For the past two decades, a homeowner thinking about selling would open Google, type something like "estate agents near me" or "best estate agent in Birmingham", and scan a list of results. That behaviour is changing.
A growing number of people are starting the same research in ChatGPT, Perplexity, or Google's AI Overviews instead. They are not typing short keyword strings. They are asking questions the way they would ask a person. "Which estate agents are well regarded in Solihull?" "What should I look for when choosing an agent to sell my home?" "How long does selling a house typically take?"
The AI does not return a list of links. It synthesises an answer and, where relevant, names specific agencies, cites specific sources, and makes specific recommendations. The homeowner reads the answer and forms an impression before they visit a single website.
Only 22% of buyers say they have used AI in their property searches, according to GetAgent CEO Colby Short, writing in March 2026. That figure is low by global standards but it is moving fast. AI chatbot traffic grew 80% year on year from April 2024 to March 2025 while traditional search engine traffic declined over the same period. The direction of travel is clear.
How AI search works differently
Traditional search engines rank pages based on keywords, backlinks, and user behaviour signals. When you rank highly, your link appears in a list. The user decides whether to click.
Generative AI search engines work differently. When someone asks an AI a question, the AI does not simply return links. It reads content from sources it considers authoritative, synthesises that content into a coherent answer, and presents that answer directly. The user does not see a list of options. They see a response.
This creates a fundamentally different visibility problem. In traditional search, being on page two means fewer clicks but some visibility. In AI search, if you are not cited in the answer, you do not exist for that query. There is no page two. There is the answer and there is absence.
The discipline that has emerged to address this is called Generative Engine Optimisation, or GEO. Independent tracking by Searchable.com in May 2026 found that UK estate agents were cited in zero out of three responses across ChatGPT, Perplexity and Google AI Overviews when prompted for property recommendations, described as the only vertical actively losing share in AI search. The agencies being recommended in those answers were not the best performing agencies in the UK. They were the agencies with the strongest structured digital presence.
What UK proptech is doing
The UK proptech market is beginning to respond to the AI search shift, though the responses are uneven and the framing varies considerably.
At the property search layer, portals and newer entrants are integrating conversational AI interfaces into property discovery. Jitty, a UK proptech firm, launched an AI search engine that analyses conversational search requests rather than relying on rigid tick-box filters, and offers agents guidance on writing property descriptions optimised for AI search queries. The framing is that natural language search behaviour requires different listing content from keyword-optimised descriptions.
At the data layer, platforms like Propalt AI are building structured UK property data infrastructure designed to be usable by AI tools directly: valuations, ownership records, listings and rental history for 29 million addresses, refreshed daily, with integrations into ChatGPT and Claude. The bet is that AI search will increasingly pull from structured, verified data sources rather than unstructured web content.
At the agency marketing layer, a small number of UK-focused services have emerged specifically to help estate agencies improve their visibility in AI-generated answers. Property Wave, Estate Apps and similar providers are offering what they describe as GEO services for estate agents, focusing on structured data implementation, schema markup, and content architecture designed for AI citation.
Rex, an estate agency software provider, predicted that 2026 would be the year AI moves from experimentation to delivery in the sector, with high-performing agencies using technology to strengthen rather than replace human connection.
Where most estate agents actually stand
The gap between what the most forward-looking proptech companies are building and where the average UK estate agency currently sits is significant.
Most independent and mid-sized agencies have websites that were built for traditional search. They rely on keyword-optimised content, Google My Business profiles, and portal presence to generate inbound enquiries. These assets remain valuable for traditional search. They do not automatically transfer to AI search visibility.
TendorAI's analysis of over 12,000 UK regulated firms in May 2026 found that the estate agencies most likely to appear in AI-generated recommendations shared three characteristics: verified membership of a recognised redress scheme visible on their website, structured schema markup on branch pages, and a consistent recent stream of verified reviews on third-party platforms. These are not complex or expensive requirements. They are largely structural. And most agencies have not implemented them.
The reason is not ignorance of the trend. It is that the consequences of absence from AI search are not yet visible in agency reporting. An agency that does not appear in a ChatGPT answer does not receive a notification. The homeowner who was directed to a competitor by an AI recommendation does not leave a trace in any CRM. The loss is structurally invisible, in the same way that the instruction not won from a mishandled inbound enquiry is structurally invisible.
What AI search engines actually cite
Research into what drives AI citation in local service categories, including estate agency, points to a consistent set of factors.
Third-party verification matters more than self-described credentials. Reviews on Google, Trustpilot, and recognised estate agency platforms carry significantly more weight than testimonials on an agency's own website, because AI systems can cross-reference external platforms but cannot verify self-published content. TendorAI's guidance suggests that five recent reviews from the past six months carry more weight than fifty older reviews. AI systems weight recency heavily.
Structured data enables AI comprehension. Schema markup that identifies a business as an estate agent, specifies its service area, lists its branches, and connects it to verified accreditation schemes gives AI systems the structured information they need to include a firm in a relevant answer. Without it, the AI must infer this information from unstructured content, which it does less reliably.
Content depth signals authority. Property firms that release market insights or appear regularly in the trade press tend to see AI citation first, according to Net Yield's analysis of generative search behaviour in property. GEO rewards clarity and authenticity rather than marketing spend. A consistent body of accurate, well-sourced content on topics relevant to sellers and buyers in a specific geography is more likely to be cited than general marketing copy.
The implication for agencies
Traditional search remains the dominant channel for property-related queries in the UK in 2026. But Gartner projected a 25% decline in traditional search engine volume by 2026 as queries shift to conversational interfaces, and the traffic data supports the direction if not the precise figure.
AI systems build citation patterns from sources they learn to trust. Agencies that establish a structured, verified, content-rich presence now are the agencies that will appear in the answer when AI-mediated property searches reach the volume that makes absence genuinely costly. That moment has not arrived yet. The agencies that act before it does will be in the answer when it comes.