🇦🇺 Australian-built  ·  Hosted in Australia  ·  NDIS & SCHADS compliant
Provider Login 0415 576 533 CareIQ IT →
CAREIQ AI · CARE SIGNALS

AI that reads every clinical note — and shows you exactly what it found.

Care Signals scans every progress note your support workers submit. When it finds language associated with unreported incidents, medication risks, or safeguarding concerns, it flags the exact phrase, explains why it matters clinically, and tells you what to do next. Not a vague alert. A specific finding.

Australian-built 2 months free No setup fee Active on every plan
▲ 3 unreviewed
Margaret T.  ·  SIL House Brisbane
⚠ Nutrition / Clinical Deterioration
SEV 7/10 88% confidence
Flagged phrase from note
"wouldn’t eat dinner… said she didn’t want to bother anyone"
Why this matters clinically
Phrases combining meal refusal with social withdrawal can indicate depression or acute physical deterioration, particularly in clients with dementia.
Recommended action
Conduct a wellbeing check within 4 hours. Document in care plan. Contact GP if appetite or withdrawal persists beyond 24 hours.
✓ Acknowledge ⚠ Raise Incident Dismiss

Always on. Always reading. Shows its work.

Every note submitted by a support worker is analysed automatically — before it reaches your triage queue. No staff behaviour change required. No extra forms. No manual review of every note.

👁️
Reads every note automatically

The moment a support worker submits a progress note, Care Signals analyses it in the background. There is no opt-in, no tagging, no extra step for staff. Every note — every shift, every client — is read. Signals with confidence below 60% are silently discarded, so your triage queue stays clean.

📝
Shows the exact phrase, not a category label

Other tools tell you a note contains a "risk flag." Care Signals shows you the verbatim excerpt — 5 to 15 words lifted directly from the note — so you know immediately what triggered the concern without reading the whole note. Nothing is paraphrased. Nothing is summarised away.

⚙️
Gives a specific action with a timeframe

Each signal includes a plain-English explanation of why the phrase is clinically concerning, followed by a specific recommended action — not "monitor client." Something like: "Complete a falls risk assessment and update the care plan within 24 hours." The timeframe is part of the output, every time.

The transparency that changes clinical governance.

Most risk-monitoring tools operate as a black box: a note goes in, a flag comes out. Care Signals is designed to do the opposite — every finding is fully traceable back to the words that triggered it.

When a support worker submits a progress note, Care Signals sends the note content to a clinical analysis model trained to identify eight categories of risk in the Australian NDIS disability and aged care context.

For each concern it detects with sufficient confidence, it produces a structured signal containing four pieces of information: the type of concern, the verbatim triggering phrase, a plain-English clinical explanation, and a specific recommended action with a timeframe.

The signal is stored against the note and the client, made immediately visible in the triage queue, and — if your notification settings allow — pushes an alert to the responsible coordinator.

1

Support worker submits note

No extra fields. No prompts. The note is written exactly as it always has been.

2

Analysis runs automatically in background

Runs asynchronously — no delay to the support worker, no page reload, no waiting.

3

Signal appears in triage queue

Sorted by severity then recency. Unreviewed signals are listed first. Each one shows verbatim phrase, why it matters, and what to do.

4

Coordinator takes action in one step

Acknowledge, raise a formal incident, or dismiss. The signal records who reviewed it and when. The audit trail is automatic.

⚠ Unreported Incident — Fall Risk
Severity 8/10 Confidence: 91%
Flagged phrase — verbatim from the note
"Harold grabbed the rail and had to be steadied, said his legs felt weak"
Why this is clinically concerning
Descriptions of a participant needing physical steadying suggest an unrecorded near-miss fall. Combined with a complaint of leg weakness, this indicates potential neurological or cardiovascular change requiring formal assessment before the next shift.
Recommended action
Complete a formal near-miss fall incident report within 2 hours. Conduct a falls risk assessment and notify the on-call nurse. Update the mobility section of the care plan before assigning the next AM shift.
Raised by: Care Signals  ·  Note by: S. Johnson  ·  14 Jun 2026, 09:47 am  ·  Client: Harold B.

Care Signals is included on every CareIQ plan. $6/user/month — no setup fee.

