Events and roundtables
How can banks, building societies and specialist lenders deliver faster, better-evidenced mortgage decisions without weakening human accountability or customer outcomes? Five invitation-led discussions will explore a shared evidence foundation across residential purchases and remortgages and buy-to-let purchases and remortgages. Bring a decision bottleneck, a control question and a measure worth testing. The aim is to define practical lender-specific tests, not promise the same route or result for every mortgage. Dates, venues and participation are not confirmed.
Programme
Series white paper
A lender-led discussion on earlier evidence, less rework and accountable residential and buy-to-let assessment.
A late property query can leave a buyer wondering whether a home, a school move or a chain will be lost. Brokers chase answers while lender teams reconcile records and repeat checks. The question for lenders is simple. Can usable evidence reach the right person earlier and improve offer time, net processing cost and customer confidence?
Trusted Offer Record (developed by Crosstide as Property Spine) is a proposed way to connect property evidence, its source and permission to use it with the work a lender needs to complete. It combines Crosstide engineering, property intelligence and Thistle risk and governance expertise, with roles agreed per engagement. It is not a new core lending system.
Why now. Home-buying reform is pushing earlier information and digital sharing, and professional AI standards now expect accountable review. The difference here is a focus on closing evidence gaps and handoffs, not generating another summary or speeding up an unchanged process.
The invitation. Book a 45-minute exploration with David Titterton of Crosstide. Bring one bottleneck and leave with a candidate cohort and the first baseline measures for a separately scoped discovery. No identifiable customer records are needed.
Three mortgage journeys
Two editorial discussion illustrations alongside the existing synthetic buy-to-let case. Unresolved evidence triggers review in every example.
Owner-occupied purchase
Discussion illustration
A fictional first-time buyer is purchasing a flat. The application and property record identify different units and lease information is incomplete. The reviewer resolves identity and evidence gaps before the assessment proceeds.
Residential remortgage
Discussion illustration
A fictional homeowner reports an extension. An older record does not establish the current layout or condition. The reviewer checks what has changed and what updated assessment the lender requires.
Synthetic buy-to-let remortgage
Synthetic case in the approved-user portal
DEMO-002 declares rent above an area estimate and the tenancy agreement is missing. The reviewer requests evidence rather than inferring the answer.
Open the case in the portal
Neither a source summary nor an available data record constitutes a valuation, an affordability assessment or a mortgage offer.
Interactive synthetic example
A mortgage application rarely waits for one decision alone. It can wait while people reconcile a property record, chase a missing document or work out whether an estimate is reliable enough to use. This example follows the evidence through those handoffs, showing where trusted data and bounded AI support could help and where a person must stop, question or escalate.
Existing interactive example | Synthetic buy-to-let remortgage
Imagine a broker preparing a landlord’s remortgage application. The declared rent is £1,650 a month, but the area estimate is £1,300 to £1,400 and the current tenancy agreement has not been supplied. These are synthetic figures from DEMO-002, not market statistics. Neither accepting the declaration nor substituting the estimate answers the real question of what evidence supports the tenancy income.
The lender’s specialist needs to understand the conflict before relying on the rent. The broker needs a clear request for the current agreement and supporting evidence, rather than several disconnected follow-ups. The customer needs an understandable explanation of what is outstanding and why the application cannot yet progress. AI can help organise a source-grounded draft, but it cannot verify missing evidence or make the lending decision.
The approved-user portal takes the discussion beyond a diagram. Inspect the case evidence with its source, dates, confidence and permitted use, compare what the application says with what the record supports and follow the lender, broker and customer handoffs. Separate exercises let you record a reviewer’s rationale, explore identity or stale-evidence exceptions and download a synthetic audit trail. These exercises do not remove the unresolved checks from the original case.
In the broker workspace, customer-update drafts require approval of the exact wording before a simulated send. The queue shows who owns an outstanding action and when a follow-up is due. Together, these views make the practical questions visible at each handoff. What is known? What is missing? Who acts next? What can responsibly be said to the customer? Any reduction in delay, cost or repeated requests remains a hypothesis to measure, not a demonstrated result.
