Skip to content

Buying guide

How to choose an aged care AI vendor

Aged care providers are being offered AI tools by care software companies, start-ups, consultants and large technology firms. These are the questions to put to every one of them, ourselves included, and a simple way to score the answers.

Reviewed 15 November 2026

Aged care operations manager comparing two vendor proposals at a meeting table with a checklist

In short

To choose an aged care AI vendor, judge them on how they handle residents' and clients' information and what happens when the AI gets something wrong, as much as on features.

Ask where data is stored, whether it is used to train their models, which sub-processors see it, how staff review AI output, how urgent matters reach a person, what is logged, and how you get your data back if you leave. Check accessibility, language support, how pricing is structured and who supports you after go-live.

Be wary of vendors who claim certifications they cannot show, promise integrations before checking your systems, quote savings without measuring your baseline or let AI make care decisions. Score each vendor on the same table, weight the factors that matter most to your service, and ask for written answers so they become part of the contract.

Start with your own requirements

Before talking to vendors, write one page on what you want the AI to do, who will use it, what information it will touch and what must never happen (for example, a family being told something about a resident's condition by an AI). This keeps demonstrations on your problem and makes answers comparable.

Involve the people who will live with the decision: the privacy officer, the IT lead or managed service provider, a care manager, someone from the admin team and, for larger providers, the risk or governance committee. The Guidance for AI Adoption from the National AI Centre (Department of Industry, Science and Resources) sets out six practices for responsible AI use, starting with deciding who is accountable. That is a useful structure for the internal conversation.

Data and privacy questions to ask every vendor

These questions matter most because aged care AI handles health information, which is sensitive information under the Privacy Act. Ask for written answers.

  • Where is our data stored and processed, including backups? Which countries?
  • Is our data, or the content of calls and documents, used to train your models or anyone else's? Can that be switched off in the contract?
  • Which sub-processors (cloud hosts, AI model providers, transcription services) receive our data, and where are they located?
  • How long is data kept, and can we set the retention period?
  • Who in your company can access our data, and is that access logged?
  • How will you tell us about a data breach, and how quickly?
  • Will you support our privacy impact assessment with documentation?

Safety, oversight and operations questions

The second group of questions is about what happens day to day, and on the bad days. The OAIC's guidance on commercially available AI products says organisations should consider whether a product has been tested for its intended use and how human oversight can be built into processes. Turn that into concrete questions.

  • Human review: which outputs are checked by staff before they reach a family, client or record?
  • Escalation: how does the system recognise urgent or distressed situations, and how fast does it reach a person?
  • Audit logs: can we see what the AI did, when, with which information, and who approved it?
  • Testing: how will the system be tested with our scenarios before go-live, and who signs off?
  • Exit and data return: if we leave, in what format do we get our data, how quickly, and how is it deleted?
  • Accessibility: does the interface meet WCAG 2.2 for staff and families using screen readers or larger text?
  • Languages: which languages are supported for callers or families, and how is quality checked?
  • Pricing structure: set-up, monthly, per call, per minute or per user, and what triggers extra charges?
  • Support: who do we call when something goes wrong, during which hours, and in which time zone?

Apply this to your service

Talk it through with a Yes AI consultant in a free 30-minute call, including whether AI fits your service at all.

Red flags

Some answers should slow a decision down. None automatically rules a vendor out, but each needs a satisfactory explanation in writing.

  • Certifications claimed (for example SOC 2 or ISO 27001) that the vendor cannot show a current certificate or report for
  • An integration with your care management or rostering system promised before anyone has checked that system's API and your licence
  • Savings or return figures quoted before your own call volumes or admin hours have been measured
  • AI that makes or recommends clinical decisions, medication decisions or care decisions without a qualified person deciding
  • Testimonials or case studies that cannot be verified, or references the vendor will not let you contact
  • Vague answers on data location, model training or sub-processors
  • No clear way to export your data or end the contract
  • Pressure to sign quickly or lock in a long term before a trial

A scoring table you can use

Score each vendor from 0 to 3 on each factor (0 no answer or unacceptable, 1 weak, 2 acceptable, 3 strong), multiply by the weight, and add up. Adjust the weights to suit your service: a provider with many culturally and linguistically diverse families might give languages a weight of 3. A low score on any weight-3 factor should be resolved before signing, whatever the total.

