Which housebuilders does AI recommend?

October 2026Jack O'Leary

Buying a new-build home is one of the biggest decisions most people make. Buyers spend months on it: which builder, which development, which scheme, and whether the house will be built well and looked after once they've moved in.

More of that research now starts in AI. A buyer can ask ChatGPT, Gemini or Google which builders to consider and who to avoid, and get a shortlist back in seconds, with reasons.

That shortlist forms before a buyer visits a website, books a viewing or speaks to a sales adviser. A builder that isn't on it may never get the chance to make its case.

So we built a UK housebuilder AI visibility index, the same way we track AI visibility for clients. We asked ChatGPT, Gemini and Google's AI Overviews 36 questions real buyers ask, about choosing a builder, first-time buying, build quality and incentives, around ten times each. Across 981 answers, we tracked how often the 50 biggest housebuilders in the UK were recommended.1 Then we set the top five against their Google traffic and their latest results.

The results don't just follow size. The UK's second-biggest builder comes fifth. Bellway builds 15% of the top five's homes but earns 22% of their AI citations. And Taylor Wimpey, second of the five in Google, is fourth for AI citations. Size and search rankings help, but they don't guarantee you'll show up where buyers are looking.

The index

We've included the top 25 below:

AI visibility rank	Housebuilder	Named in (% of answers)	Size rank (by housing turnover)
1	Barratt Redrow	84%	1
2	Bellway	66%	5
3	Taylor Wimpey	62%	3
4	Persimmon	60%	4
5	Vistry	28%	2
6	Bloor Homes	27%	7
7	Miller Homes	24%	10
8	Cala	21%	8
9	Berkeley	15%	6
10	Keepmoat	13%	12
11	Story Homes	10%	17
12	Hill	9%	9
13	Crest Nicholson	9%	13
14	Gleeson Homes	8%	15
15	Avant Homes	6%	14
16	Allison Homes	5%	42
17	Morgan Sindall (Lovell)	5%	11
18	Croudace Homes	4%	36
19	Robertson	4%	38
20	Anwyl	3%	28
21	Hopkins Homes	3%	33
22	Springfield	3%	18
23	Bewley Homes	2%	39
24	Emerson (Jones Homes)	2%	41
25	Castle Green Homes	2%	44

7DOTS analysis, October 2026. Share of 981 homebuyer answers in ChatGPT, Gemini and Google AI Overviews that name each builder, including its sub-brands. Size rank: Building, Top 50 Housebuilders 2025.¹

Four things stand out.

Barratt Redrow dominates. It's named in 84% of answers and named first in about half of them.
There's a big four, then a cliff. Bellway, Taylor Wimpey and Persimmon are each named in 60% to 66% of answers. Nobody else gets above 28%.
On quality, size stops counting. Asked which builders have the best build quality, AI named Bloor in 89% of the 27 answers and put Berkeley first in 63%. Persimmon was named in 37% and Vistry in 19%.
Most of the market is invisible. 39 of the 50 builders are named in fewer than one answer in ten, and five aren't named at all.

Google, AI and the business

AI visibility doesn't sit on its own. We set the top five against two other measures: the Google search traffic their consumer websites earn, and how their sales are moving.

Google and AI don't agree

Builders' own websites matter in both channels. So we compared each group's share of the top five's Google organic traffic with its share of the citations AI gives to their websites.

Housebuilder	Share of top five's homes	Share of Google organic traffic	Share of AI citations to builder websites
Barratt Redrow	27%	41%	37%
Vistry	24% (7.5% open market)	7%	7%
Persimmon	18%	17%	19%
Taylor Wimpey	16%	23%	15%
Bellway	15%	12%	22%

7DOTS analysis, October 2026. Homes: latest full-year completions ^{2-4, 6,7}. Google: estimated UK organic traffic to each group's consumer websites, September 2026, Semrush (Ahrefs gives the same order). AI: citations to each group's websites in 981 homebuyer answers, Peec, 20 October 2026.

Bellway wins more of AI than of Google. It earns 12% of the top five's Google traffic but 22% of AI citations. Its scheme pages do much of that work (see below).
Taylor Wimpey is the reverse. It's second in Google with 23% of the traffic, but fourth in AI with 15% of the citations.
Barratt Redrow, Persimmon and Vistry land in roughly the same place in both.

Ranking in Google helps, but it doesn't carry straight across. AI picks the page that answers the question, and that isn't always the page that ranks.

Set against the business

We also put each group's latest results next to how its Google share has moved over the past year.

Housebuilder	Completions, latest full year	Change on previous year	Google organic share, Sep 2025 to Sep 2026
Barratt Redrow	17,667 (to June 2026²	+5%	36% to 41%
Persimmon	11,905 (to Dec 2025)³	+12%	15% to 17%
Taylor Wimpey	10,614 (to Dec 2025)⁴	+6%	27% to 23%
Bellway	9,695 (to July 2026)⁶	+11%	13% to 12%
Vistry	4,065 open market (to Dec 2025)⁷	-11%	9% to 7%

Each company's latest full-year results, change as reported by the company (see Sources below). Google share: Semrush, UK.

Barratt Redrow and Persimmon are up on every measure. Both grew completions, both gained Google share, and both earn an AI share in line with their size or better.
Vistry is down on every measure. Open-market completions fell 11% in 20257, its Google share slipped, and it reported an adjusted first-half loss of £83.3m in 20268. Its September strategy review repositions it as a specialist mixed-tenure housebuilder.8
Taylor Wimpey is the one to watch. It grew completions in 2025 but lost four points of Google share, and first-half 2026 UK completions (excluding joint ventures) fell 3.5%5. AI already gives it less weight than Google does.
Bellway's growth isn't all buyer demand. Completions rose 11%, but bulk sales doubled to 615 homes6 within that. Its strong AI showing gives the open-market side a head start.

This is side by side, not cause and effect. Results come out once or twice a year, year ends differ, and plenty drives sales besides visibility. But it shows why AI and search visibility are worth tracking next to the numbers the board already watches.

What separates the leaders

In our GEO work, we look at every AI answer as five stages, and a brand can drop out at any of them. Each stage has a pillar we work on:

Reach (Technical Foundations): can AI reach and read your website?
Select (Content Alignment): are your pages useful as an input to an answer?
Corroborate (External Mentions): do other sources back you up?
Describe (Brand Perception): how does AI describe you to the buyer?
Hand off (User Experience): can buyers who arrive from AI go on to enquire?

The first four explain most of what we see in this index. We cover them starting with the content on the page and finishing with what sits behind it. User Experience depends on a builder's own analytics, so it's outside a study like this one.

They answer the buyer's question (Content Alignment)

AI cites the pages that answer the question it's been asked. The most-cited scheme page in the study is Bellway's First Homes page, cited 44 times. Next come Barratt Homes' pages on offers, part exchange and low deposits.

Bellway's First Homes page, the most-cited scheme page in our study. The first two paragraphs answer the buyer's question: what the scheme is, the discount, and who qualifies.

The page doesn't open with a brand message. It opens with the answer, in the buyer's words, with the numbers AI needs to quote. That's why Bellway, which builds 15% of the top five's homes, earns 22% of the citations to their websites.

2. Others vouch for them (External Mentions)

AI checks what other sources say before it recommends anyone. In this sector, the Home Builders Federation's website is cited more often than any other single website in the study, including every builder's own. Its customer satisfaction star ratings are how AI justifies a shortlist.

Asked where to look for a new build in the Midlands, ChatGPT put Spitfire Homes, a regional builder outside the Top 50, at the top of its list. Its reason was a five-star HBF rating, reported in the trade press. Persimmon, Taylor Wimpey and Vistry didn't appear in any of its ten answers to that question.

One of ten ChatGPT answers to "Which housebuilder should I look at for a new build home in the Midlands?". Its top pick is justified by a Home Builders Federation five-star rating, with a trade press article as the source. Captured in Peec, 20 October 2026.

You can see why in the searches AI runs behind the scenes. Given the same question, ChatGPT didn't search for housebuilders. It searched for reviews, quality and customer satisfaction.

The searches ChatGPT ran to answer the Midlands question. Every one is about reviews, quality or customer satisfaction. Captured in Peec, 20 October 2026.

For quality questions, buyer forums carry real weight too. Reddit is cited in one answer in five and Facebook groups in one in seven. This is where mid-sized builders catch up with the giants. Asked which builders are most reputable, AI named Bloor in 96% of answers and Cala in 74%, alongside the big four. Asked which hold a 5-star HBF rating, it named Miller in 96%. These are single questions, around 25 answers each, so treat them as examples rather than scores. When the question is about reputation rather than range, AI leans on ratings and reviews, not completions.

3. AI knows what they're for (Brand Perception)

AI matches builders to the buyer's need, so a clear reputation for something specific is worth more than general size. Cala is named in every answer about buying in Scotland, and Miller in nine out of ten. Bloor is named in every ChatGPT answer about the Midlands. Berkeley is named first for build quality and features strongly in the South East and London commuter towns.

The flip side is being known for the wrong thing. Even when a homebuyer asks about Vistry by name, partnerships come up.

Gemini's answer when asked whether Vistry is a good company to buy a new home from. Transcript excerpt, 20 October 2026, captured in Peec.

Tone matters too. In answers about build quality and complaints, AI's sentiment towards Vistry scores 46 out of 100, the lowest of the top five, against 60 for Barratt Redrow.

