Before You Buy AI Tools: 7 Questions to Ask Your Team First

The very first things to do before buying AI tools are agree why you want AI, find out what your team is already using, name an owner, decide what information is off-limits, talk honestly with staff about what changes for them, choose one problem to test, and agree how you'll measure the result. None of that needs a credit card. For an established business with 10 to 200 staff, these conversations are what separate a tool that sticks from a subscription that quietly gathers dust.

Think of this as the pre-flight checklist: the questions to ask before implementing AI in your business, and who should be in the room. For the bigger picture on costs, tools and governance, see our complete guide to AI for small and medium businesses in Australia.

Why the first step isn't a tool

Buying an AI tool first is a bit like buying a forklift before checking whether you have a warehouse. In our experience, when AI trials fizzle in small and medium businesses, it's rarely because the software was bad. More often, nobody agreed what it was for, who owned it, or what "working" would look like.

Australia's privacy regulator makes a similar point. The OAIC's guidance on privacy and commercially available AI products states that "AI products should not be used simply because they are available." That's sound privacy advice, and sound business advice too.

Seven questions to answer before you buy AI tools

Work through these with your leadership team first, then with the people who'll use the tools. Most SMEs can cover them in a handful of meetings.

1. Why are we doing this, really?

Start with the business outcome, not the technology. "Win more tenders without adding headcount", "get invoices processed faster" or "answer customer enquiries the same day" are useful answers. "Because our competitors are" is not.

Ask each leader to write down the one thing they'd most like AI to fix. Five different answers isn't a problem: it's your shortlist.

2. What are our people already using?

In many businesses, the AI journey has already started, approved or not. Staff use free chatbots on personal accounts to draft emails, summarise documents and tidy spreadsheets. This "shadow AI" can be a significant hidden risk, because nobody knows what client information has gone where.

Run a short, no-blame survey: which tools, for which tasks, on which accounts. You'll learn where the real demand sits and who your informal champions are.

Then check what you already pay for. Microsoft says Copilot Chat is included at no additional cost with qualifying Microsoft 365 business subscriptions, and Google now includes Gemini features in its Workspace business plans. Before buying anything new, find out what's already switched on (or switched off) in your own admin settings.

The National AI Centre's free Guidance for AI Adoption includes an AI register template, which is a tidy way to record what you find.

3. Who owns AI in our business?

Someone senior needs to be accountable for AI decisions, risks and results. In a business of 10 to 200 staff, that's usually a director, the general manager or the operations manager, not the most enthusiastic junior with a chatbot habit.

Deciding who is accountable is the first of the six essential practices in the National AI Centre's guidance, released in October 2025. The owner doesn't need to be technical. They need the authority to say yes, no and "not yet", and the time to follow through.

4. What information is off-limits?

Agree your data rules before anyone signs up to anything. List what must never go into an AI tool without approval: client files, health information, employee records, pricing and anything covered by a confidentiality agreement.

For many established SMEs, privacy law is part of this. The Privacy Act generally applies to businesses with annual turnover above $3 million, plus some smaller ones such as private sector health service providers. The OAIC's guidance recommends, as best practice, that organisations don't enter personal information, especially sensitive information, into publicly available generative AI tools.

If a tool you're considering would use personal information to make decisions that could significantly affect people, such as screening job applicants, note the new transparency rule. From 10 December 2026, businesses covered by the Privacy Act must set out in their privacy policy the kinds of personal information used, and the kinds of decisions made, where a computer program uses personal information to make, or do something substantially and directly related to making, decisions that could reasonably be expected to significantly affect people's rights or interests. This is general information only, so check with a privacy adviser about your situation.

5. What changes for our people?

This is the conversation many businesses skip, and it often decides whether AI gets used at all. Some staff will be excited. Others will quietly wonder whether AI is coming for their job. Say out loud what the goal is, what won't change, and how people will be supported to build new skills.

Be specific. "We want AI to draft the weekly site report so supervisors spend less time at a laptop" lands far better than "we're rolling out AI". Invite the people who do the work to shape how it's used.

There's a legal side too. Modern awards and enterprise agreements generally include consultation terms that require employers to consult employees about major workplace changes, including changes to technology, that are likely to have significant effects on them. A small pilot may not meet that threshold, but changes to roles or headcount are more likely to, and changes to regular rosters or ordinary hours of work have their own consultation requirement under modern awards and enterprise agreements. The Fair Work Ombudsman's consultation and cooperation best practice guide is a good starting point, and it's worth getting workplace relations advice before any change that affects roles.

6. Which problem goes first, and how will we know it worked?

Pick one specific, repetitive task that costs real time across a team, then measure it before you change anything: how long it takes, how often it happens and how much rework it creates. Without a baseline, you'll never know whether the tool earned its keep.

Good first candidates happen weekly or more, follow a similar pattern each time, and involve reading, writing or moving information between systems. For a structured way to rank options by value and effort, our guide to AI readiness assessments walks through it.

7. What will it really cost, and who signs off?

Licences are the visible cost. The bigger costs are usually time: setting tools up, training each team, writing the rules, checking output and keeping workflows running. Agree a budget envelope that covers all of it, and agree who approves spending beyond a small pilot.

Then set your buying criteria before you sit through a single vendor demo.

A one-page pre-purchase checklist

Use this as the agenda for your first AI conversations. Each one should end with something written down.

Once this is done, you're ready to compare tools properly, and vendor conversations get shorter and far more productive (for you, at least, if not for the salesperson).

Frequently asked questions

Should we write an AI policy before buying AI tools?

Yes, at least a first version. A one or two page policy covering approved tools, information that's off-limits and who checks output heads off most early problems. The National AI Centre's guidance includes a free AI policy guide and template you can adapt.

Do we have to consult staff before introducing AI?

It depends on the change and on your award or enterprise agreement. Consultation terms generally apply to major changes that are likely to significantly affect employees. Even where it isn't required, involving staff early tends to make adoption smoother. Get workplace relations advice for your situation.

How long should this preparation take?

For most SMEs, a few weeks of focused conversations rather than months of planning. The aim is enough clarity to run one well-chosen pilot, not a perfect five-year strategy.

How NovoPro can help

If you'd like help working through these questions, a NovoPro AI & Automation Audit is built for this stage. We look at how your teams really work and aim to give you a ranked list of the opportunities most likely to make or save money, before you commit to new tools. Many businesses also run AI Training and Adoption alongside, so their people build skills and confidence from day one.

Not sure where to begin? Book a 15-minute Problem First Call and bring the one problem you'd most like to fix.

Written by Kesley Kernes, Founder and Principal Consultant of NovoPro. Kesley holds an MBA and Lean Six Sigma certification, previously established an RPA Centre of Excellence, and helps small and medium businesses across Newcastle, the Hunter and Australia use AI and automation to make and save money.