Good morning future-focused leaders.

As AI becomes part of everyday communication, it is getting easier to spot when a message was generated by it (and harder to enjoy reading). Much of it has the same flat, generic tone that feels like what has been described as the “10,000 bowls of oatmeal” problem: everything blends into an average voice that no one actually writes in.

That is why it felt refreshing when Dr Paul Nicolarakis recently reached out to me on LinkedIn and signed off with: “P.S. This was a 100% personally hand-written message 😊”. I liked it enough to adopt the idea myself.

(On a similar note, how remarkable that we now need footnotes to prove a human actually wrote the words!)

What else we cover this week:

• AI phone agents arrive in aged care

• Three simple ways to improve your AI results

• Upcoming: Free AI integration webinar for aged care

• Upcoming: On AI, at Invox’s National CHSP Conference 

• AI combats loneliness with value-based matches

• Smart-home tech supports ageing in place

• Over-50s embrace AI tools—especially voice assistants

• Older adults co-design assistive robots

• And more...

^{LATEST DEVELOPMENTS}

ON THE HORIZON 

☎️ AI phone calls: From Airbnb to Aged Care

In brief: A couple of weeks ago I read an article where Airbnb CEO Brian Chesky explained that their AI customer service agent has already cut guest calls to human agents by 15% and will soon handle trip changes and cancellations. He noted that large language models are becoming commoditised, and the real advantage lies in sector-specific interfaces that streamline workflows. Quite an interesting read.

I felt that this idea has direct relevance for aged care in Australia, where phone calls and visit management consume large amounts of staff time. Intelligent call agents could not only reduce that burden, but also allow older people more control over their services—for example, by easily rescheduling their appointments without having to remain on hold.

What’s happening in Australia

As covered in a previous edition of AgeFriendly.AI, Silverchain is currently piloting an AI-powered virtual assistant in WA, in partnership with QLD medtech company Talius. The system confirms, reschedules, or postpones appointments through voice calls. It is being tested over three months to inform a potential national rollout.

Another example comes from Argentic Labs, co-founded by Dr Paul Nicolarakis. This Spring they are launching Zuji, a phone-based AI assistant designed for home aged care teams. Paul invited me to test the platform.

Zuji automates routine call tasks such as:

• Well-being check-ins

• Medication reminders

• Satisfaction surveys

• Home visit rescheduling

• Batch alerts

The system integrates directly with care software, supports realistic Australian voices (but also comes with multilingual options), and records and transcribes all calls. It also links back to care teams by triggering follow-up actions such as alerts or emails.

Argentic Labs arranged for the AI agent to call my mobile. I went in with low expectations, but I was genuinely surprised by how natural the conversation felt. Acting as a home care client, I even tried to derail the flow by changing the subject mid-call and raising a concern that would normally require a different response, such as saying I hadn’t been eating properly over the past few days. The agent adjusted its path rather than breaking down, showing real potential to reduce administrative load on staff while also supporting client wellbeing.

Why This Matters

Calls are the backbone of home aged care, but they are time-consuming and costly. Both Silverchain’s pilot and Argentic Labs’ upcoming platform show how AI agents can manage routine communication, ease pressure on staff, and give clients more control. These developments point to a broader shift: AI in aged care (like AirBnB) is moving to targeted, practical tools.

Watch this space. The way providers handle phone calls today may look very different within just a few years.

READY TO USE TODAY

✅ The three things you need to do to get better results from AI

In brief: This piece responds to a reader question that came in after last week’s newsletter, asking about differences between AI models. But here’s the thing: if you’ve been using AI tools and still feel underwhelmed, the issue probably isn’t the model—it’s how you’re prompting.

The details:

1. Treat AI like a knowledgeable—but new—colleague.
Most people fall into what developers call blind prompting—they type a quick request, get a decent-enough answer, and move on. But AI isn’t guessing. It’s following instructions. Like any new hire, it works best when you’re clear about the task.
Prompting should be treated as a process:

• Be specific about the AI’s role.

• Spell out what you want in terms of tone, format, length, and output.

