NDIS 7-Year Records and 90-Day Claims: Building a Claim Evidence Workflow That Survives a Request

From 1 December 2026, NDIS providers must retain records for seven years, and the window to submit a claim shrinks from two years to ninety days. Late or poorly evidenced claims will be treated as debts, and providers are already being warned that administrative oversight will not count as an exceptional circumstance. For a provider still running claims through a mix of paper notes, shared drives, and whoever remembers to invoice, that’s a genuinely different operating requirement than the one they’ve been working under.

The reform changes two things at once: how long evidence has to survive, and how fast it has to move from a support being delivered to a claim being lodged. Most providers can handle one of those changes. Handling both together is what actually requires redesigning the workflow rather than tightening the existing one.

What has to be retained, and in what form

The obligation covers more than invoices. Service notes, roster records, signed agreements, evidence that supports were linked to a participant’s plan goals, and the claim documentation itself all need to be retrievable for seven years. A provider that’s currently storing some of this in a practice management system, some in email, and some in a support worker’s personal notes has seven years of inconsistency to reckon with, not just seven years of storage.

The practical test is whether a full claim file, for any support delivered at any point in that window, could be pulled together and produced within a matter of days if the NDIA asked for it. For most providers today, the honest answer is that it would take longer than that, and take someone reconstructing the file from several different places rather than retrieving it from one.

Why ninety days changes the billing cycle end to end

Under the old two-year window, a backlog of unbilled or unevidenced support could sit for months without becoming an emergency. Under the new window, that same backlog becomes a claim that can no longer be lodged at all, with the value of the support delivered turning into a genuine loss rather than a delayed invoice.

That timing pressure runs backward through the whole cycle. Service notes need to be completed close to the point of delivery, not caught up in a batch at the end of the week. Roster approvals need to happen promptly enough that billing isn’t waiting on a signature from three weeks ago. Evidence of goal alignment needs to be captured as part of the service note itself, not reconstructed later when a claim is being prepared. Each of these steps used to have slack built in. None of them can anymore.

Building a claim evidence workflow that survives a request

A workable workflow treats each support delivered as generating one file that moves through a defined sequence: service note completed and logged, roster or shift confirmed, evidence of plan alignment attached, claim prepared and lodged, and evidence archived in a single retrievable location. Every step has an owner and a target turnaround measured in days, not weeks.

The businesses managing this well track a small number of numbers weekly: days from support delivered to claim lodged, the proportion of claims with complete documentation on first submission, and the number of claims sitting unresolved past an internal warning threshold well before the ninety-day cutoff. A provider watching these numbers catches a slow-moving problem while it’s still an internal process fix, not an unrecoverable claim.

Which parts of this move to a dedicated admin role

Support delivery, goal-setting, and any clinical or care judgment stay squarely with frontline and clinical staff. Everything downstream of that, chasing an incomplete service note, checking a roster against what was actually delivered, assembling the file for a claim, filing evidence in the right participant record, and tracking claims against the ninety-day clock, is process-driven work that a dedicated claims or documentation coordinator serving NDIS providers can own directly, onshore or offshore.

Take a mid-size support provider running eighty active participants across a team of support workers. Before redesigning the workflow, claims were prepared in batches by whichever admin staff member had time that week, with no single person tracking which files were complete and which weren’t. A dedicated coordinator running the same eighty participants through a defined weekly cycle, chasing incomplete notes within days rather than weeks, cut the average time from support delivered to claim lodged by more than half.

What the first ninety days of building this looks like

The first few weeks are spent mapping the current path a support takes from delivery to claim, participant by participant if needed, against what the seven-year and ninety-day requirements actually demand. From there, a coordinator takes over the mechanical parts of the cycle: chasing notes, checking rosters, assembling files, and tracking the claims clock, while clinical and frontline staff continue owning delivery and any genuine judgment calls about a participant’s plan.

By the end of that period, most providers have a single, consistent claim evidence file structure running across their whole participant base, with a coordinator who can produce any file on short notice rather than reconstructing it under pressure.

What to check before December 2026

• Pick five recent claims at random and time how long it would take to produce a complete, retrievable evidence file for each.
• Map where service notes, rosters, and claim evidence currently live, and whether that’s one place or several.
• Identify who owns tracking claims against the ninety-day window, and whether that’s currently happening at all.

If your claim evidence still lives across several different places and several different people’s memory, that redesign is worth starting well before December, not after a claim gets rejected. HIPPO’s earlier work on a repeatable NDIS billing and reconciliation engine covers what that coordinator role looks like day to day. Book a Connect Session