Everyone's on personal ChatGPT accounts
Half the team already use it, on personal logins, with no agreement about what goes in and nothing the business can switch off when they leave.
AI set-up & automation
AI that does the boring bits, safely.
Tools chosen and set up on business accounts, an AI-use policy your staff can actually follow, and two or three automations that save real hours — quotes, the inbox, documents.
Sound familiar?
Where most small businesses are with AI
Somewhere between "everyone's quietly using it" and "we really should do something". Both are fixable.
A client's details, pasted somewhere they shouldn't be
A whole email thread dropped into a chatbot to "tidy it up" — name, address, medical history and all. Nobody meant harm; nobody had been told the rule.
The quote takes an hour to write
Same structure, same clauses, same photos — typed from scratch every time, usually at night, which is why it goes out late.
"We should be doing something with AI"
Said in every meeting for two years. Nobody owns it, so nothing happens — or something happens with no policy behind it.
What we set up
Six things that make AI useful
Not a strategy deck. Accounts, a rule, a session and a few automations — in place and in use within weeks.
The right tool, on a business account
Microsoft 365 Copilot, Claude or ChatGPT — whichever fits how you work. Business tiers, where the vendors state your data isn't used to train their models, on accounts the business owns and can switch off.
An AI-use policy
One page: which tools, on which accounts, what never goes in, who to ask. Written for your business, not downloaded, alongside your other security and compliance policies.
A staff session
A session with your team on what the tool is good at, what it gets wrong, and the three things they'll use it for on Monday.
First automations
Two or three, picked for hours saved: quote drafting from job notes and photos, inbox triage that sorts and drafts replies, document templates that fill themselves.
For practices: AI scribes, governed
If an AI note-taker is in the room there's consent to get, a storage question to answer and Ahpra guidance to follow. We set it up so the practitioner stays in charge of the record.
Monthly check-in
What's being used, what isn't, what's next — and the policy updated when the tools or the rules change. Optional, month to month, and included in the Complete IT plan.
Done already
An electrical contractor, two automations
For an electrical contractor we manage, we set up quote drafting and inbox triage on business accounts, with a policy and a staff session behind them. Quotes start from a draft built from the job notes instead of a blank page; the inbox sorts itself into quote requests, supplier email and the rest, with replies drafted for the owner to check and send.
- Business accounts, owned by the businessSet up in the business's Microsoft 365 tenant, not a personal login — so access ends when employment does.
- The owner stays in the loopNothing goes out without a person reading it. The AI drafts; a human sends.
- A policy the team actually readOne page, plain words, agreed before the first automation went live.
- Checked monthlyWhat's used, what's ignored, what to build next.
The rules that apply
AI doesn't get a pass on the law
The privacy and professional rules you already work under apply to a chatbot as they apply to email. The short version, and where the long one lives.
What never gets pasted in
Client records, health information, tax file numbers, passwords, and anyone's personal information beyond what the task needs. If it would be a problem in an email to a stranger, it's a problem in a chatbot.
Privacy, and the December 2026 change
If the Privacy Act covers your business — it covers every health service provider, whatever its size — you need to know where personal information goes. From 10 December 2026, privacy policies must also explain any automated decision-making that uses personal information and significantly affects people. We word yours carefully; the OAIC's guidance is the reference.
Health practitioners: Ahpra guidance
Ahpra and the National Boards have published guidance on using AI in practice: the practitioner stays responsible for the record and the care, and patients are told and asked before an AI scribe is used. We set tools up so that's the easy path.
How it works
Look, set up, check in
-
01
Look
A short look at how work flows: where the typing is, where the copying between systems is, and what people already use AI for.
-
02
Set up
Accounts, the policy, the staff session and the first two or three automations — built on your business tenant and tested with the people who'll use them.
-
03
Check in
A monthly check-in if you want it: what's being used, what isn't, what to build next — and the policy kept current. Already part of the Complete IT plan.
What it costs
One set-up, one optional habit
Prices are what you pay — Tidewire isn't registered for GST, so none is added. Tool subscriptions are billed by the vendor to the business.
AI set-up
From $1,500 one-off
Tools on business accounts, the AI-use policy, a staff session and the first two or three automations — scoped and quoted before anything starts.
- Tool chosen and set up on business accounts
- AI-use policy written for your business
- Staff session
- Two or three automations built and tested
Monthly check-in
$200 a month · optional
For businesses that want the automations kept working and the next one built. Month to month, no lock-in.
- Usage reviewed each month
- Policy updated as tools and rules change
- Next automation scoped
- Included in the Complete IT plan
Bigger builds — connecting AI to your job-management system, say — are quoted first. If you're not happy in the first 90 days, every monthly fee you've paid comes back — set-up fees and advertising spend excepted.
FAQ
Fair questions
Which tool should we use?
It depends on where your work lives. If the team is in Outlook, Teams and Word all day, Microsoft 365 Copilot sits inside those — and it needs a tidy tenant underneath. For drafting, documents and longer reasoning, Claude or ChatGPT on a business plan is often the better fit. We recommend one, not three, and set it up on accounts the business owns.
Is our data safe?
Safer than on personal accounts. On business tiers the vendors state that your prompts and files aren't used to train their models, and the business controls the accounts. That still doesn't make a chatbot the place for client records, health information or passwords — the policy covers what stays out, and sensitive systems aren't connected unless there's a reason.
Do staff need training?
A session, not a course. Most people have used a chatbot; what they haven't had is someone show them the three things it's genuinely good at in their job, what it gets wrong, and the rule about what never to paste in. That session is part of every set-up.
Can you automate our quoting?
Usually, yes. Job notes and photos go in, a draft quote in your format comes out, and you check it before it's sent. It won't price the job for you, and it shouldn't: it removes the typing, not the judgement. Connecting it to simPRO, Tradify, ServiceM8 or AroFlo is quoted as a project. More for trades.
Is this only for offices?
No. The first business we did this for was an electrical contractor. Quotes and the inbox are the same problem at a desk or on a roof, and the phone is where most of it gets used.
What does the monthly check-in do?
Keeps it alive. A look at what's being used, a fix for what's broken, the policy updated when the tools or the rules change, and the next automation scoped. Without it, set-ups tend to drift back to personal accounts. Everything else we do.
Start with the boring bits
A free 15-minute chat about where the hours go. You'll get a straight answer on whether AI helps — and sometimes the answer is "not yet".
Book a free 15-minute chator call 0439 649 099 · email hello@tidewire.com.au