The lifetime property record. From repeated checks to reusable evidence

Every mortgage application asks a version of the same question. What is this property, what condition and value evidence supports the loan and who can rely on it? Today the answer is usually rebuilt from scratch, by different parties and at different points in the chain. Evidence gathered for a buyer, a broker or a valuer rarely arrives in a form the next decision-maker can reuse with confidence.

The cost shows up in time and failure. The government home buying and selling reform roadmap says that once an offer is accepted a purchase takes around 120 days on average to complete, that the journey is around 60% longer than in 2007 and that around one in three transactions fall through. [13]

These figures describe the purchase market. They are not mortgage-offer processing times and they do not apply directly to remortgages or product transfers. The House Buyer Bureau figure is a commercial estimate rather than an official statistic and the government savings are forward-looking. Their value here is to show why late surprises matter, not to forecast a Property Spine saving.

The scale of the system raises the stakes. The FCA and Bank of England report £1,760.6 billion of outstanding residential mortgage loans and £77.4 billion of gross advances in Q2 2026. [1] Small amounts of repeated work or avoidable delay, multiplied across that volume, become a real cost to lenders, intermediaries and customers. Equally, small control failures at that scale become a real risk.

What the room should take away.

Repetition is also a quality problem, not only a cost. Each time a fact is re-keyed, re-ordered or re-interpreted there is a fresh chance of a mismatch between the address a broker entered, the title a conveyancer reviewed and the unit a valuer inspected. A record that carries identity and provenance forward could reduce the number of places where those errors enter. That is a design hypothesis, not a measured result.

The opportunity is therefore twofold. Fewer repeated requests could reduce cost and waiting, while consistent identity and provenance could reduce the late surprises that the government roadmap links to fall-throughs. [13] Neither benefit has been measured for Property Spine and both should be tested rather than assumed.

Several independent signals now point the same way. Policy, standards and lender practice are each moving towards earlier, structured and shareable property information.

Policy. The June 2026 roadmap sets out a direction towards upfront property information and digital property logbooks. [13] OPDA reports that the preceding consultation received 1,133 responses, with 79% agreeing that upfront information should be mandatory and 82% supporting digital property logbooks and property packs. [2] These are consultation responses reported by an industry body, not a representative survey or a measured saving.

Standards. OPDA released its v3.5 schema in September 2025, including standardised survey and valuation information. [18] A common data language could reduce translation effort between parties. It does not by itself give anyone permission to access or rely on that information.

Lender practice. In November 2025, Nationwide reported that more applicants were receiving mortgage offers within 20 minutes of starting their application, supported by automated income verification and automated valuations. Nationwide did not disclose the number or percentage of applicants achieving that turnaround. [16] The same lender statement identifies income verification as one contributor, so this is not evidence of a property-data-only saving. [16] Nationwide also joined OPDA in 2024, showing lender interest in shared property-data standards. [19]

Registration. HM Land Registry reports more than 100,000 fewer requisition letters during 2025 to 2026, alongside customer training and sharing avoidable-requisition data. [11] It is a registration-system result, but it illustrates that structured feedback on avoidable errors can reduce repeat work.

Cross-sector data policy. The July 2026 Smart Data call for evidence cites a projection of £71.2 billion in net social value from four schemes over 2028 to 2043. [3] That is a conditional cross-sector estimate rather than a property benefit, but it shows government intent to build consent-based sharing with governance at its core.

Signals worth separating.

Legislation is now part of the same picture. The Data (Use and Access) Act 2025 received Royal Assent on 19 June 2025 and creates powers for new Smart Data schemes like Open Banking, alongside digital verification services and changes to data protection law. [22] Commencement is staged by regulation, so the Act enables future schemes rather than creating a property scheme today.

The July 2026 Smart Data call for evidence cites £71.2 billion of projected net social value over 2028 to 2043 and a £9.6 billion annual GDP contribution by 2043 across four candidate schemes. [3] These are conditional cross-sector estimates. They are useful as a signal of policy ambition, not as a property-market forecast.

Read together, the signals suggest a practical sequence.

The registration system is moving in step. HM Land Registry holds more than 27 million titles covering more than 90% of the land area of England and Wales. [30] Its 2026 business plan says it will automate some of the highest-volume applications, starting with residential mortgages. [31] That is a plan rather than a delivered result, but it signals an ambition for a more machine-ready public record.

A lifetime property record is not one giant database. It is a governed way of linking evidence to a property and carrying the conditions for its use. The Unique Property Reference Number, or UPRN, is a strong candidate join key because it identifies an addressable location consistently across datasets. It supports matching evidence to a location, subject to checking that the right property or unit has been linked. [25] It says nothing about value, condition or suitability.

Each fact in the record needs six attributes to be reusable.

Candidate content includes title and tenure, lease terms where relevant, energy performance, planning and building-control history, environmental signals, survey and valuation information, and, for rented property, licensing and tenancy evidence. Borrower, affordability and legal checks remain outside the property record and stay with their accountable owners.

Market context can enrich the record without becoming case evidence. A property intelligence provider reports 185,595 new listings and 106,296 sales agreed across Great Britain in August 2026, a 57.3% period conversion ratio. [4] Data of this kind can frame local market questions. It does not verify a particular property, tenancy or valuation.

