20 April 2026
The three tiers of AI in aged care
A simple way to work out where your organisation sits — and what to do next.
If you work in aged care, you've probably noticed AI has gone from "interesting idea" to "already happening in our organisation" in a pretty short space of time. Someone on your team is almost certainly using it. Your board has probably asked about it. And somewhere between the hype and the horror stories, you're trying to work out what it actually means for your organisation.
To help make sense of it, I find it useful to think of AI in aged care as falling into three tiers. Most organisations I talk to are somewhere on this spectrum, often without realising it.
Before we get to the tiers, though, it helps to start with a picture of what most aged care organisations actually look like today — before AI enters the conversation at all.
To help make sense of it, I find it useful to think of AI in aged care as falling into three tiers. Most organisations I talk to are somewhere on this spectrum, often without realising it.
Before we get to the tiers, though, it helps to start with a picture of what most aged care organisations actually look like today — before AI enters the conversation at all.
|
Subscribe to my newsletter - tap below ↓
|
Every aged care organisation is some version of this diagram.
This is the landscape AI lands on. It doesn't arrive into a blank page — it arrives into this. And where AI sits in the picture is really what separates the three tiers.
- The blue circles are the software systems you use — your care management platform, your rostering system, your finance system, your HR system, and so on.
- The green figures are your people — staff, managers, admin, clinicians.
- The dashed lines show who has access to what. Some people use one system, some use several. Some systems are accessed by many people, others by only a few.
- The blue lines between circles are the places where your systems actually talk to each other. You'll notice not all of them do.
This is the landscape AI lands on. It doesn't arrive into a blank page — it arrives into this. And where AI sits in the picture is really what separates the three tiers.
Tier 1 - Generic AI Tools
|
Subscribe to my newsletter - tap below ↓
|
Tier 1 - Generic AI Tools
In Tier 1, AI sits with the people. Your staff sign up for tools like ChatGPT, Claude, Copilot or Gemini — free, or a small monthly fee per person — and start using them alongside whatever systems they already use. The systems themselves don't change; the people just have a new helper on the side.
People use these tools to draft emails, tidy up newsletters, summarise meeting notes, get a first cut of a policy, or help shape a funding narrative. And honestly, a lot of the time it's working — individual productivity goes up, people feel less stuck, and the quality of written work often improves.
The catch is that at this tier, AI use tends to be scattered and invisible to management. Different people using different tools in different ways, sometimes pasting in information that really shouldn't leave the organisation. Privacy is the big risk here, and it's not theoretical.
If this sounds like your organisation, the sensible next step isn't to shut it down — it's to get a simple set of Responsible Use Guidelines in place. Something short that tells your team what's okay to do, what isn't, and which tools are preferred. This one step alone takes you from "hoping for the best" to "managing the obvious risks."
In Tier 1, AI sits with the people. Your staff sign up for tools like ChatGPT, Claude, Copilot or Gemini — free, or a small monthly fee per person — and start using them alongside whatever systems they already use. The systems themselves don't change; the people just have a new helper on the side.
People use these tools to draft emails, tidy up newsletters, summarise meeting notes, get a first cut of a policy, or help shape a funding narrative. And honestly, a lot of the time it's working — individual productivity goes up, people feel less stuck, and the quality of written work often improves.
The catch is that at this tier, AI use tends to be scattered and invisible to management. Different people using different tools in different ways, sometimes pasting in information that really shouldn't leave the organisation. Privacy is the big risk here, and it's not theoretical.
If this sounds like your organisation, the sensible next step isn't to shut it down — it's to get a simple set of Responsible Use Guidelines in place. Something short that tells your team what's okay to do, what isn't, and which tools are preferred. This one step alone takes you from "hoping for the best" to "managing the obvious risks."
Tier 2 — Custom AI for your organisation
|
Subscribe to my newsletter - tap below ↓
|
Tier 2 — Custom AI for your organisation
In Tier 2, AI moves off the individual and onto one (or more) of your systems. This tier actually shows up in two different ways, and it's worth understanding both.
