We scored 10 of the UK's biggest estate agents for AI readiness. None came close to passing.

When a buyer asks ChatGPT to recommend a good estate agent in their area, the answer doesn't come from a Rightmove listing or a glossy window card. It comes from whatever the model can read, parse, and trust on the open web. We scanned 10 of the UK's biggest residential estate agents to see how ready their websites are for that kind of discovery. The headline finding: not a single agent scored above 63 out of 100.

The numbers

Ten agents. Average score 54 out of 100. The highest score was 63 (Fine & Country). The lowest was 43 (Jackson-Stops). Five agents scored below 55. None reached the "good" band of 70 or above.

For context, our scoring rubric awards points for the fundamentals of AI discoverability: crawlable structure, schema markup, clear factual content, and signals that help LLMs understand who you are and what you do. A score in the mid-50s means a site is doing the basics but is missing most of the signals that drive AI citations.

The full ranking

Why this matters for estate agents right now

The property market is one of the highest-intent search categories on the internet. A buyer asking "which London estate agent specialises in period homes under £1.5m?" or a seller asking "who are the best agents in SW6?" wants a name, not a list of links to scroll through. Generative search is increasingly delivering exactly that: a direct answer with one or two agent names.

Whether your firm gets cited in those answers depends on how much useful, structured, trustworthy information about your business the model can find on the open web. Right now, the sector is leaving a significant amount of that visibility on the table.

How we scored them

Every site is run through the same 34-check rubric, grouped into five categories: crawler access, structured data, content clarity, AI-specific signals, and authority & trust. Each check is weighted by how much it actually moves the needle for AI retrieval today. The rubric is published in full on the methodology page.

The full ranking, every individual check, and per-site scorecards live on the benchmark page. What follows is the editorial cut: the patterns that stood out.

Three patterns that hold across the sector

1. Content quality is a genuine strength - but schema is not.

The good news: most agents in this set write usable, factual copy. Homepages explain what the firm does and where it operates. Content broadly answers the kinds of questions buyers and sellers actually ask. Fine & Country, Savills, Foxtons, Hamptons, and Purplebricks all earned credit for content that "answers real user queries."

The bad news: that content is largely invisible to structured data parsers. Fine & Country - the top scorer - is still missing Organization schema markup. Purplebricks is missing it too. Savills lacks complete Open Graph tags and a logical heading structure. When an LLM tries to understand who a business is and what it offers, schema is the shorthand. Without it, the model is reverse-engineering your identity from prose.

2. AI bot policy is an afterthought for most of the sector.

Foxtons and Hamptons - two of the higher scorers - both flag the same priority improvement: no explicit stance on AI training bots, and retrieval and training bots treated identically. This is one of the most common AI readiness gaps we see across every sector, and it matters more than it might look.

Most operators want to allow retrieval bots (the bots that power real-time AI answers in ChatGPT, Perplexity, and similar) while being more deliberate about training bots (which scrape content to build future models). The robots.txt mechanism to express that distinction exists. Almost nobody in this sector is using it. That's a policy decision masquerading as a technical gap.

3. The brand-recognition gap doesn't map to AI readiness.

The most recognisable brands in the sector are not the most AI-ready ones. Knight Frank, one of the world's most prestigious estate agency brands, scored 51 - below Foxtons (58), Hamptons (57), and Purplebricks (55). Chestertons (50) and Strutt & Parker (54) both trail Fine & Country despite comparable or greater brand equity.

This pattern holds in every sector we've benchmarked: brand heritage does not transfer to AI discoverability. An LLM can't read a reputation. It reads structured signals, and on that measure the field is surprisingly flat and surprisingly low.

What the top performers get right

Fine & Country (63) and Savills (61) share a few habits the rest of the sector would benefit from copying:

• Their homepages lead with facts. The first paragraph tells you what the firm is, where it operates, and what kind of property it handles. There's no editorial preamble before the useful information.
• AI retrieval bots can reach their content. Clean robots.txt, no legacy rules quietly blocking modern AI user agents.
• Their content structure is closer to answering specific questions than telling a brand story. That's the shift the whole sector needs to make.

What every agent in this set can fix this quarter

Three changes would move the needle materially for every firm in this ranking, none of which require a redesign or a new website:

1. Add Organization schema in JSON-LD. A single structured data block on the homepage that declares your firm's name, type, locations served, and areas of specialism. One-day engineering task. The biggest leverage point on the entire rubric.
2. Take an explicit position on AI training bots in robots.txt. Decide whether you want training bots accessing your content. Then say so, explicitly, in your robots.txt. The industry convention is to allow retrieval bots (GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot) and make a deliberate call on CCBot and similar training crawlers.
3. Fix heading structure and Open Graph. Savills - one of the best performers - is still missing clean H1–H6 hierarchy and complete Open Graph tags on key pages. These are quick wins that improve both AI readability and social sharing in one pass.

One number worth sitting with

The average score across the 10 largest residential estate agents in the UK is 54 out of 100. That is better than the luxury London hotels benchmark we ran earlier this year (average: 48) - but it still means the typical agent website is failing more than half the AI readiness checks we run. In a market where AI-powered search is increasingly the first place buyers and sellers form opinions, that gap has a commercial cost.

See the full ranking

Every score, every check, and every per-site finding is published on the UK estate agents benchmark page. If you work at one of these firms - or compete with them - it is worth ten minutes.