• Review and refine based on what comes back.

2. Give examples—and counterexamples.
Telling AI what you want is helpful. But showing what not to do is often just as important. Be precise. Point out what to ignore or avoid.
Claude’s internal system guide actually includes more “don’t” rules than “always” ones. It might feel overly detailed, but these instructions reduce confusion and lead to more accurate results. So next time, include both positive and negative examples in your prompt.

3. Let the AI admit when it doesn’t know.
AI will sometimes guess if it’s unsure—and sound confident doing it. That’s risky.
Ask your AI to say “I don’t know” when it’s uncertain. And tell it to ask clarifying questions before answering anything ambiguous. This simple step can cut down on errors and save you time.Decrease in use of physical restraints and pressure ulcers in residents.

Why it matters: Prompting is more than writing the right words, or the right prompts—it’s about managing a system. Better prompts mean better outcomes and fewer surprises.

^{COMING UP}

🧩 Free webinar: your AI roadmap for aged care

 Aged care providers are turning to AI—but many face uncertainty when implementing it. This free webinar on 28 August introduces a practical toolkit to guide safe, smart AI integration under the new Support at Home model. All attendees receive templates, tools, and planning resources to take back to their teams.

📍 Hosted by Ageing Australia & AI Adoption in Aged Care Workgroup
🛠️ Includes a 5-part AI toolkit
📅 Register now 

🎤 Invox’s National CHSP Conference 2025: AI & the Human at the Centre

Only 3 weeks to go for Invox’s CHSP Conference in Melbourne (4–5 Sept). The conference promises to bring together sharp minds, innovators, and sector leaders for two days of practical insights and strategy. On the first day (4 September) I’ll be presenting on AI in Aged Care—looking at tools, ethics, and why people need to stay at the centre. If you’re attending, connect with me on LinkedIn and let’s meet in person!

📍 In-person & online
📅 See the program and register

^{QUICK HITS}

📱 Researchers at Kennesaw State University are developing Sandrapp (Supporting Adults Needing Direct Relationship App), an AI-powered tool aimed at reducing loneliness among older adults. The app uses AI to connect people with volunteers, family stand-ins, or community members who share similar values, beliefs, and life experiences. The system is intended to facilitate real-world interactions. If an older adult enjoys talking about faith or particular traditions, the app’s matching process seeks out companions with those interests. I think this is a great example of how artificial intelligence can be applied to strengthen social connection without creating more screen dependency.

🏠 Smart-home tech supports ageing in place, as older adults are increasingly turning to connected devices to maintain independence, comfort, and safety at home. NYT’s Wirecutter recently covered a wide range of tools — from voice assistants and smart lights to emergency alert systems and pill dispensers — that support daily routines and reduce risks. While the guide has a promotional tone, it still offers a useful overview of what’s possible, and opens the door to meaningful conversations about how technology might support living at home longer.

🗣️ If you assume older adults are reluctant to use AI, this data might surprise you. A recent University of Michigan survey found that over half of Americans aged 50+ use AI, with voice assistants being twice as popular as chatbots. Many use AI for entertainment and safety: 1 in 3 older adults reported using AI-powered home security devices, and 96% of them felt safer as a result. As adoption grows, there’s real potential for aged care providers to explore how these tools might complement existing supports or align with their clients’ own preferences.

🤖 At Baptist Senior Family in Pittsburgh, older adults are collaborating with Carnegie Mellon researchers to co-design assistive robotics — including voice-activated feeding devices and tools to support memory and daily tasks. It’s a strong example of what’s possible when we design with older people, rather than studying them from the sidelines or assuming what they might want. (Read the article, but feel free to ignore the somewhat reductive headline — participating in research is meaningful, but it's hardly the quintessence of purpose for older adults.)

I'm not here to hype AI. I'm here to help you understand it, use it, and learn as it evolves. Whether you're testing a new tool, using it to lighten your workload, or keeping pace with the changes, I hope you found something here worth your time.

Feel free to forward this to your network or share it with your team.

See you next Tuesday,
George