AI agents for NDIS providers

We build for NDIS providers.

We are experts in your field. Most AI firms need you to explain what makes an incident reportable, why the 24 hour clock matters, or what happens to a shift when a clearance lapses. We already understand your industry.

What we look at

Sure, your quality team is on top of it. But is the organisation?

Every provider we speak to has two or three people holding the whole compliance picture in their heads. That works until one of them takes leave and a reportable incident gets missed. We can fix that.

Below is what a first look usually turns up. Yours will differ in the detail and look much the same in shape.

provider-readinessExample, not a real provider
Ready for one agent. Not ready for five.Good people, good systems, and nothing joining them up.
What we look atHow we usually find itHow we fix it
Worker clearances and expiriesA spreadsheet and a reminderCredentials agentWatches every expiry and tells the coordinator weeks out, and keeps the history for seven years.
Incident analysis for systemic issuesNot startedIncident agentReads across your incidents for the pattern the Rules require you to find.
Shift notesWritten, not searchableShift evidence agentTurns them into evidence you can search, so they hold up under an incident or a plan review.
Claim rejections and unclaimed supportsFound at month endClaims agentReconciles weekly against what was delivered and funded, and chases rejections with the reason attached.
Restrictive practice recordsOn paperGroundwork firstA scan is a picture of text. Get these into a system and the monthly report becomes automatable. Not an agent job yet.
Who signs when it goes wrongUnderstood, not written downNot an AI jobHalf a day with your exec to write it down. Every agent we build needs a named person at the end of it.
Client management and rostering systemsIn place and usedNothing to fixThese stay. Everything above is built on top of them.
Where we would start

With a deadline somebody else set. Clearances and reportable incidents both run on clocks you do not control, so they fail loudly and they are worth fixing first.

Almost every provider we look at sits here. Good people, systems that work, and no thread running between them. That is a solvable problem, and it starts with one agent rather than five.

The panel above is an illustration of the pattern we see. It is not a real provider and not a result.

The three things that decide it

Whether an agent works here has almost nothing to do with the AI.

Can it get to the evidence?

An agent cannot check a clearance it cannot see. If the real expiry dates live in one person’s spreadsheet, that is the first job, and it is not an AI job.

Is the work the same every time?

Chasing a signature is the same every time. Deciding whether a practice was restrictive is not. We build the first kind and we say so about the second.

Who puts their name to it?

Every agent we build hands its work to a named person before anything leaves the building. If nobody owns the decision today, an agent will not fix that.

What we build

One agent, start to finish.

The table above names the agents a provider in that shape would need. This is what one of them actually does, so you can judge the rest by it.

Incident agent
  1. A support worker writes up what happened. In their own words, in the system they already use, on their phone at the end of a shift.
  2. The agent reads it against the Rules and says whether it looks reportable, and which limb it falls under. If it is unsure it says so rather than guessing.
  3. It drafts the notification with the prescribed elements filled in from what it already knows about the participant, the worker and the setting, and lists what is still missing.
  4. It tells the person who has to sign that the clock is running, and how much of it is left.
  5. A person decides and signs. Always. The agent has no authority to notify the Commission on its own.
  6. Afterwards it reads across every incident you have logged for the pattern the Rules require you to find and almost nobody has the time to look for.

Six steps. The agent does five of them. The one that carries professional judgement stays with your people, and that is true of every agent we build.

How we know this

We read the framework. All of it.

Fifteen instruments, the Practice Standards, the Rules and four quarters of Commission reporting. Three things came out of it that changed what we build.

Your highest volume document is not in the framework at all.

Shift notes appear in none of the fifteen instruments. Nobody mandates them. But they are the evidence base under your incident records, your plan reviews and your claims, which means the risk they carry is not your registration. It is your revenue.

Searched the full instrument corpus and found absent, corroborated by our own provider research

There is a legal obligation with no tooling behind it.

