22 September 2025
Buy, Build, or Wait: A guide to AI decision-making for aged care
Two weeks ago, someone at a conference asked my opinion on smart AI glasses for aged care. Last week, it was whether we should be investing in care robots. Meanwhile, my inbox is full of questions about updating AI policies and whether staff should be using AI tools for documentation.
This is the reality of AI in aged care right now; we're bouncing between futuristic possibilities and immediate practical concerns, often missing the middle ground where real decisions need to be made.
The AI landscape can feel overwhelming. Each week brings announcements of breakthroughs from the likes of OpenAI, Google, and others, while tech vendors rush in with solutions (if we can first define our problems clearly). At the same time, no one suddenly has extra hours to dedicate to AI and technology adoption.
The AI maturity gap
Speaking to providers around the country, I can see we're caught in sorts of an AI maturity gap. On one side, we have generic, one-size-fits-all tools like large language models that don't always understand the nuances of aged care. On the other side, we have custom AI development that requires some level of technical expertise, budgets and time, which most of us simply don't have.
So, based on AI's actual current capabilities, how do we navigate this gap in-between? The answer lies in understanding when each approach makes sense.
Option 1: Buy
Purchasing an existing solution is often your best bet, and frankly, it's where most providers should start. This path makes sense when you're dealing with relatively common operational tasks.
The key questions to ask yourself are straightforward: Is there an affordable, commercially available solution that meets about 80% of your needs? Are you trying to solve a problem that many other organisations face?
If you're wrestling with consolidating data from multiple systems for reporting, or you're spending hours answering the same policy questions from staff, chances are good that someone has already built a tool that can help. Many AI vendors (including here in Australia) are pouring significant resources into improving their products, so you get the benefit of their ongoing development without having to fund it yourself.
I've seen providers get excellent results with purchased solutions for tasks like voice-to-text documentation, AI-powered policy question answering, and automated compliance analysis. The secret is being realistic about what "good enough" looks like rather than holding out for perfection.
Consider collaboration
One approach that's showing promise is providers banding together when engaging with tech developers. I've already seen successful examples of this collaborative approach, and the benefits are significant.
When providers come together, whether it's two organisations or a larger group, you have much more leverage in negotiations with technology vendors. You can share costs for customisation, pool resources for training, and learn from each other's implementation experiences. More importantly, you can speak with a unified voice about what problems you are actually trying to solve, rather than leaving developers to guess.
If you're interested in exploring collaborative approaches or connecting with others who might share similar AI needs, I'm happy to introduce you to my network of both tech developers and forward-thinking providers who are working on these challenges. Sometimes the best solutions come from conversations between people facing the same problems.
Option 2: Build
Custom development is the high-stakes option, and I'll be honest, it's probably not right for most aged care providers. But there are situations where building makes strategic sense.
Consider this route if AI is genuinely core to your organisation's innovation strategy, or if you're trying to solve a problem that no existing solution addresses. Maybe you're a larger provider with unique service delivery models, or you've identified a workflow challenge that's specific to your client demographic or geographic area.
The critical questions here are tough ones: Will building this give you a genuine long-term competitive advantage? Are you prepared to invest not just the upfront costs, but the ongoing resources needed for maintenance, updates, and technical oversight?
This is where I see providers get into trouble. They assume that because they have IT staff, they have AI development capability. Those are very different skill sets. Plus, there's the very real risk that by the time your custom solution is ready, someone else has released a commercial product that does the job better and cheaper. (Trust me, I’ve been burned myself.)
Option 3: Wait
In rare occasions, the smartest move is no move at all, or at least not yet. Waiting (if not by indecision,) can be strategic when the timing isn't right.
This approach makes sense when your use case is too niche for existing tools, when the costs of building clearly outweigh the benefits, or when there's no immediate operational risk in delaying. AI is evolving at breakneck speed. What seems expensive and out of reach today might become standard and affordable within twelve months.
But here's the thing about waiting. It can't be passive. The Aged Care Data and Digital Strategy emphasises building digital maturity and preparing for technology adoption, and this is exactly what providers should be doing during a strategic pause.
Use this time productively. Clean up your data systems. Map out your actual workflows (not what the policy manual says they should be, but what actually happens). Assess where your biggest pain points are. Train your staff and build their confidence with the tools you're already using.
When the right AI solution does come along—and it will—you'll be in a position to implement it successfully rather than struggling to retrofit it onto unprepared systems and processes.
It's an exercise, and there's room for mistakes
One final thought that I think gets lost in all the AI hype: this is still largely experimental territory. Even the most successful AI implementations I've seen have involved plenty of trial and error, course corrections, and learning from what didn't work.
That's not a reason to avoid trying, but a reason to approach AI adoption thoughtfully, with realistic expectations and a willingness to adapt as you learn. Whether you buy, build, or wait, make sure you're doing it as part of a broader strategy to enhance the human elements of care.
