Your company has decided AI is important. You need someone to make it happen. You write a job ad, post it on Seek, and wait. Six weeks later, you’ve either hired the wrong person or you’ve hired nobody because your expectations were unrealistic.

This is how most Australian companies approach their first AI hire. Here’s a better way.

Decide What You Actually Need

The most common mistake is conflating different AI roles. A data scientist, a machine learning engineer, a data engineer, and an AI product manager are very different people with very different skills.

Data scientist: Analyses data, builds models, generates insights. Best first hire if your primary need is understanding patterns in your data and making predictions.

ML engineer: Builds and deploys production AI systems. Best first hire if you already know what you want to build and need someone to build it reliably.

Data engineer: Builds data pipelines and infrastructure. Best first hire if your data is messy, fragmented, or inaccessible. Without clean data, your data scientist or ML engineer can’t do anything useful.

AI product manager: Translates business problems into AI solutions. Best first hire if you have multiple potential AI use cases and need someone to prioritise and scope them.

For most Australian mid-market companies making their first AI hire, a data scientist with some engineering skills is the right choice. They can analyse your data, identify opportunities, build prototypes, and help you make the case for further investment.

Write a Realistic Job Description

I’ve reviewed hundreds of Australian AI job listings. Most are wish lists rather than job descriptions. They ask for expertise in every framework, every technique, and every domain, plus five years of experience with technologies that are three years old.

Nobody matches those listings. The good candidates know it and apply anyway. The great candidates look at the unrealistic requirements and conclude the company doesn’t understand what it’s hiring for.

A realistic job description for a first AI hire should list: the specific business problems you want them to work on, the data environment they’ll be working in (be honest about its quality), the tools and infrastructure currently available, the team they’ll work with, and three to five genuinely essential skills.

Everything else goes in a “nice to have” section or doesn’t appear at all. You’re hiring a person, not a catalogue of capabilities.

Set Realistic Salary Expectations

The Australian AI salary market is competitive. For a mid-level data scientist or ML engineer in a capital city, expect to pay between $150,000 and $190,000 base salary. Senior roles command $200,000 to $260,000. If you’re outside a capital city, you might pay slightly less but you’ll also have a smaller candidate pool.

If your budget is significantly below market rates, be upfront about what you offer instead. Flexible working, interesting problems, equity (if you’re a startup), professional development budget, or a faster path to leadership than larger companies offer.

Don’t waste candidates’ time or your own by interviewing people you can’t afford. The AI talent market is small enough that companies with a reputation for lowballing struggle to attract candidates.

Interview for the Right Things

Technical interviews for AI roles at many Australian companies are broken. They either test theoretical knowledge that’s irrelevant to the actual job or they test implementation skills using whiteboard coding that doesn’t reflect real work.

A better approach for your first AI hire.

Take-home challenge using your real data. Give candidates a sample of your actual data (anonymised if necessary) and a business question. Ask them to explore the data, identify relevant patterns, and present findings. This tests practical skills, communication, and how they think about problems.

System design discussion. Describe a real business problem you’re facing and ask how they’d approach building an AI solution. You’re testing their ability to scope a problem, identify data requirements, choose appropriate techniques, and anticipate implementation challenges.

Communication assessment. Ask them to explain a technical concept to a non-technical audience. This is crucial for a first AI hire because they’ll spend as much time communicating with business stakeholders as they spend writing code.

Culture and collaboration. Your first AI hire will work closely with people who don’t understand AI. How do they handle questions they consider basic? How do they push back on unrealistic expectations? How do they explain why something is harder than it looks?

Set Them Up for Success

Hiring the right person is half the battle. Keeping them productive and engaged is the other half.

Give them data access immediately. The number of AI hires who spend their first month waiting for database access or security clearances is staggering. Have everything ready before they start.

Pair them with a business domain expert. Your new AI hire doesn’t understand your business yet. Pairing them with someone who deeply understands your operations accelerates their learning dramatically.

Define clear first-quarter objectives. Not “explore AI opportunities” but specific deliverables: “Analyse customer churn data and present findings. Build a prototype prediction model. Recommend next steps.” Clear objectives prevent drift and provide early wins that build organisational confidence.

Protect their time. Your first AI hire will be a magnet for every “quick question about AI” from across the organisation. This is flattering for about two weeks and then destructive. Set boundaries around their time to ensure they can do deep work.

Plan for growth. Good AI people leave when they stagnate. Have a discussion about career development in the first month. Where could this role grow? What additional hires might follow? What skills development will you support?

Your first AI hire is an investment in your company’s future capability. Get it right and you build a foundation for genuine AI adoption. Get it wrong and you lose twelve to eighteen months and a lot of money. Take the time to do it properly.