The Reportable Incident Clock: What NDIS Providers Lodge, and When
The reportable incident clock does not start when the report is written. It starts the moment a support worker becomes aware — which is usually hours before anybody opens a form.
That gap is where most of the risk in NDIS incident management actually lives. Not in the form itself, which is short. Not in the notification, which is a button. In the hours between something happening and somebody writing it down properly.
What has to be lodged, and by when
The NDIS Commission’s rules are specific, and they are the reason this is not a job for memory:
- 24 hours for the most serious incidents — the death of a participant, serious injury, sexual misconduct, or assault.
- Five business days for the other reportable incidents — for example unauthorised use of a restrictive practice, or an incident involving a child.
- A register of every incident, reportable or not, kept and reviewed — because patterns across small incidents are what audits and safeguarding reviews actually look for.
Providers know these numbers. The failure is almost never ignorance of the deadline. It is the distance between the shift and the person who files.
What goes missing is the evidence, not the notification
By the time an incident is lodged, the notification itself is usually fine. What is thin is everything underneath it: the shift note written at the end of a nine-hour day, the witness account collected two days later, the photo that is still on someone’s phone, the exact time the participant was last observed.
An auditor reads the file, not the form. A file that says “participant became distressed” with no contemporaneous notes is a file that says nothing — and it is the version that ends up in a safeguarding review six months later.
A notification lodged on time with no evidence behind it is a deadline met and a risk kept.
What an agent does about it
The Incident Report Agent works at the point of care rather than the point of filing. A worker describes what happened in plain language, on a phone, at the end of the shift. The agent structures it into the reportable format, captures the witnesses, times and attachments alongside it, flags whether the incident is reportable and which clock applies, and reminds the right person before the window closes.
What reaches the Commission is the same account the worker gave at the time — not a reconstruction typed up a week later by somebody who was not there.
The uncomfortable part
Most providers will say their staff know what to report. Ask a support worker at 9pm on a double shift what constitutes a reportable incident and you will get a hesitant answer. That is not a training failure. It is a design failure: the decision is being made by the least available person, at the least available moment.
If the process needs a worker to remember policy in the middle of an incident, it will fail quietly and often.
What we build
We build automation and digital employees for Australian NDIS providers. Incident reporting is one of twenty specialised AI agents, and it connects to the same shift and participant record as claiming and rostering. See the NDIS provider setup.
Start with one month of incidents
Send us last month’s incident register. We will show you which entries would have been stronger if they had been captured at the shift, and what an agent would have held onto — dates, times, witnesses, attachments — without adding a single form to the worker’s day.