Book a Demo →

Eight clinical risk categories, monitored on every note.

Care Signals is trained on the specific language patterns that appear in NDIS disability and aged care progress notes when clinically significant events occur but are not formally reported.

🚫
Unreported Incidents

Falls, near-miss falls, injuries, altercations, property damage, and medication errors described in note content without a corresponding incident form raised. This is the most commonly missed category.

Trigger patterns
fell hit grabbed the rail had to be steadied bruise noted
💊
Medication Concerns

Missed doses, refusal to take medication, apparent incorrect dose descriptions, side-effect language, and notes that mention medication without a corresponding MAR entry. Flags before the 24-hour window closes.

Trigger patterns
refused medication spat it out missed morning meds rash after tablet
🛠️
Behavioural Escalation

Aggression toward staff or other participants, self-harm indicators, significant deviation from a participant's documented behavioural baseline, and language suggesting acute emotional distress.

Trigger patterns
aggressive screaming self-harm hit staff distressed
💋
Clinical Deterioration

Mobility changes, cognitive changes, unexplained weight loss or gain, reduced responsiveness, and vital signs concerns described in note language but not formally captured as a vital signs entry.

Trigger patterns
confused wandering legs weak not responding breathless
🪶
Skin Integrity Concerns

Wounds, pressure areas, unexplained bruising, redness, skin tears, and any description of skin condition that has not been formally documented as a wound management entry.

Trigger patterns
redness on heel bruise on arm skin tear wound looked worse
🍕
Nutritional & Infection Concerns

Meal refusal, significant appetite change, dysphagia language, fever indicators, signs of infection (redness, discharge, odour, increased confusion), and dehydration language that warrants clinical follow-up.

Trigger patterns
wouldn't eat choking feverish smelled of infection not drinking
8
Clinical risk categories monitored
60%
Minimum confidence threshold before flagging
<2s
Analysis latency — runs asynchronously, no note delay
100%
Of submitted notes analysed — no sampling

One queue. Every concern. Across all clients.

Care Signals aggregates every unreviewed finding into a single triage view — sorted by severity first, then recency. Coordinators start with the most urgent concerns and work through to lower-priority signals, without switching between client profiles.

Each signal links directly to the originating clinical note and to the client's profile. Raising an incident from the triage queue pre-populates the incident form with the flagged phrase, the supporting context, and the recommended action — so the formal record takes seconds, not minutes.

Filter by severity level, signal type, client, date range, or review status
Three actions per signal: Acknowledge · Raise Incident · Dismiss
Each reviewed signal records the reviewer's name and timestamp automatically
Raise Incident pre-fills the incident form from signal data — no double-entry
Full audit trail: clinical note → signal → incident in one traceable chain
⚠ Care Signals — Triage Queue
All Sev 7+ Unreviewed
8
Harold B.
Unreported incident — fall risk
09:47 am
7
Margaret T.
Nutrition / clinical deterioration
10:12 am
5
Dorothy W.
Medication concern — refusal
11:35 am
3
Francis O.
Skin integrity — redness noted
Yesterday
4 signals shown · 2 unreviewed · Sorted by severity
★★★★★

"Before Care Signals, we were relying on coordinators to read every note looking for concerns — which meant things got missed during busy periods. Now every note is read automatically. When something comes through, we know exactly what was said, why it matters, and what to do. That's given us real clinical governance confidence — and during our last NDIS audit, the reviewers specifically noted how thorough our incident trail was."

Operations Manager
SIL Provider — Queensland

Not all risk monitoring is the same.

The difference between a useful finding and a useless flag comes down to specificity. Here is what Care Signals gives you that a standard keyword-alert system does not.

Care Signals Standard keyword alerts
What you see when a flag is raised Verbatim excerpt from the note — the exact phrase that triggered it "Risk flag detected" or a category label
Clinical explanation Plain-English explanation of why that specific phrase is concerning None — coordinator must determine relevance themselves
Recommended action Specific action with a timeframe (e.g. "within 2 hours") Generic ("review this note", "discuss with team")
Confidence scoring 0–100% confidence score, flags only above 60% threshold Binary match — keyword present / not present
Severity ranking 1–10 severity score — triage queue sorted automatically Not available or manually assigned
Direct incident creation Raise Incident pre-fills the incident form from signal data Separate manual workflow
Audit trail Reviewer name + timestamp recorded per signal, linked to note and incident Often limited or manual

Clinical governance that scales with your roster.