The first way is AI that's already built into the software you use. A lot of the care management, rostering and HR platforms in the Australian market have started shipping AI features inside their existing products — AI-assisted progress notes, summarised client histories, meeting transcription and summary tools, AI search across your policies, draft responses to common queries. You haven't changed systems; your existing vendor has added an AI layer on top of the thing you already pay for. Sometimes it's included, sometimes it's an add-on.
The second way is a custom AI assistant set up specifically for your organisation — trained on your policies, your procedures, your service information and your templates. Instead of a staff member asking ChatGPT a general question, they can ask your AI things like:
Both versions are Tier 2 because the AI is attached to a system, not to an individual. The answers come from your actual documents and workflows, not generic content from the internet. This is particularly useful for staff onboarding, plain-English policy access, service eligibility guidance, care plan and progress note drafting, and drafting first-cut responses to complaints and feedback. It also reduces your reliance on a few key people who "just know how things work here" — which is a real risk in our sector, where experienced staff are hard to hold onto.
The cost varies depending on which version you're looking at — an add-on feature from your existing vendor might be a small incremental fee, while a custom-built assistant for your organisation involves a modest per-user monthly fee plus a setup investment.
Privacy risk is lower than Tier 1 because the data stays in an environment you control. It's still worth asking the right questions about data storage and retention — especially with vendor-added AI features, where the detail of how your data is used matters a lot — but you're now operating with a lot more visibility.
The next step here, if you're curious, is twofold: ask your existing vendors what AI features they already offer (you may be paying for some of this already), and pick one high-friction process in your organisation that isn't already covered by those features and build a small pilot around it.
In Tier 2, AI moves off the individual and onto one (or more) of your systems. This tier actually shows up in two different ways, and it's worth understanding both.
The first way is AI that's already built into the software you use. A lot of the care management, rostering and HR platforms in the Australian market have started shipping AI features inside their existing products — AI-assisted progress notes, summarised client histories, meeting transcription and summary tools, AI search across your policies, draft responses to common queries. You haven't changed systems; your existing vendor has added an AI layer on top of the thing you already pay for. Sometimes it's included, sometimes it's an add-on.
The second way is a custom AI assistant set up specifically for your organisation — trained on your policies, your procedures, your service information and your templates. Instead of a staff member asking ChatGPT a general question, they can ask your AI things like:
- "What's our process for responding to a complaint?"
- "Is this client eligible for XYZ service under our criteria?"
- "Draft a care plan using our template for a client in XYZ situation."
- "What does our policy actually say about X?"
Both versions are Tier 2 because the AI is attached to a system, not to an individual. The answers come from your actual documents and workflows, not generic content from the internet. This is particularly useful for staff onboarding, plain-English policy access, service eligibility guidance, care plan and progress note drafting, and drafting first-cut responses to complaints and feedback. It also reduces your reliance on a few key people who "just know how things work here" — which is a real risk in our sector, where experienced staff are hard to hold onto.
The cost varies depending on which version you're looking at — an add-on feature from your existing vendor might be a small incremental fee, while a custom-built assistant for your organisation involves a modest per-user monthly fee plus a setup investment.
Privacy risk is lower than Tier 1 because the data stays in an environment you control. It's still worth asking the right questions about data storage and retention — especially with vendor-added AI features, where the detail of how your data is used matters a lot — but you're now operating with a lot more visibility.
The next step here, if you're curious, is twofold: ask your existing vendors what AI features they already offer (you may be paying for some of this already), and pick one high-friction process in your organisation that isn't already covered by those features and build a small pilot around it.
Tier 3 — Integrated AI suite
|
Subscribe to my newsletter - tap below ↓
|
Tier 3 — Integrated AI Suite
In Tier 3, AI wraps around the whole picture. It isn't bolted onto one system or one person — it sits across care management, HR, finance and rostering, or forms a layer that connects the platforms you already use.