The organisations that will succeed aren't necessarily the ones that move fastest or spend the most money on AI. They're the ones that make decisions based on their actual needs, their genuine capabilities, and their commitment to improving outcomes for the people they serve.
This is the reality of AI in aged care right now; we're bouncing between futuristic possibilities and immediate practical concerns, often missing the middle ground where real decisions need to be made.
The AI landscape can feel overwhelming. Each week brings announcements of breakthroughs from the likes of OpenAI, Google, and others, while tech vendors rush in with solutions (if we can first define our problems clearly). At the same time, no one suddenly has extra hours to dedicate to AI and technology adoption.
The AI maturity gap
Speaking to providers around the country, I can see we're caught in sorts of an AI maturity gap. On one side, we have generic, one-size-fits-all tools like large language models that don't always understand the nuances of aged care. On the other side, we have custom AI development that requires some level of technical expertise, budgets and time, which most of us simply don't have.
So, based on AI's actual current capabilities, how do we navigate this gap in-between? The answer lies in understanding when each approach makes sense.
Option 1: Buy
Purchasing an existing solution is often your best bet, and frankly, it's where most providers should start. This path makes sense when you're dealing with relatively common operational tasks.
The key questions to ask yourself are straightforward: Is there an affordable, commercially available solution that meets about 80% of your needs? Are you trying to solve a problem that many other organisations face?
If you're wrestling with consolidating data from multiple systems for reporting, or you're spending hours answering the same policy questions from staff, chances are good that someone has already built a tool that can help. Many AI vendors (including here in Australia) are pouring significant resources into improving their products, so you get the benefit of their ongoing development without having to fund it yourself.
I've seen providers get excellent results with purchased solutions for tasks like voice-to-text documentation, AI-powered policy question answering, and automated compliance analysis. The secret is being realistic about what "good enough" looks like rather than holding out for perfection.
Consider collaboration
One approach that's showing promise is providers banding together when engaging with tech developers. I've already seen successful examples of this collaborative approach, and the benefits are significant.
When providers come together, whether it's two organisations or a larger group, you have much more leverage in negotiations with technology vendors. You can share costs for customisation, pool resources for training, and learn from each other's implementation experiences. More importantly, you can speak with a unified voice about what problems you are actually trying to solve, rather than leaving developers to guess.
If you're interested in exploring collaborative approaches or connecting with others who might share similar AI needs, I'm happy to introduce you to my network of both tech developers and forward-thinking providers who are working on these challenges. Sometimes the best solutions come from conversations between people facing the same problems.
Option 2: Build
Custom development is the high-stakes option, and I'll be honest, it's probably not right for most aged care providers. But there are situations where building makes strategic sense.
Consider this route if AI is genuinely core to your organisation's innovation strategy, or if you're trying to solve a problem that no existing solution addresses. Maybe you're a larger provider with unique service delivery models, or you've identified a workflow challenge that's specific to your client demographic or geographic area.
The critical questions here are tough ones: Will building this give you a genuine long-term competitive advantage? Are you prepared to invest not just the upfront costs, but the ongoing resources needed for maintenance, updates, and technical oversight?
This is where I see providers get into trouble. They assume that because they have IT staff, they have AI development capability. Those are very different skill sets. Plus, there's the very real risk that by the time your custom solution is ready, someone else has released a commercial product that does the job better and cheaper. (Trust me, I’ve been burned myself.)
Option 3: Wait
In rare occasions, the smartest move is no move at all, or at least not yet. Waiting (if not by indecision,) can be strategic when the timing isn't right.
This approach makes sense when your use case is too niche for existing tools, when the costs of building clearly outweigh the benefits, or when there's no immediate operational risk in delaying. AI is evolving at breakneck speed. What seems expensive and out of reach today might become standard and affordable within twelve months.
But here's the thing about waiting. It can't be passive. The Aged Care Data and Digital Strategy emphasises building digital maturity and preparing for technology adoption, and this is exactly what providers should be doing during a strategic pause.
Use this time productively. Clean up your data systems. Map out your actual workflows (not what the policy manual says they should be, but what actually happens). Assess where your biggest pain points are. Train your staff and build their confidence with the tools you're already using.
When the right AI solution does come along—and it will—you'll be in a position to implement it successfully rather than struggling to retrofit it onto unprepared systems and processes.
It's an exercise, and there's room for mistakes
One final thought that I think gets lost in all the AI hype: this is still largely experimental territory. Even the most successful AI implementations I've seen have involved plenty of trial and error, course corrections, and learning from what didn't work.
That's not a reason to avoid trying, but a reason to approach AI adoption thoughtfully, with realistic expectations and a willingness to adapt as you learn. Whether you buy, build, or wait, make sure you're doing it as part of a broader strategy to enhance the human elements of care.
The organisations that will succeed aren't necessarily the ones that move fastest or spend the most money on AI. They're the ones that make decisions based on their actual needs, their genuine capabilities, and their commitment to improving outcomes for the people they serve.