Example vendor scoring table (weights are a starting point)
FactorWeight (1 to 3)What a strong answer looks like
Data location3Named regions in writing, including backups and sub-processors
Training on your data3Contractual commitment that your data is not used to train models
Sub-processors2Full list with locations, and notice of changes
Human review and escalation3Defined review points and tested escalation to named staff
Audit logs2Logs you can access, showing actions, data used and approvals
Exit and data return2Export format, timeframe and deletion confirmation in the contract
Accessibility1Stated WCAG 2.2 conformance level with known gaps listed
Languages1Supported languages listed and tested with your scenarios
Pricing structure2All charges and triggers set out in writing
Support2Australian business hours support, named contact, escalation path
Evidence2Verifiable references and a trial on your own scenarios
Aged care board members reviewing a document together at a meeting table

Run a short trial before committing

A demonstration shows what the vendor wants you to see. A trial on your own scenarios shows how the product handles your residents, clients, families and staff. Keep it narrow: one use case, a defined period, written success measures agreed beforehand, and a baseline taken before it starts.

Include awkward cases in the trial: a caller who is upset, an enquiry in another language, a family member asking for clinical information the AI must not give, a document with a missing field. Seeing how the system handles these cases is the main reason to run a trial.

Holding us to the same standard

We are one of the vendors you might be comparing, so the same questions apply to us. We hold no SOC 2, ISO 27001 or similar certification. We have no partnerships or ready-made integrations with care management, rostering or finance software; connections are custom work on a quote after checking each vendor's API and terms. We do not quote savings before measuring your baseline, and our tools do not make care decisions. Ask us for written answers to every question on this page, as you would anyone else.

The vendor checklist tool on this site turns these questions into a form you can send to each supplier.

How it works in practice

  1. 1

    Write your requirements

    One page: the task, the users, the information involved and what must never happen.

  2. 2

    Send the same questions to every vendor

    Use the questions on this page or the vendor checklist, and ask for written answers.

  3. 3

    Score the answers

    Use the weighted table, with your own weights, and resolve any low scores on weight-3 factors.

  4. 4

    Trial the top one or two

    A narrow trial on your own scenarios with success measures agreed in advance.

  5. 5

    Put answers in the contract

    Data location, training, sub-processors, breach notice, exit and data return belong in the signed agreement.

Common questions

What should I ask an AI vendor about data privacy?

Ask where data is stored and processed, whether it is used to train models, which sub-processors receive it and where they are, how long it is kept, who can access it, and how you will be told about a breach. Get the answers in writing.

Does an aged care AI vendor need SOC 2 or ISO 27001?

Neither is legally required, but if a vendor claims one, ask for the current certificate or report and check its scope covers the service you are buying. A vendor without certification should be able to describe its security controls in writing.

Is it a problem if the AI vendor stores data overseas?

Not automatically, but Australian Privacy Principle 8 applies when personal information is disclosed overseas, and you remain accountable. Many providers prefer Australian hosting for health information. Know exactly where data goes and reflect it in your privacy policy.

Should we do a privacy impact assessment before buying AI?

For tools that handle residents' or clients' health information, it is good practice. The OAIC publishes a guide to undertaking privacy impact assessments, and a good vendor will give you the documentation you need to complete one.

How do we compare AI vendors fairly?

Send every vendor the same questions, score the written answers on a weighted table, and trial the strongest one or two on your own scenarios with measures agreed in advance.

What are the red flags when choosing an AI vendor?

Unverifiable certifications or testimonials, integrations promised before checking your systems, savings quoted before measuring your baseline, AI making care decisions, vague answers on data location and training, and no clear exit.

Sources and further reading

Related pages

AI Privacy Guide for Aged Care

Aged care records hold some of the most sensitive information a person has. Before any AI tool touches it, you need to know where the information goes, who can see it and what the law expects.

AI Governance for Aged Care Boards

Directors of aged care providers already carry legal governance duties for quality and safety. AI is now part of the operations those duties cover, and boards need a simple, regular way to oversee it.

Aged Care Software Integrations

An AI tool is only useful if its work ends up in the systems your staff already use. This page explains the four usual ways of connecting AI to aged care software, which method suits each type of system, and what to check before anyone promises an integration.

AI Risks and Limits in Aged Care

We build AI tools for aged care providers, and we think the risks deserve as much attention as the benefits. This page sets out what can go wrong, how to reduce it, and the jobs AI should not be given.

Talk it through with us

A free 30-minute video call about where admin time goes in your service and which task is worth automating first. No obligation, and no resident information needed.

Call Book free assessment