The clearest example in our sector is claims and funding. A Tier 3 system can pull from rostering, clinical records and finance at the same time, prepare claims based on services actually delivered, spot gaps or duplicates before submission, and cross-check claimed services against what was delivered and what's eligible. That's the difference between manually chasing revenue leakage after the fact and having it flagged in close to real time.
Beyond claims, Tier 3 is where you start to see AI-assisted rostering, automated operational and compliance reporting, and alerts on anomalies across services and workforce — things that currently require someone to notice a pattern in a spreadsheet after the month has closed.
The cost is higher, and it's usually a monthly figure rather than a per-user one. But the value is at a different level too: less revenue leakage, faster reporting, a much tighter feedback loop between what's happening on the ground and what leadership can see, and audit-readiness as a by-product rather than a scramble.
Risk at this tier is generally low (if the vendor is reputable). The questions worth asking are the boring ones: where is our data stored, who owns it, what happens if we leave, and how does this integrate with what we already have? An unreliable Tier 3 vendor can create more problems than the tier solves.
If you're genuinely considering this, the next step is honest: assess your system readiness and look seriously at vendor options. Most organisations aren't ready for Tier 3 yet, and that's fine.
So where are you?
Most aged care organisations I talk to are somewhere between Tier 1 and Tier 2 — usually further along at Tier 1 than they realise, and thinking about Tier 2. Very few are genuinely at Tier 3 yet, despite what the marketing suggests.
The useful thing about the tiers isn't picking the "right" one. It's being honest about where you actually are, and deciding what the next sensible step looks like for your organisation.
Where to from here?
These conversations are genuinely some of my favourite ones to have. If you want to think out loud about where your organisation sits and what might be worth doing next, I'm up for a chat.
I've got a pretty good sense of which vendors are doing interesting work in aged care at each tier, so if a specific introduction would help, I'm happy to make it.
And, if this piece has sparked something, share it internally. Most of the good decisions in this space start with a conversation inside your own team.
In Tier 3, AI wraps around the whole picture. It isn't bolted onto one system or one person — it sits across care management, HR, finance and rostering, or forms a layer that connects the platforms you already use.
The clearest example in our sector is claims and funding. A Tier 3 system can pull from rostering, clinical records and finance at the same time, prepare claims based on services actually delivered, spot gaps or duplicates before submission, and cross-check claimed services against what was delivered and what's eligible. That's the difference between manually chasing revenue leakage after the fact and having it flagged in close to real time.
Beyond claims, Tier 3 is where you start to see AI-assisted rostering, automated operational and compliance reporting, and alerts on anomalies across services and workforce — things that currently require someone to notice a pattern in a spreadsheet after the month has closed.
The cost is higher, and it's usually a monthly figure rather than a per-user one. But the value is at a different level too: less revenue leakage, faster reporting, a much tighter feedback loop between what's happening on the ground and what leadership can see, and audit-readiness as a by-product rather than a scramble.
Risk at this tier is generally low (if the vendor is reputable). The questions worth asking are the boring ones: where is our data stored, who owns it, what happens if we leave, and how does this integrate with what we already have? An unreliable Tier 3 vendor can create more problems than the tier solves.
If you're genuinely considering this, the next step is honest: assess your system readiness and look seriously at vendor options. Most organisations aren't ready for Tier 3 yet, and that's fine.
So where are you?
Most aged care organisations I talk to are somewhere between Tier 1 and Tier 2 — usually further along at Tier 1 than they realise, and thinking about Tier 2. Very few are genuinely at Tier 3 yet, despite what the marketing suggests.
The useful thing about the tiers isn't picking the "right" one. It's being honest about where you actually are, and deciding what the next sensible step looks like for your organisation.
Where to from here?
These conversations are genuinely some of my favourite ones to have. If you want to think out loud about where your organisation sits and what might be worth doing next, I'm up for a chat.
I've got a pretty good sense of which vendors are doing interesting work in aged care at each tier, so if a specific introduction would help, I'm happy to make it.
And, if this piece has sparked something, share it internally. Most of the good decisions in this space start with a conversation inside your own team.