This is one worked example of the broader proposition. The owner-occupied purchase and remortgage scenarios above are discussion illustrations, not additional implemented portal workflows. The public stages below are scripted. Portal exercises use synthetic evidence and simulated actions, send no real customer messages and issue no mortgage offer.
Synthetic DEMO-002 is a buy-to-let remortgage. A validated property identity connects the evidence, using a UPRN where appropriate. Identity alone does not prove lending suitability.
Source, timestamp and permitted use must be checked.
Aligned roles and handoffs at each stage. Synthetic DEMO-002 illustrates a conflict that needs human review, not a completed offer.
Stage 1 · Property identity
Better-evidenced reviews and fewer avoidable checks, without weakening policy.
Specify property identity, evidence rights and acceptable sources.
Handoff · Lender → Broker
Evidence checklist and permitted-use requirements.
Stronger submissions, less repeated chasing and clearer next actions.
Match the property and assemble source-backed evidence within permitted use.
Handoff · Broker → Lender
Property identity, available evidence and declared gaps.
Less avoidable waiting, clearer explanations and appropriate support.
Confirm the property and share relevant information through the broker.
Handoff · Customer → Broker
Case details and any communication or support needs.
DEMO-002 · Fictional buy-to-let borrower. Days illustrate a possible sequence, not actual waiting, a service promise or measured savings. Later stages are conditional. Evidence remains incomplete.
Original scenario. Rental, tenancy and licensing gaps remain unresolved.
Illustrative day 1
1 · Open and match · Current
Original conditional sequence. Existing evidence gaps remain.
Illustrative days 2 to 3
2 · Assemble evidence
Original conditional sequence. Existing evidence gaps remain.
Illustrative day 4, if evidence arrives
3 · Ground the summary
Original conditional sequence. Existing evidence gaps remain.
Illustrative day 5, subject to review
4 · Review and refer
Original conditional sequence. Existing evidence gaps remain.
Later, only if all requirements are met
5 · Communicate outcome
Original conditional sequence. Existing evidence gaps remain.
Original conditional sequence. Existing evidence gaps remain.
No additional disruption active. The original rental-evidence checkpoint remains below.
Scenario overlays only. Restoring the original scenario does not resolve an exception or erase recorded follow-ups. Reload clears this role-play. Case evidence, other exercises and actual borrower updates are unchanged.
Every value is invented. Missing items have no source document or assessed use rights. Fixture dates do not establish currency today.
Flagged exception · £1,650 declared rent differs from the £1,300 to £1,400 area estimate and no tenancy agreement is supplied. Alex Morgan, a fictional rental specialist, must review it. The estimate cannot verify tenancy income.
At least 10 characters. Explain the evidence gap and reason for the next action.
Separate page-local role-play. Clears on reload. No real documents, messages, AI calls or lending decisions. Case records and other exercises are unchanged.
Broker → Lender. Property identity and evidence rights
DEMO-002 · All sources and dates are fictional. Freshness is measured at the fixture assessment date, not today. Changing a handoff never refreshes or verifies evidence.
Assigned demonstration owners, not notifications or real staff assignments. Completion supplies evidence for human review, never an automatic offer.
DEMO-002 rental-evidence pathway. Invented assumptions compare repeated late requests with one complete, permitted request before submission. Neither pathway resolves the fixture or produces an offer.
Rental specialist review and licensing assessment still required. No offer issued.
Rental specialist review and licensing assessment still required. No offer issued.
Assumed difference · 2 repeat requests, 3 working days of repeat-request waiting and 30 minutes of handling.
Late requests equal one initial request plus the chosen repeats. Early requests equal one consolidated request. Waiting equals repeats multiplied by working days per repeat. Handling equals requests multiplied by minutes per request.