As a provider grows — more staff, more clients, more shifts — the clinical governance burden grows proportionally. The coordinator who could read every note when the organisation had 12 participants cannot do the same at 40. Neither can they identify patterns across clients and shifts that no single note reveals.

Care Signals scales without adding headcount. Whether you have 7 staff or 70, every note is read to the same standard. The triage queue surfaces only the signals that warrant human attention — not the full note volume.

During NDIS audits and Quality and Safeguards Commission reviews, the Care Signals audit trail demonstrates systematic clinical oversight — every note reviewed, every concern actioned, every decision recorded.

See It in Your Trial →

Audit Confidence

NDIS Audit Trail
Every signal shows who reviewed it, when, and what action was taken — satisfying Quality and Safeguards Commission evidence requirements for clinical monitoring.
Incident Traceability
The chain from clinical note → Care Signal → incident report is traceable in one click. No gap between a concern being recorded and a formal incident being raised.
Coverage Guarantee
Every submitted note is analysed — not a sample, not the high-priority clients only. If a note was submitted, it was read. This is demonstrable in the audit log.

Common questions about Care Signals.

No. Support workers write their progress notes exactly as they always have. Care Signals runs automatically in the background after submission — there is no opt-in, no extra form, no prompt to the worker. The entire system is invisible to the person writing the note. This is by design: the moment staff know their notes are being scanned, note quality changes in unpredictable ways. Passive analysis gives you a more accurate clinical picture.
The signal appears in the Care Signals triage queue, sorted by severity. The reviewing coordinator sees the exact flagged phrase, the clinical explanation, and the specific recommended action. They then choose one of three responses: Acknowledge (the concern is noted and managed within the existing care plan), Raise Incident (pre-fills and submits a formal incident report linked to the originating note), or Dismiss (the signal is not clinically significant in this client's context). Every action is recorded with the reviewer's name and timestamp.
Care Signals can be enabled or disabled per company from the platform settings. The 60% confidence threshold filters low-quality signals automatically. If you operate in a specialist service context — for example, a provider focused entirely on behaviour support where distress language appears frequently as baseline — you can contact CareIQ to discuss threshold adjustments for your account. Signal types themselves (fall risk, medication concern, skin integrity, etc.) map to the standard NDIS incident reporting categories, which are not service-specific.
Manual review at scale is not consistent. A coordinator reviewing 40 notes after a busy shift is reading quickly, under fatigue, and may not recognise the combination of phrases that together constitute a clinical risk. Care Signals reads every note with the same level of attention regardless of shift volume, time of day, or staffing pressure. It also identifies cross-sentence patterns — for example, meal refusal mentioned early in a note combined with withdrawal language mentioned later — that can be missed when skimming. The coordinator's time is redirected from reading every note to reviewing only the subset with a confirmed concern.
Yes. Care Signals is active on every CareIQ plan that includes the clinical notes module — every plan, no exceptions. There is no separate charge for Care Signals, and no plan tier that locks it behind an upgrade. Clinical governance should not be a premium add-on.
Care Signals is designed to fail silently. If the AI service is unavailable, or if analysis returns no signals above the 60% confidence threshold, no signal is raised and the note is stored normally. The system never blocks note submission — care documentation is never held up by Care Signals processing. Signals with confidence below 60% are discarded rather than flagged, to keep the triage queue clinically meaningful rather than noisy.
CareIQ uses OpenAI's API with default API data handling policies in place. Clinical note content is transmitted to the model for analysis and is not retained by CareIQ beyond what is stored in your account database. CareIQ does not use your clinical note data to train models. For enterprise deployments with specific data sovereignty requirements, contact CareIQ to discuss arrangement options.

Every note your team writes, read with clinical precision.

Care Signals is included on every CareIQ plan. $6/user/month, 2 months free, no credit card required to start.

No setup fee  ·  2 months free  ·  Australian hosted  ·  Cancel any time