evidoria

← Back to browse

Good practice Imported

Magic Notes — Swindon's AI Meeting-Notes Tool Cuts Social Work Admin Time by More Than Half

United Kingdom · Swindon · See the United Kingdom profile

Evidence: Observational / pre–post Top 14% 80/100 · Ask Evidence Copilot about this practice

Swindon Borough Council piloted Beam's Magic Notes AI transcription tool with 19 social workers across 184 meetings in 2024, cutting Care Act assessment time from 90 to 35 minutes and report-writing from four hours to 90 minutes; the tool is now used by 85 UK councils.

19 social workers
Social workers in Swindon trial (Apr–Jul 2024)
184 meetings
Meetings covered in trial (Apr–Jul 2024)
90 → 35 minutes
Average assessment time before/after (2024)
4 hours → ~90 minutes
Average report-writing time before/after
44–50%
Admin-time reduction in Ealing rollout (2024)
100+ social workers
Social workers covered in Ealing rollout
4.26 out of 5
Average user satisfaction score
85 councils
UK councils using Magic Notes (late 2024)
Magic Notes — Swindon's AI Meeting-Notes Tool Cuts Social Work Admin Time by More Than Half

Details

Maturity
Scaling
Promoter
Swindon Borough Council
Period
2024–present
Keywords
social care, public administration, case management

Context

Swindon Borough Council trialled Magic Notes, an AI transcription tool built by UK welfare-tech company Beam, between April and July 2024, to help social workers with the administrative burden of writing up assessments and reports.

Objectives

The trial aimed to test whether AI-generated transcription and draft case notes could reduce the time social workers spend on Care Act assessments, Mental Capacity Assessments and related report-writing while keeping a human reviewer in the loop.

Activities

Nineteen social workers used Magic Notes across 184 meetings to automatically transcribe conversations and draft case notes and reports, with all outputs reviewed by a human before entering the case record.

Results

The trial cut average assessment time from 90 minutes to 35 minutes and report-writing time from four hours to roughly 90 minutes; a related rollout in Ealing recorded 44-50% admin-time reductions among more than 100 social workers, user satisfaction averaged 4.26 out of 5 across trial councils, and Swindon secured a further six-month contract with Beam after the pilot.

Conclusions

Practitioners also reported accuracy issues — unwarranted assumptions about service users' needs, misspelled or repeated names, and difficulty in multi-person case meetings — underscoring the need for human review, though the tool has since spread to 85 UK councils including Barnet, Camden, Kingston, Peterborough and Oxfordshire.

Implementation

Indicative cost
Low (< €50k)
Time to results
Medium (1–3 years)
Staffing & skills
19 Swindon Borough Council social workers (trial participants), Beam (UK welfare-tech vendor), human reviewers signing off AI-drafted case notes before they enter the case record

Conditions for success

  • Mandatory human-in-the-loop review before any AI output enters a case record
  • A follow-on contract with the vendor after the pilot
  • Use across genuine Care Act and Mental Capacity Assessment meetings to test real workloads

Common failure modes

  • Unwarranted assumptions about service users' needs
  • Misspelled or repeated names
  • Difficulty handling multi-person case meetings

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful