United KingdomPublic sector2021 – ongoingDepartment for Work and Pensions (DWP)
Child Maintenance Group returned-letter address extraction
DWP automated address extraction for a returned-letter workload of around 5,000 letters monthly
Problem
The Child Maintenance Group received around 5,000 undeliverable letters each month. Staff previously opened letters and investigated addresses manually.
Approach
DWP and Accenture developed the extraction system in 2021. Computer vision located address and reference-number regions in scanned letters, and optical character recognition converted them into text. The service ran Monday to Friday, with failed extractions referred to staff.
Results
Takeaway
The system supplied extracted information to the Child Maintenance Group. DWP did not publish an extraction accuracy or measured time saving.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
DWP documented production use of computer vision and OCR for returned letters. It reported around 5,000 incoming returned letters monthly and described human review when extraction failed.
Full details are in the original sources.