10 to the 23 AI logo
Work With 1023AI

Domain Atlas

Caseworker documentation & copilots

Generative assistants drafting the records institutions run on — where today's convenience becomes tomorrow's contaminated memory if review erodes.

What AI is doing here

Generative case documentation

Generative

Transcription and drafting of assessments and case notes, reviewed by workers before entering the record.

General staff copilots

Generative

Office assistants (drafting, summarizing, search) used informally across casework and administration.

What has gone wrong — and right

Documented deployments, presented as stylized model organizations with full citations.

Magic Notes (Beam)

United Kingdom (local authorities)

A generative documentation assistant spreading through UK social care — the live test of human-in-the-loop at scale.

Who is in the system

  • Frontline workers. Caseworkers, screeners, eligibility staff — the operator network whose judgment the system augments or erodes.
  • Supervisors & QA. The institutional correction layer: overrides, second reads, quality review.
  • Agency leadership. Owns procurement, policy, and the authority map; answers for the system publicly.
  • Served people & families. Those the decisions land on. Deliberately outside the PAN dynamics — their outcomes are measured, never simulated.
  • Vendors. Build and update the systems; hold the information asymmetry procurement must govern.

Dominant pressures

  • Caseload surge. Demand outruns staffing; per-case attention shrinks and review becomes triage.
  • Reviewer bottleneck. One fixed-capacity checking stage sits between AI output and consequence; everything queues behind it.
  • Deadline pressure. Statutory or managerial timeliness rules reward fast approval of machine output over slow disagreement.
  • Staff turnover. Experienced skepticism leaves; new staff calibrate their trust on the tool itself.

Questions leaders should be asking

  1. 1. What fraction of AI-drafted records are substantively edited before acceptance — and is that measured?
  2. 2. Are AI-drafted entries labeled in the record, so downstream readers and systems know their provenance?
  3. 3. Who audits accepted notes against their source (audio, interview) on a schedule?
  4. 4. What happens to review quality when caseloads spike?

For the actions behind these questions, see the Practice Library.