practice portfolios before making contact. AI search has become the starting point for that exploration, and the practices it surfaces early shape the entire shortlist.
Hiring an architect is not an impulse decision. It is one of the most considered professional appointments a homeowner or business will make. The project could be a rear extension, a barn conversion, a new-build home, or a commercial fit-out. Whatever the brief, the client will spend significant time in a research phase before they reach out to anyone.
That research used to happen through word of mouth, design magazines, and Google image searches. The process has shifted. Clients now open ChatGPT, Perplexity, or Google AI Overviews and ask questions like "What type of architect do I need for a listed building extension?" or "Which practices in South Manchester specialise in sustainable residential design?"
AI compiles its response from structured data. It looks for practices with clear schema markup that identifies them as architects, outlines their specialisms, confirms their professional credentials, and defines their geographic coverage. Practices without this data are simply absent from those early-stage research conversations. The client never sees them. The shortlist forms without them.
For architecture specifically, this matters more than almost any other profession. The research phase is longer, the decision is more considered, and the client is actively looking for a match between their vision and the practice's expertise. AI is the tool they use to find that match.
Architecture queries in AI search are distinctive. They are not urgent like plumbing or emergency-driven like electrical work. They are exploratory and research-focused. Clients want guidance on what is possible, what it costs, who specialises in their project type, and which practices are credentialed. These are the queries where structured data determines which practices get mentioned:
Notice the pattern. Every one of these queries depends on structured data to produce a useful answer. A beautifully designed website with a portfolio of stunning projects is not enough if the data behind it is invisible to AI. Schema markup is what translates your visual portfolio into machine-readable intelligence.
Architecture practices have a genuine advantage in schema markup because schema.org includes a dedicated Architect type. This is not a workaround or a generic classification. It is a specific schema type built for architecture practices. Combined with the right supporting types, it creates a complete structured profile that AI platforms can read, reference, and recommend.
Your portfolio and project specialisms are your strongest differentiator in AI search. But AI cannot browse a gallery of images. It reads structured data. Service schema is how you translate your design portfolio into something AI can understand, categorise, and recommend to the right client at the right moment.
Architecture is inherently specialism-driven. A practice that excels in residential extensions may have no experience in commercial warehouse conversions. A conservation architect working on listed buildings requires completely different skills to one designing contemporary new builds. Clients understand this, and their AI queries reflect it.
When a client asks "architect experienced in Victorian terrace extensions in Didsbury", AI needs structured data to make a match. If your Service schema includes a detailed entry for period property extensions, and your areaServed covers Didsbury and South Manchester, AI can connect those dots. Without that schema, your years of experience with exactly that type of project remain invisible to the client.
The same applies to every specialism your practice offers. Sustainable design, passive house construction, planning application support, interior design, landscape integration, conservation architecture - each of these should exist as its own Service entry in your schema. The richer and more specific each entry, the more query types AI can match your practice to.
Think of it this way: your portfolio page shows visitors what you can do. Your Service schema tells AI what you can do. Both need to exist. The portfolio converts once someone finds you. The schema is what gets you found in the first place.
RIBA chartered status is one of the most recognisable trust markers in UK architecture. Clients actively search for RIBA chartered practices because the qualification signals professional standards, ongoing development, and accountability. When that search happens through AI, the hasCredential markup becomes critical.
AI platforms cannot read a RIBA logo. They cannot interpret a badge graphic in your website header. What they can read is hasCredential schema that explicitly states your RIBA chartered status, your ARB registration, and any other professional accreditations your practice holds. When a client asks "RIBA architects near me" or "chartered architect for a new build in Cheshire", AI filters its recommendations based on this structured credential data.
The same principle applies to ARB registration. Every practising architect in the UK must be registered with the Architects Registration Board, but AI has no way of knowing that about your practice unless the credential exists in your schema. Including both RIBA and ARB in your hasCredential markup creates a layered trust signal that elevates your practice above competitors who rely on visual badges alone.
A homeowner commissioning an architect may spend three to six months in the research phase before making contact with a single practice. They explore design styles, project feasibility, planning constraints, and budget expectations - all through AI. By the time they reach out, their shortlist is already formed. Schema markup is what determines whether your practice is on that list or excluded from the conversation entirely. In no other profession does the research window last this long, and in no other profession does AI have this much influence over the eventual decision.
The majority of architecture practices in the UK are invisible to AI search, and the reasons are surprisingly consistent. It is rarely a lack of talent, reputation, or even web presence. It is a set of assumptions about how discoverability works that made sense five years ago but no longer hold true in an AI-driven search landscape.
The most common barrier is an over-reliance on visual portfolio work. Architecture is a visual profession, and practices invest heavily in photography, renders, and case study imagery. That investment pays off when a visitor reaches your website. But AI platforms do not browse image galleries. They cannot interpret a beautifully photographed kitchen extension or a rendered section drawing. Without structured data describing the project type, location, and design specialism behind those images, AI has nothing to work with. Your portfolio exists for human visitors. Schema exists for AI. You need both.
A second assumption is that professional body membership handles discoverability. Many practices believe that being listed on the RIBA Find an Architect directory, or appearing in the ARB register, is sufficient for AI platforms to recognise their credentials. It is not. AI search does not crawl third-party directories and cross-reference them with your website. It reads the structured data on your own domain. If your RIBA chartered status is not declared in hasCredential schema on your own site, AI treats your practice as unverified.
There is also a deeply held belief, particularly among established practices, that reputation and referral networks make digital discoverability unnecessary. Word of mouth has always been powerful in architecture, and it still is. But the referral pipeline does not account for the growing number of clients who start with AI. A homeowner planning a loft conversion does not ask friends first any more. They ask ChatGPT. If your practice is absent from that conversation, no amount of reputation compensates for the lost visibility.
Platform limitations play a role too. Many smaller practices use WordPress themes, Squarespace, or Wix to build their websites. These platforms produce visually polished sites, but most themes ship without any schema markup at all. Some include basic Organisation or LocalBusiness schema, but almost none implement the dedicated Architect type, Service entries, or hasCredential markup that AI needs to understand an architecture practice properly. The site looks professional to a visitor but reads as generic to a machine.
Finally, there is a size perception problem. Smaller practices often assume AI search is only relevant for large commercial firms with national reach. The opposite is closer to the truth. AI search is intensely local and specialism-driven. A two-person studio specialising in residential conservation work in a specific borough is exactly the kind of practice AI wants to recommend, provided the data is there. Firm size is irrelevant. Data completeness is everything. The practices that appear in AI results are not necessarily the biggest or the best known. They are the ones whose structured data gives AI something to work with.
When multiple practices serve the same area and claim similar specialisms, AI does not pick one at random. It runs a multi-signal evaluation, comparing the depth, specificity, and credibility of each practice's structured data to determine which ones earn a recommendation and which get filtered out.
The first signal is schema type specificity. A practice using the dedicated Architect type immediately ranks higher for architecture queries than one using the generic ProfessionalService or LocalBusiness type. The Architect schema tells AI exactly what your business is. Generic types force AI to guess, and when it is choosing between a confirmed architect and an ambiguous professional service, the confirmed architect wins every time.
Service entry detail is the second signal, and it is where most competitive differentiation happens. Consider two practices that both serve South Manchester and both offer residential extensions. Practice A has a Service entry that reads "residential design". Practice B has a Service entry that reads "Victorian terrace rear extension with full planning support, including pre-application consultation, party wall guidance, and building regulations compliance for period properties." When a client asks AI for help with their Victorian terrace extension in Chorlton, Practice B is a direct match. Practice A is a vague possibility. AI recommends the specific match.
Credential layering is the third signal. A practice with RIBA chartered status in its hasCredential schema has one trust marker. A practice with RIBA chartered status, ARB registration, and a regional design award has three. AI stacks these credentials when assessing trustworthiness. Each additional verified credential increases the likelihood that AI will recommend your practice over a competitor with fewer. This is not about collecting badges for display. It is about giving AI more structured evidence that your practice meets professional standards.
Geographic granularity matters more than most practices realise. Listing "Manchester" as your areaServed is functional but imprecise. Listing Didsbury, Chorlton, Withington, Sale, Altrincham, Stockport, and Salford as individual areas within your coverage gives AI far more connection points. When a client asks about architects in Didsbury specifically, the practice with Didsbury in its areaServed data appears. The one with just "Manchester" may not.
Review data quality and recency form the final layer. AI treats recent, detailed reviews as stronger signals than older or generic ones. AggregateRating schema that reflects a high score from a meaningful number of reviews tells AI your practice is actively delivering quality work. A practice with fifteen five-star reviews from the last twelve months outweighs one with five reviews from three years ago, even if both show the same average rating.
Walk through the full picture. Two practices both claim to handle residential extensions in South Manchester. Practice A has generic Architect schema, a single Service entry for "residential architecture", areaServed set to "Manchester", no hasCredential markup, and no review schema. Practice B has Architect schema with detailed Service entries for Victorian terrace extensions, new-build residential, and sustainable retrofit, areaServed listing eight specific suburbs, hasCredential for RIBA and ARB, and AggregateRating from twenty-two recent reviews. AI does not see two equal options. It sees one complete, trustworthy profile and one incomplete one. Practice B gets the recommendation. Practice A does not appear at all.
A free AI Visibility Snapshot is the starting point. Schema implementation starts from £295. Monthly monitoring is £79 per month, no contract.
For an architecture practice, consider the value of a single commission. Even a modest residential extension project generates fees of several thousand pounds. A new build or commercial project can be worth significantly more. If schema markup connects you with one additional client enquiry per quarter who would otherwise have gone to a competitor, the return on investment is substantial. Most architecture practices in the UK have not implemented this yet. That gap is an opportunity, but it will not remain open indefinitely as more practices recognise how AI search works.
Get a free report showing how ChatGPT, Google AI Overviews and Perplexity see your architecture practice right now. Discover what prospective clients are being told when they ask AI for an architect in your area.
