AI is good at
- Writing first drafts of emails, quotes and notes
- Summing up long documents or email threads
- Answering common questions from your own information
- Pulling details out of forms and PDFs
- Rewording things to be clearer or kinder
- Home
- Safe AI
Safe AI, set up carefully.
Good AI is set up one step at a time, with your people in charge and your data kept safe. Here’s exactly how I do it, in plain English.
My promise
Five promises to every client
These aren’t extras. They’re how every job starts, whether it’s a single workshop or a custom tool.
- 1Your people stay in chargeAI drafts, a person checks. Nothing goes to a customer unless your team is happy with it. You decide what AI can and can’t do.
- 2Your data stays yoursI use business-grade tools that don’t learn from your information, and set things up to follow the NZ Privacy Act.
- 3Start small, no big bangWe begin with one job, prove it works, then go from there. No ripping out your systems.
- 4Everything written downA one-page plan in plain English: what the AI does, what it doesn’t, and where your data goes.
- 5Nobody gets left behindEvery person who uses it gets trained, with a local person to call when they’re stuck.
Plain facts
What AI does, and what it doesn’t
AI is a tool, like a calculator for words. It’s great at some jobs and not suited to others. Knowing the difference is what keeps it safe.
People are needed for
- Final decisions, especially about money or people
- Checking facts, figures and anything legal
- Tricky, upset or sensitive customers
- Knowing your customers and your community
- Judgement, care and common sense
AI can sound sure of itself and still be wrong. That’s why a person checks anything that matters before it goes out.
Your data
Data and the NZ Privacy Act
In New Zealand, the Privacy Act 2020 sets the rules for how businesses handle personal information. Those rules apply when you use AI too. I set things up with them in mind.
- Only what’s needed. AI gets the information it needs for the job, and no more.
- Business-grade tools. I use settings and plans where your information isn’t used to train the AI.
- You know where it goes. Your plan says which tools are used and where your data is stored, including if it’s held overseas.
- Kept secure. Access is limited to the people who need it.
- People can ask. Customers can still ask what you hold about them and have it corrected.
This is a general guide, not legal advice. The Office of the Privacy Commissioner has more on privacy and AI at privacy.org.nz.
Before we go live
What I check before anything goes live
Nothing gets switched on until it’s been tested with your team and you’re happy with it.
Tested on real work
We try it on your real emails, questions or documents and check the answers together.
Data checked
What information goes in, where it’s stored, and who can see it. All written down.
A person in the loop
We agree what needs a person to check it, and make sure that step is built in.
Clear hand-over
Customer-facing tools always offer a real person, and say clearly that they’re AI.
Team trained
Everyone who uses it knows how it works, how to check it, and who to call.
Easy to switch off
If something isn’t right, it can be paused straight away while we fix it.
Free for clients
An AI use policy for your team
Every client gets a simple AI use policy template. It’s a short, plain-English page that tells your team how AI is used at your business.
- Which AI tools your team can use for work.
- What’s fine to share, and what should never go into an AI tool.
- When a person must check AI’s work before it goes out.
- Who to ask when you’re not sure.
Questions
Common questions about safe AI
Will AI replace my staff?
No. It takes tasks, not people. When the busywork shrinks, your team has more time for customers and the work they’re good at. And I train everyone, so nobody gets left behind.
Does the AI learn from my business data?
Not with the tools and settings I use. I choose business-grade AI where your information isn’t used to train the AI, and I write down exactly which tools are used.
Is it OK to put customer information into AI?
Sometimes, with the right tool and setup. Your plan and AI use policy spell out what can be shared and where. When in doubt, leave it out, and ask.
What if the AI makes a mistake?
That’s why a person checks anything important before it goes out. AI drafts, people decide. We also review how it’s going each month and improve it.
Do my customers need to know we use AI?
Being open is the best approach. Customer-facing tools like a website helper say clearly that they’re AI, and always offer a real person.
Let’s make AI work for your team
Book a free 30-minute AI chat. No jargon, no pressure, just a friendly look at where AI could help your people do their best work.