SWEET: The Social Work Evidence Engagement Tool

NIHR-funded project developing an AI tool to help social workers find and apply relevant research evidence, legislation, and practice guidance in their day-to-day work.

Funder: National Institute for Health and Care Research (NIHR)
Period: 2026 – 2028
Role: Work Package Lead (system development, evaluation, and human-in-the-loop testing)
Partners: Cardiff University (Computer Science and Social Sciences), Neath Port Talbot Children’s Social Care
Research theme: NLP for Education, Healthcare & Social Impact

Social workers face significant practical barriers to using research evidence: time pressure, fragmented information sources, and the challenge of translating general findings into specific practice decisions. SWEET addresses this by developing a specialised AI tool for reliable retrieval and summarisation of relevant research, legislation, and practice guidance in response to social worker queries.

The project uses design ethnography to understand evidence needs in real practice settings, working closely with Neath Port Talbot Children’s Social Care. At Cardiff Computer Science, the focus is on NLP system development, evaluation, and the design of human-in-the-loop testing protocols to ensure the tool supports rather than replaces professional judgement.