AccAnalysisAccAnalysis
Healthcare & Social Care

The session recorded, transcribed and turned into the report it has to produce

A disability support centre was running assessments, support planning and evaluation sessions on paper, then typing them up. We built the workflow that captures the session and populates the record from it.

At a glance
Status
Delivered
Region
Japan
Engagement model
Dedicated team

Status: Delivered. Described from our own delivery records. Client identity, brand and commercial terms are withheld. Figures, where given, cover the period stated and nothing beyond it.

The organisation

Context

A disability support centre conducting three kinds of structured session: initial assessment, individual support planning, and regular evaluation and support reviews.

Each produces documentation that is both a care record and a compliance artefact.

The brief

The problem

  • Sessions were recorded on paper and digitised afterwards, so every session was handled twice.
  • The administrative burden sat on the same staff conducting the sessions, which is the worst possible allocation of their time.
  • Transcription after the fact loses detail, and the detail lost is not random — it is whatever the note-taker did not think mattered at the time.
  • Report data was re-entered from notes, introducing variation between records that should be structured identically.
The work

What we built

A web application covering all three session types

With the structure of each session reflected in the system rather than in a paper form.

Automated recording and transcription

Of the session, integrated with language models for analysis.

Automated population of report data

From the analysed transcript into the structured record — with the practitioner reviewing and confirming rather than typing.

Method

How we delivered it

  1. 1

    Map the three session types

    With the practitioners who run them, including what each report must contain and why.

  2. 2

    Build the structured record first

    So there is a defined target for anything the automation produces.

  3. 3

    Add capture and transcription

    Tested on real session audio for the language and terminology actually used.

  4. 4

    Add analysis and population

    Always as a draft for practitioner confirmation — never as a final record.

  5. 5

    Pilot with a small group of practitioners

    Across a full review cycle.

Sequence

How it was phased

PhaseDurationWhat happens
1Session mapping
2–3 wks

Three session types, report requirements, compliance needs

2Structured record
3–4 wks

Data model, forms, permissions

3Capture & transcription
3–4 wks

Recording, transcription, terminology handling

4Analysis & population
3–4 wks

Draft report data, practitioner review step

5Pilot
3–4 wks

Small practitioner group, full review cycle

Indicative phasing for work of this shape. Actual duration varies with data quality, access and decision speed.

Hand-over

What the client keeps

  • The application and its source
  • The session and report data models
  • Transcription and analysis configuration
  • Practitioner documentation
Stack
Web applicationOpenAI modelsTranscriptionPythonStructured data model
Keep reading

Related

Related case studies

Tell us what your current system can't do, and we'll tell you what it would take to change that.