4. They're a clear business to a machine (Technical Foundations)

First, what isn't the problem: all the top five's websites let AI crawlers in and work without JavaScript.

The difference is in how clearly each business defines itself. Bellway's website tells machines its founding date, awards, regions and social profiles. Barratt Redrow runs three consumer brands, and each site is at least marked up as an organisation. Vistry's three consumer brands, Bovis Homes, Linden Homes and Countryside, describe themselves only as websites, with nothing in their structured data linking them to Vistry. Semrush scores them 47, 42 and 34 out of 100 for AI visibility, against 61 to 76 for every other main consumer brand in the top five.

What each homepage tells a machine about the business. Bellway describes an organisation with a history, awards and regions. Bovis describes a website. Checked 20 October 2026.

Why the second-biggest builder comes fifth

Vistry shows all four factors at once. It completed 15,658 homes in 2025, more than anyone except Barratt Redrow. But around three quarters were built for housing associations, councils and rental investors, and only 4,065 were sold on the open market7. Count only those, and Vistry's share of the top five's homes falls to about 7.5%. Its share of AI mentions in homebuyer answers is 7%. Measured by the homes a buyer can actually buy, Vistry sits almost exactly where you'd expect.

In other words, AI isn't getting Vistry wrong. It's describing the business Vistry presents most clearly. Its corporate homepage leads with "mixed-tenure communities". Its most-cited page explains its partnerships model. Ask AI about partnerships and affordable housing, and Vistry is named in 92% of answers and named first in 82%.

The homebuyer side is harder for AI to piece together. Bovis Homes, Linden Homes and Countryside each have their own website, none of them tells a machine it belongs to Vistry, and Linden and Countryside share the same homepage description. When a buyer asks about Vistry by name, ChatGPT goes looking for reviews and complaints, and finds a review profile for the group rather than for the brands buyers actually deal with.

ChatGPT's answer to "Is Vistry a good company to buy a new home from?". Its searches (right) look for reviews, complaints and build quality, and its sources include a group-level Trustpilot profile. Captured in Peec, 20 October 2026.

For a business that wants to sell more homes on the open market, that's the gap to close. The partnerships story is working. The homebuyer story needs the same clarity.

What this means for housebuilders

A builder's position in AI answers isn't fixed by its size. It's shaped by what AI can find, what others say, and how clearly the business explains itself. That makes it something builders can manage.

Answer every buyer question. First Homes, part exchange, Deposit Unlock, shared ownership: each scheme needs its own page that answers the question directly, with the numbers AI needs to quote.
Earn the vouch. The HBF star rating, review sites, new-build listings and buyer forums are where AI checks a builder's reputation. Make sure your rating and reviews are visible and current, and that third-party listings describe you accurately.
Know your sentiment. AI's tone on build quality and complaints varies a lot between builders. Find out what AI says about your snagging, aftercare and customer service, trace the objections back to the forums and reviews they come from, and address them at the source.
Be known for something. Check what AI thinks you're known for, by region and by buyer type, and whether that matches where you want to sell.
Connect your brands. If you sell through more than one brand, make the link to the group clear, on the page and in your structured data, so AI can piece the business together.

How we approach GEO

This piece is a small example of how our GEO team works. For a client, we track the questions their buyers actually ask across the AI engines that matter, trace the sources behind every answer, and check what AI can read on their website. Then we work through the five pillars, from Technical Foundations to User Experience, to find where the brand drops out of the answer and what to fix first.

See where you stand. Our free AI Visibility Snapshot shows how often you appear in AI answers compared with your competitors, how AI describes you, and where to start. Request your Snapshot, or read more about our GEO service.

How this report was compiled

Questions: 50 in total. The index uses 36 homebuyer questions across four themes: choosing a builder by region, first-time buying, build quality and complaints, and part exchange and incentives. Another 14, comparing named brands and asking about partnerships and affordable housing, feed the Vistry analysis and the source counts.
AI engines: ChatGPT, Gemini and Google AI Overviews, UK, each question asked about 10 times on 20 October 2026, tracked in Peec. Google doesn't show an AI Overview every time, so the 36 homebuyer questions gave 981 answers rather than 1,080.
Builders: the Building Top 50 Housebuilders 20251, each measured across its brands.
Website checks: robots.txt, responses to AI crawlers, content without JavaScript and structured data on each top five consumer website, 20 October 2026. Brand scores from Semrush AI Visibility.
Search data: estimated UK organic traffic to each group's consumer websites from Semrush, September 2025 and 2026, checked against 24 months of Ahrefs data. The tools disagree on whether total traffic rose or fell, but agree on who gained share, so we report share.
Business data: completions and results from each company's latest published results, including Bellway's full-year results (13 October 2026) and Vistry's half-year results (24 September 2026).2–8
Limits: a snapshot from one day, so small differences between builders shouldn't be over-read.