Plan Reassessment Drafter | Real Minds AI
Healthcare & Disability /Drafting live field guide · 8 min

Plan Reassessment Drafter

Turns a participant's file into a first-draft plan reassessment report your coordinator edits down, instead of starting from a blank page.

theater/demos/healthcare_plan-review-drafter.html · sandbox · read-only
Open
FIG. 1

The live demo, running on fabricated data. Open it to step through the full flow — every output is shown for a person to approve before anything happens.

How it would work

Reads the participant file and therapy reports, drafts the reassessment narrative with every figure traced to its source, then hands it to a coordinator to check and approve before it goes near the NDIA.

Input 01
The file + the therapy reports

The participant file summary and prior plan, the OT functional assessment and physio progress report, support notes, and the latest utilisation figures per support category.

Agent 02
Extracts, traces, drafts

Pulls goals, goal progress, therapy use, utilisation and new needs — each tagged to the report it came from — then drafts recommended supports across Core, Capacity Building, Support Coordination and Capital with figures and a plan period.

Output 03
A draft, with its working shown

A complete first-draft reassessment with every figure and its source laid out for the coordinator to review, edit and approve before anything is submitted to the NDIA.

Where it works well

It makes the transcription disappear and hands the coordinator an editable draft instead of a blank page.

  • Best for support coordinators and LACs carrying a queue of reassessments with reasonably structured files.
  • Earns its keep when the goal, its progress and the supporting report are all findable: the draft is mostly right and the work becomes checking, not typing.
  • At a queue of reassessments a month, the recaptured hours go back into the participant conversation, not the paperwork.

The slow, invisible cost a reassessment hides is transcription — a coordinator reading an OT assessment, a physio report and two quarters of support notes, then hand-copying goals, attendance and utilisation into a blank document before the real thinking starts.

Where it works badly

It is confidently wrong when the file is stale — and a clean draft looks more trustworthy than the records behind it.

  • Weak where the file is the problem — a scanned-image OT report, free-text notes that never separate a goal from an observation, last year's utilisation.
  • Wrong tool for the genuinely contested reassessment — a change-of-circumstances case or an unusual support argued for the first time, where the value is the conversation, not the draft.
The honest test

If you'd struggle to point at the report that justifies a recommended support, the draft can't justify it either.

Point it at last year's OT assessment or a utilisation snapshot nobody has refreshed and it will recommend supports for a situation that has already moved on. That is the trap.

What it doesn't do — and shouldn't

It drafts. The coordinator approves. The NDIA delegate decides. That boundary is deliberate.

WHAT IT DOES
Surfaces each extracted figure with the source report it came from
Shows an extraction confidence so the reviewer sees how complete the read was
Flags items needing extra proof — e.g. an AT request above the $1,500 low-cost threshold
WHAT IT WON’T
Decide what is reasonable and necessary
Submit anything to the NDIA
Clear a flagged item — flagging is not clearing

A reassessment shapes a real person's supports for the next plan period, and the recommendation carries obligations under the NDIS Code of Conduct. A wrong number lands on the participant, and the NDIA delegate — not the tool — makes the funding decision. The accountable person stays on the decision because the consequence is theirs.

What your data has to look like

A participant file where goals, current therapy evidence and a recent per-category utilisation snapshot are all retrievable.

48%
Typical readiness
across orgs we see, before the first job
Prior plan and its goals
Usual weak point
Current therapy evidence tied to goals
Needs shaping
Support notes that record delivered vs funded
Needs shaping
Recent utilisation snapshot per category
Needs shaping
Plan-management type on file
Usually ready
The real first job

Most providers have some of this and not all of it in good shape — goals recorded once and never tracked, notes that don't separate a goal from an observation, utilisation living in a portal export nobody has pulled. Getting that into a consistent, current, findable state is usually the real first job, and it's about how information is captured at the point of support — larger and more valuable than the drafting layer on top of it.

Right fit if…
You carry a steady queue of reassessments, not the occasional one-off
Your participant files keep goals, therapy reports and utilisation findable
You can pull a current utilisation snapshot per support category today
Most of your reassessments are recurring, evidence-backed renewals
Walk away if…
Your files are mostly scanned images and free-text progress notes
Utilisation lives in a portal export nobody has refreshed this plan period
Most reassessments are contested change-of-circumstances cases
You want a tool that decides reasonable-and-necessary for you
Open questions

The worried-buyer questions, answered straight

It can, which is why nothing it writes is final. The draft proposes supports and budget figures by reading the file; the coordinator checks every line against the evidence and their knowledge of the person before anything is submitted. It flags items that need extra proof — for example an assistive- technology request above the $1,500 low-cost threshold gets marked as needing written evidence from an AT advisor — rather than passing them through. The reasonable-and-necessary judgement stays with the person who knows the participant, and the NDIA delegate makes the funding decision.
Partly, and it will tell you where it couldn’t. It shows an extraction confidence and a source tag against each figure so you can see what it pulled from where, and it should surface a gap rather than invent a number where a note is illegible or contradicts another report. But a file that is mostly unstructured scans produces a thin draft that takes as long to fix as to write. The honest test: open three recent files and check whether a goal, its progress and the supporting report are actually findable.
No. It drafts; people decide. The coordinator or LAC still reviews the evidence, talks to the participant and shapes the recommendation, and the NDIA delegate still makes the funding decision. What it removes is the hours spent transcribing figures from reports into a blank document — time the coordinator redirects to the participant conversation and the judgement calls that need a human.
Current enough that the utilisation figures, therapy attendance and any new needs reflect this plan period. If it drafts from last year’s OT assessment or a stale utilisation snapshot, it will confidently recommend supports for a situation that has already changed. Before a reassessment is drafted, the latest therapy reports and an up-to-date utilisation pull should be in the file — the draft is only as honest as the most recent record it read.
The build runs against your own SharePoint and Word documents, and the model processes the file to produce the draft; it is not used to train a shared model. Because participant files are sensitive health and disability information under the Privacy Act and the NDIS Code of Conduct, we scope data handling, retention and access with you up front. Where the data lives and who can see it is part of the engagement, not an afterthought.
What it takes to build
4–6 weeks · 4 phases
Reused from template~65%
Bespoke to this skin~35%
stack · Claude · SharePoint · Word
What it would cost

Fixed scope, fixed price, fixed dates.

01
Bite-sized first piece
One contained change, low risk
02
Pilot build
Most builds land here
03
Embedded support
Scale on proof

Considering this for your participants?

The honest place to start is a bite-sized first piece — one contained change, low risk. Tell us where the reassessment queue hurts; we'll play it back, scope it, and show you what's possible.

More in Healthcare & Disability
Care Policy Concierge
View →
Reportable-Incident Signal Triage
View →
Prior-Auth Compiler
View →
Clinic Policy Concierge
View →
How We Work Proof Talk to us
How We Work Proof Talk to us
Ask us anything