A land registry extract uploaded to a new purchase file is read by AI: owners, parcel details, mortgages, liens and warning notes. Each encumbrance becomes a dated closing task, and the lawyer reviews a one-page title summary instead of a retyped file.
Typical results
≈20 min
Paralegal check and sign-off per extract
≈5.5 hrs/wk
Saved at a pace of six purchase files a week
100%
Encumbrances turned into dated tasks
Figures reflect typical results for this kind of automation.
The challenge
Every property purchase starts with a land registry extract that a paralegal retypes into the transaction file: owners and their shares, parcel numbers, registered mortgages, liens and warning notes. A busy extract takes 60 to 90 minutes, a missed lien can surface days before closing, and open items are tracked in scattered notes.
Before and after
Minutes of prep per land registry extract
−73%Retyping errors per 20 extracts
−75%The automation
Uploading the extract PDF to the purchase file triggers AI extraction that reads owners, ownership shares, parcel details, mortgages, liens and warning notes into structured fields in the transaction record. Each encumbrance becomes a task with an owner and a due date counted back from the closing date, such as a mortgage release letter from the seller's bank or a lien removal. Fields with low extraction confidence are highlighted for the paralegal to confirm against the PDF.
How it works
- 01
Extract uploaded
The paralegal uploads the land registry extract PDF to a new purchase file, which starts the extraction.
- 02
AI reads the registry entries
Owners, shares, parcel details, mortgages, liens and warning notes are read into structured fields in the transaction record.
- 03
Low-confidence fields flagged
Any field the AI is unsure about is highlighted so the paralegal confirms it against the PDF.
- 04
Encumbrances become tasks
Each mortgage, lien or warning note becomes a task with an owner and a due date counted back from closing.
- 05
Title summary for the lawyer
The lawyer reviews a one-page title summary and an open-items checklist before advising the buyer.
Starts from
- Land registry extract PDF
- Closing date
Delivers
- Transaction record fields
- Dated closing checklist
- One-page title summary
The impact
Title review preparation drops by about 70%, from roughly 75 minutes per extract to around 20 for checking and sign-off. The lawyer reads a one-page title summary with an open-items checklist, and every mortgage release or lien removal has an owner and a deadline well before closing.
Questions about this automation
What does the AI pull from the extract?
It reads the registered owners and their shares, the parcel and sub-parcel numbers, the area, registered mortgages with the lender and amount, liens, warning notes and any other encumbrances. Each item lands in its own field in the transaction record, so it can feed tasks, documents and the title summary.
How are closing tasks and due dates set?
Each encumbrance type maps to a standard task, such as requesting a mortgage release letter from the seller's bank or confirming a lien removal. Due dates are counted back from the closing date in the file, and the task owner is the paralegal or lawyer assigned to the matter.
What if the extract is a poor scan?
Scanned extracts go through OCR before extraction. Any field the AI reads with low confidence is highlighted for the paralegal, who confirms it against the original PDF before the tasks are created. Nothing is treated as final until a person has signed off on the title summary.