The Incident Rules require you to analyse your own incidents for systemic issues. Across three research streams we did not find one provider with a working way to do it. Everyone is compliant on paper and nobody is actually doing the analysis, because there is nothing to do it with.

NDIS (Incident Management and Reportable Incidents) Rules, r 12(5)

The screening wave that started in 2021 is landing now.

Those first clearances are expiring, and pending applications are up 92.8%. Separately, the Rules expect you to be able to reconstruct who held a risk assessed role on any given day, going back seven years. Most providers cannot answer that question at all.

NDIS Worker Screening Rules r 20 · NDIS Commission Quarterly Performance Report

The foundations

Enterprise plumbing, without the enterprise project.

You are handing an agent your compliance evidence. These are the parts that make that a reasonable thing to do.

Partners, and not tied to any of them

We are partners with Anthropic, OpenAI, AWS and Vercel, and certified with Anthropic. Our engineers are Claude Certified Architects who have sat and passed the CCAF exams. None of that locks you in. We build on the platform your organisation already runs, or the one that suits the job, including one you are contractually stuck with.

A partnership is not a security audit, and a certification is an engineering one. Both are real, and they are different things.

Your data stays where your policy says

Your code and your runtime live in your own hosting account. Identity, grants and audit run on our control plane, and that is self hostable if your policy requires it. We deploy into whatever region you require, including an Australian one.

Region is agreed per deployment and confirmed in writing before anything goes live.

The agents remember

An agent that forgets which clearances it already chased is worse than no agent. Ours run on Arcana, our own memory service, so context carries between shifts, months and audits instead of being re explained every time.

Arcana benchmarks at approximately 99% recall on LoCoMo, a published long conversation memory benchmark. State of the art as at August 2026.

Permissions you can take back in minutes

Grant someone access and the agent works for them. Revoke it and it refuses within minutes, with no redeployment. The same applies to one agent’s permission to call another. We run this live rather than describing it.

Verified end to end against the live system on 5 and 7 August 2026.

Whatever procurement asks for

We can meet what an enterprise requires, including SOC 2 for your deployment, and we are not limited on infrastructure. Your compliance rules shape the architecture. We build to them rather than asking you to work around ours.

Two standing commitments

Nothing reaches production without a person approving it. An automated test suite gates every deployment, and you keep that suite, which is how both sides know an upgrade is safe.

How the work runs

Four steps. You can stop after any of them.

01

Readiness check

Eight minutes, free, on your own. You get the panel above filled in for your organisation.

02

Find the constraint

Paid discovery. We sit with the people who do the job and find the one thing everything else waits on. The fee comes off the build.

03

Build the one that pays for itself

One agent, running in production, measured against what we baselined in step two.

04

Build out

The next agents, and a monthly review. The gains come from work you stop doing.

Start on your own

Eight minutes, and you will know which agent to build first.

No call, no pitch, no participant data. You answer questions about how your organisation runs and you get the readiness panel filled in for you.

It tells you where to start, and what to leave alone. Some of your work should stay with a person, some of it lives only in people’s heads, and some varies too much to call from the outside. The report separates those out, so the shortlist you get is work that can actually be built.

  • Six screens, about eight minutes, nothing to look up
  • A shortlist in order, and a section on what we would not build
  • Every figure marked directional. None of it presented as a saving
  • It never asks for participant data, and refuses it if you type it in
Start the check

Free · You get the report either way

Fit

When we are the right people to call.

  • You are a registered provider with staff, working across more than one team or more than one site.
  • You already run a client management system, rostering and finance, and they work. The trouble is what happens between them.
  • One or two people hold your compliance picture together, and you already know that is fragile.
  • You want the paperwork to shrink without anyone losing sight of a participant.

We build on top of the systems you already own. We leave clinical and behavioural judgement with the people qualified to make it, and your auditor stays the one who decides whether you pass.

We are operators who build. Name the job, the deadline and the person who signs it, and we can build you an agent that does it.