The UPRN is already well placed for this role. Ordnance Survey describes it as a unique identifier for approximately 40 million addressable locations in Great Britain, primarily allocated by local authorities and published as open data updated every six weeks. [25] Government open standards require public-sector systems that store or publish property and street information to use UPRN and USRN identifiers. [26]

An identifier is a key, not a conclusion. A UPRN can link title, energy, planning and valuation evidence to a location, yet it does not confirm the tenure, the condition or the value. For flats, conversions and properties that have been split or merged, the hardest question is often whether the evidence describes the right unit. A credible record therefore shows how each match was made and how confident it is.

The value of a reusable record grows each time the property meets a new decision. The question at each point is not whether evidence exists, but whether it is current, relevant and permitted for this decision.

Valuation remains a lender decision. LSL reports that automated valuation models account for about 25% of UK residential valuations, typically for lower loan-to-value lending. [12] A richer record may help a lender decide which cases suit an automated or desktop route and which need inspection. It does not remove valuation requirements, and a lender’s mortgage valuation is different from a buyer’s condition survey.

LSL also reports £27.3 billion of mortgage lending in H1 2026, with £16.1 billion of purchase and remortgage lending and £11.2 billion of product transfers. [7] The split is a reminder that existing-customer journeys are a large share of activity and deserve their own evidence design.

A synthetic illustration shows the difference. A homeowner who bought a flat in 2023 remortgages in 2026. Under a one-off model the lender orders fresh searches and a valuation and the broker re-asks for lease details. Under a governed record the lender sees which facts were checked in 2023, which have expired under its own policy and which have changed since, such as a new energy certificate or a lease extension. It then requests only what its policy says is stale or missing.

The scale of existing-customer activity makes this relevant. LSL reports £11.2 billion of product transfers within £27.3 billion of H1 2026 lending. [7] These journeys rarely need a full purchase evidence pack, yet they still depend on knowing that the property and its risks are as previously understood.

Market-wide figures point the same way. UK Finance reports around half a million refinancing transactions in Q1 2026, up by a third on a year earlier. [28] It forecasts £300 billion of gross lending in 2026, including £77 billion of external remortgaging, while house purchase stays broadly flat. A separate £261 billion product-transfer forecast is excluded from gross lending. [27] These trade-body figures and forecasts indicate substantial refinancing activity. They do not establish the proportion of all mortgage activity eligible for evidence reuse.

Reuse must also respect the purpose for which evidence was gathered. A valuation commissioned by one lender is not automatically available to another, and a condition survey for a buyer is not a mortgage valuation. The record should carry those limits with the fact so that reuse is a decision, not a default.

A shared record only works if each participant sees a reason to contribute. The incentives differ by role and a design that ignores them is unlikely to last.

Accountability should follow the decision, not the data. The supplier remains responsible for what it supplied and when, while the lender remains responsible for deciding that the evidence was sufficient for this loan. Recording both separately could help prevent two failure modes, which are suppliers being blamed for unforeseeable uses and lenders assuming that someone else checked.

Competition needs care as well. Shared identifiers and schemas are a common foundation, but pricing, risk appetite and policy rules should remain each firm’s own. A record that standardises how evidence is described without standardising how it is judged keeps the market competitive.

Regulators are encouraging this kind of controlled testing. The FCA reports a second cohort of eight firms in AI Live Testing in April 2026 and plans a good and poor practice report on AI later in the year. [24] Supervised testing of data and AI foundations is becoming a normal part of responsible adoption.

Reuse creates new risks alongside its benefits. A credible design answers these questions before it promises speed.

Absence of a record must never be treated as absence of a risk. A record that looks complete can still be wrong about the unit, the lease or the current condition. The design should make gaps visible, prompt review where evidence conflicts and preserve earlier decisions when later evidence changes.

Consumer Duty provides a useful lens where it applies, covering products and services, price and value, consumer understanding and consumer support. [8] Its applicability depends on the product, activity and customer, so it should be assessed rather than assumed for every journey.

Automated decisions sharpen the rights question. The government summary of the Data (Use and Access) Act describes a more permissive framework for solely automated significant decisions, with safeguards including information about the decision, the ability to challenge it and the ability to obtain human intervention. [23] A reusable record that feeds automated routes would need to make those safeguards practical, by showing customers which facts were used and how to correct them.

Correction is the overlooked design problem. If a reused fact is wrong, the fix must reach every later decision that relied on it, together with a note of which decisions may need review. An append-only history that keeps the original, the correction and the reason is more defensible than silently overwriting a value.

The lifetime record should be earned in steps rather than declared. Begin with one property type, one journey and synthetic cases, then expand only when evidence supports it.

Measures worth agreeing include repeat evidence requests per case, the share of reused facts that needed correction, time from submission to a clear property assessment and customer understanding of what is needed. Purchase, remortgage and product-transfer cohorts should be reported separately.

Accenture reports that only one in ten UK organisations had scaled AI or embedded it into core operations by early 2026. [10] Data foundations are part of that gap. A governed, reusable property record could be one of the foundations that allows later automation to be trusted.

The invitation is practical. Bring one evidence bottleneck, one permission question and one measure worth testing. The aim is a shared design that any participant can test, not a single platform or a promise of faster offers.

Governance should be proportionate to reliance. A fact used only to prompt a question needs lighter controls than a fact used to waive a valuation or confirm tenure. Classifying each use by consequence keeps controls focused where errors matter most.

Success should be judged on quality as much as speed. A reuse rate that rises while the correction rate also rises is a warning sign, not a success.