AI and Data Analytics for NZ SMEs

“AI and data analytics” is one of those phrases that means everything and nothing. For a New Zealand SME with 10 to 100 staff, it usually comes down to two practical questions: what do we already know that we can’t currently get at, and what do we do by hand that a machine should be doing?

This article answers both with real examples from our own clients — named businesses, actual builds, and what they cost in time.

What does AI actually do with your business data?

Your business already generates more useful data than you probably realise: quotes, invoices, support tickets, product specs, emails, the knowledge in your senior people’s heads. Historically, getting value out of it required analysts and expensive software, which put it out of reach for most SMEs. What changed with modern AI is that unstructured information — documents, PDFs, plain-language notes — became queryable without anyone building a data warehouse first.

Two examples of what that looks like in practice, both from Waterware, a family-owned plumbing and heating importer with about 35 staff:

Making buried knowledge searchable. Decades of technical product knowledge sat with a few key people, which meant real key-man risk — if they were away, answers were unavailable. We built Gandalf, an internal AI chatbot trained on Waterware’s own knowledge base. Anyone on the team can now ask a plain-English question and get the diagnosis, the fix, the spare part number, current stock, and shipment tracking from one interface. Concept to working model took four to five weeks.

Turning documents into data entry. Waterware’s finance team was manually raising sales orders and reconciling purchase orders against invoices — slow, repetitive, and error-prone. We built an automation that reads the supplier PDFs and does the raising and reconciling itself. The team got their time back and the error rate dropped with the re-keying.

Neither project needed a data science team, perfectly clean data, or a big budget. They needed a well-chosen problem and a few weeks of focused work.

Where should an SME start?

Not with tools. Start with an inventory of where time and knowledge leak:

1. Find the repetition. Anything your team does the same way every week — reconciliation, data entry, standard reports, answering the same internal questions — is a candidate.
2. Find the bottleneck people. If certain answers can only come from one or two heads, that’s both a risk and an AI use case, as Waterware found.
3. Pick one problem and prove it. A pilot that works in weeks tells you more than a strategy that takes months. It also teaches your team what the technology can and can’t do, which makes the second project better-chosen than the first.
4. Sort out the plumbing early. AI tools need access to your systems — Microsoft 365, your job or inventory software, your accounting platform — and that access needs to be secure. This is where AI work and managed IT meet, and why we treat security and data-access questions as part of the build, not an afterthought.

What results should you expect?

Be suspicious of anyone promising transformation. The honest wins are specific: hours per week back from a manual process, answers available in seconds instead of waiting for the right person, fewer entry errors, knowledge that survives staff changes. Those compound — a business that wins back a day a week of admin time has effectively hired a fifth of a person for the cost of a small software build.

Scale follows from there. Farro Fresh grew from one store and ten staff to six locations and over 350 staff with iT360 handling the technology side — growth is a lot cheaper when your systems and data keep up with it instead of fighting it.

How do you find the right first project?

This is exactly what our AI workshop is for: half a day with your team, walking your actual processes, and you leave with a prioritised list of automation opportunities specific to your business — whether or not you build anything with us.

Or just get in touch and describe the most annoying repetitive job in your business. That conversation is usually the fastest way to find out whether AI can take it off your hands.