Quick answer
Responsible AI in a maintenance business comes down to three plain rules: the models are internal and never trained on customer data; the technology runs on modern infrastructure with strong, up-to-date security and data-protection controls; and AI is used as an assistant only, with the final decision always resting with the human in charge. Those rules are what let a conservative client take the efficiency of automation without taking on a data or accountability risk. If a contractor cannot state them plainly, that is the answer to whether to trust its AI near your buildings and your residents.
How AI earns trust on your buildings.
Key points
- Responsible AI is three rules: internal models never trained on customer data, strong security, and a human making the final call.
- Not training on customer data is the line that protects a client's and residents' information from leaking into a model.
- Security and data-protection controls are what keep the whole stack safe, not an afterthought bolted on later.
- Human-in-charge means the technology assists and flags, but a person decides and stays accountable for the outcome.
- The useful test of a contractor is whether it can state plainly what its AI touches and who makes the decisions.
Bringing AI onto someone's buildings only works if the governance is clear, because the client is trusting you with information about their property and, in an aged-care or housing setting, about vulnerable residents. The efficiency is real and worth having, but not at the cost of a data risk or a decision made by a machine that nobody owns. So the responsible version is not complicated, and it is worth saying out loud rather than burying in a policy nobody reads. Here are the rules we hold to and why each one matters.
Rule one: never trained on customer data
The first and most important rule is that our models are internal to MTN and we never train them on customer data. This is the line that matters most for a client, because it means information about your buildings, your programme and your residents does not get absorbed into a model and cannot resurface somewhere else. Automation can use your data to do a job, checking an invoice, drafting a report, without that data ever becoming training material. Keeping those two things separate, using data versus training on it, is the heart of doing this responsibly.
Rule two: security and data protection
The whole stack runs on modern infrastructure protected by strong, up-to-date security and data-protection controls. That is not a line of marketing; it is the practical reason the efficiency does not come with a risk attached. Handling resident information also sits under the Privacy Act 2020, so information security is a legal duty as well as good practice, part of the same vetting-and-trust standard we hold our people to. A contractor casual about data security is a risk to your compliance, not just its own.
Rule three: the human makes the call
AI is used as an assistant only, and the final decision always rests with the human in charge. The tool can flag an invoice mismatch, a missing photo, an inconsistent scope, but a person reviews it and decides. This keeps accountability where it belongs: with people who can be asked why, not with a system that cannot. It is the rule that runs through everything from our invoicing checks to our quality reporting, and it is the difference between AI that supports a trusted contractor and AI that replaces judgement with a black box.
| Rule | What it means | What it protects |
|---|---|---|
| No training on customer data | Data is used, not absorbed | Your and residents' information |
| Strong security controls | Modern, protected infrastructure | The whole stack, and your compliance |
| Human in charge | AI assists, a person decides | Accountability and judgement |
How to judge a contractor's AI
The practical test for a client is specificity. A contractor that uses AI responsibly can tell you, in plain words, what the AI actually does, whether it touches your data and how, and who makes the final decision. A vague reassurance, or a refusal to be specific, is the answer. This is the client-facing side of how a trades business actually uses AI: the useful conversation is not whether a contractor uses AI, which everyone now claims, but whether it can be precise and honest about how.
Frequently asked questions
Do you train your AI on our data?
No. Our models are internal to MTN and we never train them on customer data. Automation may use your data to do a specific job, such as checking an invoice or drafting a report, but that data is never fed into a model that keeps it. Using data and training on it are different things, and we only ever do the former.
What does human-in-charge mean?
It means AI is used only as an assistant, and a person always makes the final decision and carries the accountability. The tool can flag an issue, a billing mismatch or a missing photo, but a person reviews it and decides what happens. Accountability stays with people who can explain why, not with a system that cannot be asked.
Is my residents' data safe if you use AI?
Yes. Resident information is handled under strong, up-to-date security and data-protection controls and under the Privacy Act 2020, and it is never used to train a model. The same vetting-and-trust standard we apply to our people applies to information security, because in an aged-care or housing setting protecting resident data is both a legal duty and a matter of trust.
What is the difference between using data and training on it?
Using data means a tool reads it to do a job, such as checking an invoice against a scope, and then that is the end of it. Training on data means feeding it into a model that retains it and could reproduce it later. We use customer data to do jobs but never train models on it, which keeps your information from ending up anywhere it should not.
How can I tell if a contractor uses AI responsibly?
Ask for specifics. A responsible contractor can tell you plainly what its AI does, whether and how it touches your data, and who makes the final decision. Clear, concrete answers are a good sign; vague reassurance or reluctance to be specific is the real answer. The test is precision and honesty, not whether the contractor says it uses AI.
Does using AI mean fewer people are accountable?
No, accountability stays fully human. The point of the human-in-charge rule is that AI never removes a person from the decision; it just helps them make it. Every invoice, quality sign-off and report still has a person behind it who owns the outcome. The technology assists the accountable people, it does not replace them or dilute who is responsible.
If you are letting AI-assisted work anywhere near your buildings and residents, the questions to ask are simple, and we answer them plainly: our models are internal and never trained on your data, the stack is properly secured, and a person makes every call. Ask us to walk you through exactly what the tools touch and what they never do, and we will.