How is AI used in data science? A practical guide for New Zealand

Picture a supermarket chain in Auckland trying to predict how many loaves of bread each store will sell next Tuesday. A data scientist pulls together years of sales records, weather data, public holidays and local events, then trains a model that forecasts demand store by store. That forecast — the part that learns patterns and makes a prediction — is AI doing a job inside a data science project. So the question "how is AI used in data science" has a fairly grounded answer: AI is one of the tools a data scientist reaches for, usually to make predictions or find patterns that would be impossible to spot by hand.

Let's pull the two terms apart before we put them back together, because the overlap confuses a lot of people starting out in Wellington, Christchurch and beyond.

Data science and AI are not the same thing

Data science is the broader craft. It covers getting data, cleaning it, exploring it, visualising it, and drawing conclusions that someone can actually act on. A lot of that work has nothing to do with AI at all — a well-made dashboard or a solid statistical test counts as data science.

Artificial intelligence is narrower. It's the set of techniques that let a computer learn from data and make decisions or predictions without being explicitly programmed for every case. Machine learning (ML) is the workhorse branch of AI you'll meet most often, and deep learning sits inside that again.

So when people ask "does data science relate to AI?", the honest answer is yes, closely, but one contains the other. AI is a capability you use within the wider data science process. We wrote a fuller breakdown in our guide to how data science and AI actually fit together if you want the longer version.

Where AI actually shows up in a data science project

A typical project runs through a few stages, and AI tends to appear at a specific point rather than everywhere.

First comes the unglamorous work: collecting data, fixing missing values, removing duplicates, and reshaping messy spreadsheets into something usable. Data scientists often say this is the bulk of the job, and no model fixes bad inputs. AI barely features here.

Then comes exploration — charts, summaries, spotting which variables matter. Still mostly human judgement and statistics.

AI enters when you want the system to learn something. A few concrete examples a beginner can picture:

• A bank in New Zealand flags unusual transactions. A model learns what normal spending looks like for each customer and raises a flag when something is off. That's anomaly detection.
• A health provider estimates which patients are likely to miss appointments, so clinics can send reminders. That's a classification model.
• A streaming service suggests what you might watch next based on what similar viewers enjoyed. That's a recommendation system.

In each case, the data scientist frames the problem, prepares the data, picks and trains the model, then checks whether it's actually any good. The AI is the engine; the data scientist is the one deciding where to point it.

Lately, generative AI has added another layer. Data scientists now use large language models to summarise text, draft code, and build tools that answer questions over a company's own documents. If you've read our walkthrough on building a retrieval-based AI application, that's a modern slice of the same discipline.

AI vs data science: which is "better"?

This question comes up constantly, and it's a bit like asking whether an engine is better than a car. One lives inside the other. Still, if you're choosing where to focus your learning, a side-by-side comparison helps.

	Data science	AI / machine learning
Main goal	Turn data into decisions and insight	Build systems that learn and predict
Day-to-day work	Cleaning, analysis, visualisation, reporting	Designing, training and tuning models
Core skills	Statistics, SQL, Python, communication	Maths, algorithms, Python, model evaluation
Typical NZ roles	Data analyst, data scientist, BI analyst	ML engineer, AI engineer, applied scientist
Good starting point?	Yes — broader and more forgiving	Easier once you have data science basics

For most people starting in New Zealand, data science is the sensible entry door. You pick up the foundations — Python, SQL, statistics, how to tell a clear story with data — and AI becomes a natural next step rather than a wall you hit on day one. Trying to jump straight into deep learning without those basics usually ends in frustration.

Can I learn machine learning in 3 months?

Short answer: you can learn to build and use machine learning models in three months. You won't become a research scientist in that time, and anyone promising that is overselling.

Here's a realistic picture of what three focused months can get you. If you already know a bit of Python and some maths, you can reach the point where you train models on real datasets, understand why they succeed or fail, and deploy a simple one. If you're starting from scratch, three months is better spent on the data science groundwork — programming, data wrangling, statistics — with ML as the final stretch.

The thing that makes or breaks a three-month timeline is structure. Self-teaching from scattered tutorials tends to leave gaps you don't notice until a project breaks. A guided path with projects and feedback moves faster because you're not guessing what to learn next. Our data science and AI bootcamp is built around that idea — real datasets, regular feedback, and a sequence that goes from fundamentals to working models in a set timeframe.

A quick, honest aside: the hardest part for most beginners isn't the maths. It's the data cleaning and the habit of checking whether a model's results actually hold up. Models are easy to run. Trusting them is the skill.

What this means if you're starting out in New Zealand

Demand for people who can work with data sits across retail, finance, health, agriculture and government here, and the roles increasingly expect at least a working knowledge of machine learning. You don't need to master every AI technique before you're employable. You need to be solid on the fundamentals, comfortable with Python and SQL, and able to explain your findings to someone who doesn't code.

Start with data science foundations, add machine learning once those are steady, and build a couple of projects you can actually talk through in an interview. If you'd like a timeline and a sense of the costs involved, our course pricing and options lay it out plainly.

AI is a tool inside data science, not a rival to it — so the best move is to learn the craft first and let the AI skills slot in where they belong. If that's the direction you want to go, take a look at the full data science and AI programme and map out your first three months.