It reads the handover plus the incident report, med chart and GP letter, drafts the care-plan update domain by domain with a source on every field, and routes it to the RN to correct and sign off.
Where this works well
The slow, invisible problem this surfaces is the gap between a clinical event happening and the care plan catching up to it. After a fall, a medication change, or an unplanned weight loss, the registered nurse has to read the incident report, the med chart, the GP letter and the handover, then assemble a structured plan update — an hour of compiling that often finishes after the shift, unpaid. The drafter does the compiling: it pulls the care needs out of those documents, attributes each to its source, and hands the RN a first draft to correct.
It earns its keep where the trigger is a discrete, well-documented clinical event and the source documents actually exist — a post-fall review with an incident report attached, a dietitian's weight-loss alert against a charted weight history. The role that benefits most is the clinical care coordinator or RN carrying a unit of residents, where care-plan reviews stack up against direct care time. At that scale, recapturing the compiling time is the difference between reviews done on shift and reviews done in "pyjama time".
Where it works badly
It will be confidently wrong when the source documents disagree or are stale. If the medication chart hasn't been updated for the new Webster pack, the draft reflects the old medications and reads as authoritative. If the handover blurs a one-off observation with an ongoing goal, the draft inherits that blur. The failure mode is not gibberish — it is a tidy, plausible plan built on a document that was already wrong, which is harder to catch than an obvious error.
It is also not worth the effort for residents whose care plans are stable and whose updates are minor and routine — the compiling there is short, and the review overhead can exceed the saving. The honest test: pull your last ten care-plan updates and ask how many were triggered by a documented clinical event with attachments, versus drifting changes nobody logged against a source. If most are the latter, the documents the drafter needs to read aren't there yet, and that is the real first job.
What it doesn't do — and shouldn't
It drafts; the registered nurse decides. The tool surfaces extracted care needs, suggests care-time allocations by domain, and flags where its draft falls short of the facility's AN-ACC care-minute target or where a source is missing — but it never sets the falls intervention, judges whether the nutrition plan is adequate, or commits anything to the resident record. The RN reviews, edits, and signs off, and only then does the plan enter the system.
That boundary is deliberate, and in aged care it is non-negotiable. The care plan is a clinical and legal document; the registered nurse holds the duty of care, the accountability under the Strengthened Aged Care Quality Standards, and the judgment about this specific resident that no document fully captures. A reportable incident under SIRS turns on clinical criteria a draft cannot assess. The tool accelerates the assembling so the RN's attention goes to the parts that need a clinician — it does not, and should not, make the clinical call.
What your data has to look like for this to work
This needs three things to be true of your records. First, the trigger documents have to exist and be attached — an incident report for a fall, a medication chart that reflects the current Webster pack, a GP letter — because the drafter can only attribute a care need to a source it can read. Second, your handover and progress notes need to distinguish a goal from an observation from an action; most don't, consistently, and that is where the draft quality is won or lost. Third, the documents need to be current to the event, not weeks behind it.
Most facilities have some of this in good shape and some not. The med chart in a system like Manad Plus may be clean while free-text handovers are inconsistent across shifts and agency staff. Getting the source documents structured and current is usually a matter of how information is captured at the point of care — not buying a new tool — and it is the work that makes everything downstream possible. It is typically a bigger and more valuable piece than the AI layer that sits on top, and it is the part we help with first.
Could the draft recommend the wrong care and put a resident at risk?
It can, which is exactly why a registered nurse signs off every plan before it enters the resident record — the tool drafts, the RN decides. Each line carries the source it came from (handover, incident report, med chart, GP letter) so the RN can check it against the original rather than trust a summary. It also flags its own gaps: where the drafted care minutes fall short of the facility's AN-ACC target, or where a source is missing, it surfaces that for the RN to confirm rather than quietly filling the hole.
Our progress notes and handovers are inconsistent between staff and shifts — will this still work?
Partly, and the honest answer is that messy free-text notes are the usual first job, not the AI. The drafter can only attribute a care need to a source if that source actually records it: a fall has to be in the incident report, a medication change in the med chart, a goal distinct from an observation in the note. Where handovers blur those together, the draft inherits the blur. We typically look at how your handover and progress notes are captured before adding any AI on top.
Does this replace the registered nurse who writes the care plan?
No. It removes the compiling — reading three or four documents and assembling a structured first draft — so the RN's time goes to the clinical judgment: is the falls intervention right for this resident, is the nutrition plan adequate, does the family's request change anything. The RN still owns the plan and the sign-off. Capacity moves from typing to deciding, not out the door.
How current does the data have to be for the draft to be trustworthy?
The draft is only as current as the documents it reads. If the medication chart predates a Webster-pack change, or the handover misses an overnight event, the draft will reflect the stale picture — confidently. It works best triggered off a fresh clinical event (a post-fall review, a weight-loss alert) with the relevant attachments dated and attached, not run against a record nobody has touched in weeks.
Where does the resident's clinical data go — does it leave our systems?
That is a design decision we make with you, not a default. The pattern is built to run against your own clinical records with a private retrieval layer, so resident data stays within your environment rather than being posted to a public chatbot. Aged-care records carry obligations under the Privacy Act and the Aged Care Act, so the data path, retention and access are scoped explicitly as part of the build.
Will using this create a SIRS or compliance problem if a draft is wrong?
The reportable-incident obligation under the Serious Incident Response Scheme sits with the provider and the clinician, and the draft does not change that — a fall is reportable on its own criteria regardless of how the plan was written. Because the RN reviews and signs off before anything is recorded, the audit trail shows a human decision, and the source attribution on each line makes that review faster to evidence under the Strengthened Aged Care Quality Standards.
Most of it is template work we’ve already done.
Fixed scope, fixed price, fixed dates.
The cost band reflects the engagement shape, not a per-feature line item. We work on fixed scope, fixed price, fixed dates — see the services catalogue for what falls inside each band.
Considering this for your org?
The honest place to start is a bite-sized first piece — one contained change, low risk. Tell us where it hurts; we’ll play it back, scope it, and show you what’s possible.