The early pathway assumes complete, correct and reusable evidence arrives before submission. This shifts preparation upstream and does not remove customer preparation, necessary review, initial-response waiting, data charges or integration costs. Waits are assumed sequential, not overlapping. Zero repeats means no assumed difference.
These figures are not total application-to-offer time, net savings, forecasts or measured outcomes. Test the assumptions against a real baseline and include decision quality, support needs and complaints before claiming value.
Separate scripted DEMO-002 exercise. A previously reviewed fictional rent document is superseded by an amendment. This is not a live data feed and does not suggest the missing workspace tenancy agreement has been supplied.
Prior exercise document. Monthly rent £1,650. Exercise-only prior review. Source case remains unresolved.
Replacement exercise document. Monthly rent £1,450. Synthetic tenancy amendment supplied by the broker. Observed 2026-10-08, retrieved 2026-10-09.
Affected checks · Declared monthly rent and Tenancy agreement. The next amendment would reopen these two checks.
Unaffected evidence · Property identity and capital-value sources retained. Licensing remains unassessed. No automatic recalculation of rental cover, eligibility or offer.
Owner · Alex Morgan (fictional rental reviewer). Broker coordinates clarification and the customer update.
Recorded actions persist only in this browser tab. Original evidence and earlier rationale are retained. Future live monitoring depends on contracted access, coverage, dates and permitted use.
Illustrative permissions, not source licences. Fixed exercise date 2026-10-09. Sharing a report does not grant another party permission to rely on it. No evidence is transferred here.
Within the illustrative permission window. Human source and property checks still required.
Fresh check required. Exercise validity has expired.
Reuse blocked. Confirm source rights.
These invented review windows are not a lender policy. An in-date record can still be wrong, superseded or unsuitable. Check the current source, property match and actual contract before reuse.
Invented terraced-property example with an assumed 80 m² floor area. All prices, dates and attributes are synthetic. No valuation, physical inspection or lender routing decision is produced.
A valuation professional must assess comparability, source reliability and limitations. No automatic adjustments or overall value score hide the unknowns.
Separate scripted DEMO-002 communication exercise. The underlying rental checks remain unresolved. This local template is not AI-generated and does not send messages.
Proposed customer-facing action
Ask for monthly rent confirmation and a current tenancy agreement. The broker coordinates the request and the lender rental reviewer assesses the evidence.
Approval and confirmations clear on refresh. The next update date is an exercise assumption, not a scheduled reminder. No borrower data, internal risk assessment or automatic offer promise is included.
Generic provider acceptance exercise, applicable when exploring a property intelligence source or another source. No actual contract, supplier assessment, integration or endorsement is recorded. Completing this checklist cannot authorise live use.
0 of 6 evidence reviews recorded in this exercise
Agree a field dictionary, geography, property types and missing-data treatment.
Confirm observation and retrieval dates, refresh frequency and change-notification limits.
Retain original source references, match confidence and correction procedures.
Agree permitted recipients, purposes, retention and any limits on lender reliance.
Confirm permitted AI use, personal-data handling, model-provider terms and retention.
Test sample records, agree accountable owners and escalation routes and document lender approval.
Notes remain only until refresh. Use fictional information. Before a live pilot, obtain actual source agreements and lender, legal, security and governance approvals.
Property intelligence describes property data, sourced AI answers and pack preparation. Property Spine complements evidence providers by connecting permitted evidence to lender requirements, accountable human reviews and controlled stakeholder handoffs. Supplier speed, coverage and satisfaction claims are not Property Spine results.
Governance applies to every handoff. Check access, source currency, permitted use and review rationale. Intended benefits are unmeasured. This scripted example runs no AI, changes no case records and issues no offer.
Scripted synthetic demonstration. No customer records, live data feeds, model call, lending approval or achieved offer-speed result.
Explore the approved-user PoC workspaceHow the room works
Output
Guardrails
Propose
Tell us the sector interest and the organisation. We convene the room, chair it neutrally and publish only what the room agrees to.