Cotality Joins OPDA as UK Property Data Race Exposes AI Valuation Blind Spots
The firm whose software touches nearly every UK valuation instruction just signed up to the open data standards body. The timing matters: agents are simultaneously discovering they have no reliable way to measure how AI search is redirecting buyers away from their listings.

A 150-Million-Record Bet on Open Standards
Cotality, the property data and analytics provider formerly known as CoreLogic, has become the latest member of the Open Property Data Association (OPDA), joining a network that already spans the major High Street banks, building societies, conveyancers and technology providers pushing to standardise how property data moves across the UK transaction chain.
The numbers Cotality brings are not small. The company holds a database of over 150 million property records covering more than 25 years of historical transactional data, and monitors around 30.9 million UK residential properties each day for events that affect value and risk, drawing on more than 100 data sources. Its analytics, AI and workflow tools support estate agents, lenders, surveyors and conveyancers, including valuation and surveying software used in nearly all UK property valuation instructions.
That last detail deserves a pause. "Nearly all" UK valuation instructions. If a firm with that footprint decides open data interoperability is the direction, the rest of the industry's argument for staying siloed gets harder to make.
What Cotality Brings, and Why OPDA Wanted It
Cotality COO Mark Blackwell said the company's existing Integrated Lender Hub, which connects lenders and surveyors and digitises parts of the valuation workflow, already demonstrates the kind of digital connectivity OPDA is trying to standardise across the whole sector. Joining the association, he said, was aligned with plans to develop interoperable data sharing and increase digital connectivity across the property transaction process.
OPDA chair Maria Harris was direct about the strategic value of the addition, noting that Cotality's "deep technical expertise and proven integration across lenders and valuation networks" would be invaluable as OPDA continues building open, trusted standards for the home buying and selling process.
The move sits inside a broader regulatory current. The Department for Business and Trade is running a Smart Data Multi-Sector Call for Evidence, open until 1 October 2026, that is explicitly scoping property data sharing as a candidate for future regulation. OPDA has already urged the property industry to engage with that consultation. Firms that have shaped the data standards before the consultation closes will be better placed when mandatory frameworks arrive.
The Valuation Stack, and Its Gaps
Cotality's AI and workflow tools support "nearly all" UK valuation instructions. Simultaneously, the RICS is developing global practice guidance on AI use in real estate valuation, issued for public consultation in Q2 2026 and expected to be published later this year. The guidance aims to support registered valuers in the responsible use of AI within valuation practice, providing a framework around professional judgement, transparency and accountability.
Those two developments, a dominant data player joining an open-standards body and the professional regulator writing the rules for AI in valuation, are converging on the same pressure point. Automated Valuation Models process around 50 million valuations a year in the UK, covering close to 75% of the mortgage market, according to industry data. The accuracy headline: broadly 80% of estimates land within 10% of a chartered surveyor's figure for standard properties. The ceiling matters. AI cannot inspect a property, see a damp problem or price a Grade II listed conversion in a low-transaction postcode. Surveyors describe disputes between buyers and sellers over AI-generated figures as one of the most significant sources of friction in the 2026 transaction market.
Open data standards, if Cotality's membership of OPDA helps accelerate them, could raise the floor on AVM accuracy by giving models access to richer, more consistent inputs. That is the structural pitch.
The Visibility Problem Nobody Is Measuring
Separate research published this week adds a layer of discomfort. A survey of 622 UK estate agents carried out by GetAgent in February found that only 7% believed AI search was improving their listing visibility, with 93% reporting no impact. Yet 84% of those agents expected AI search to become a primary way buyers look for property. Only 22% had seen a buyer reference AI when explaining how they found a property.
The problem embedded in those numbers: agents are reporting what buyers tell them, which may bear little relation to what actually drove that buyer's discovery. Rightmove became the first UK property portal to launch an app in ChatGPT in March 2026; Zoopla runs one too. When a national portal sits inside an AI assistant and resolves as an entity across the whole web, an individual branch has almost no chance of getting cited at the same scale. The agents who say AI search is not affecting their listings may simply have no mechanism to find out that it is.
The data infrastructure problem and the AI visibility problem are, at root, the same problem: fragmented, non-standardised property data that leaves agents unable to see what the machines are doing with their stock.
The Trademark Signal
One way to track where the sector thinks AI will actually matter is through trademark filings. AI Business Dispatch analysis of IPO (TMD) data shows UK Class 42 filings, the Nice class covering technology and software services, reached 6,136 in 2026-Q3, down 17.1% against the prior period. That decline may reflect a market that has moved past the land-grab registration phase and into consolidation: fewer speculative filings, more purposeful brands. Cotality's move into OPDA is consistent with that reading. This is infrastructure-building, not branding noise.
The UK property transaction chain is long, slow and still partly paper-based. The firms betting on open data standards are betting that AI tools are only as good as the data fed into them. On that point, the numbers are hard to argue with.
