Deemerwha studio/Shutterstock.com
The development of an AI (artificial intelligence) tool by HM Land Registry's Local Land Charges Programme is dramatically accelerating and enhancing the migration of local authority local land charges records.
The challenge
Local authorities across England and Wales hold millions of local land charges records in electronic, paper and other formats. Before this information can be migrated to the central Local Land Charges (LLC) Register, every line of data must be reviewed, checked for accuracy and digitised. It’s time-consuming work that requires large teams and months of effort. But the development of an innovative AI tool promises to change that.
How AI is helping
Developed by HM Land Registry’s Data Science team, our cutting-edge AI tool is transforming how the LLC Programme processes millions of pages of historical local authority data. The AI tool can identify, extract, and structure LLC information faster and more accurately than manual efforts. Using this tool will significantly speed up future LA migrations to our national LLC Register.
The London Borough of Newham was the first local authority to have its textual LLC information processed by AI and pass all data quality checks.
Initially, the LLC Programme estimated it would take 20 Data Transformation Team colleagues three months to capture Newham’s data. AI completed the task in four weeks with just four people.
Proven results
Quality fails requiring data re-capture are the most expensive and time-consuming tasks within data transformation, adding weeks to a schedule. Our new AI tool has proven it can eliminate these errors. Every batch of Newham’s data successfully passed the data quality assessments first time. Maintaining these rigorous quality standards is critical to ensure customers get the accurate information they expect.
The London Borough of Newham may have been the first local authority to have its LLC data processed using large language model AI, but it won’t be the last. We’re committed to harnessing technology wherever practical to enhance both our efficiency and effectiveness. The AI tool we have developed is enabling our data transformation team members to move from time-consuming tasks to more complex, high-value work that benefits from human expertise.
Mark Kelso, Programme Director for Local Land Charges, HM Land Registry
Our commitment to build strong data governance, improve data quality and modernise systems set out in the Newham Data Strategy is represented in this project. The input from HMLR’s LLC Data Quality and Assurance team has provided us with assurance that our data and the outputs meet our shared high standards.
Jennifer Bishop, Assistant Director Development Delivery, London Borough of Newham
The journey
The development of the new AI tool has taken more than five years of agile working to reach this point. Early AI iterations required a bespoke code for each local authority, running on physical workstations, built to suit each local authority’s requirements. Now the Data Science team has created a more robust, reusable system that operates within HM Land Registry’s secure cloud infrastructure.
The AI tool’s reliable extraction of data means typographical errors such as the letter O becoming zero, or the letter I becoming one are a thing of the past. The sophisticated system combines an enhanced Optical Character Recognition engine, a large language model and several bespoke models to enhance accuracy. Strict controls prevent the generation of false information or rogue data like menus becoming part of the dataset, helping to safeguard the integrity of the LLC Register.
Only one team member is needed to process the data, before it’s cross-checked by the LLC Data Quality and Assurance team. To date there have been no quality or data completeness failures.
What’s next?
Following Newham’s success, we have also used our new AI tool with Dacorum Borough Council and Cotswold District Council. The Data Science team has identified other local authorities with similar datasets so they can rerun the tool and continue to enhance its capability such as looking at solutions for handwritten documents and poorly scanned records which can be challenging to digitise.
The future of AI within HM Land Registry
AI will play an important part in delivering HM Land Registry’s Strategy 2025+. While we recognise its incredible opportunities to shape the way we work and serve our customers, we acknowledge the need to integrate it responsibly. Alongside the great work happening with the LLC data migration, we have been testing how different types of AI tools could work for us in other areas, particularly to unlock the value of our data for wider usage.
Lynne Nicholson, Deputy Director for Data, HM Land Registry
We welcome your comments about this blog in the comments below. Please note that we are unable to discuss individual cases through the comments section and would request that all such queries be directed to our Contact Us web form where you will receive a response as soon as possible.