Employment Hero has been technology-forward since day one; we’ve made it our business to digitise manual employment processes. When AI came along, we decided to embrace it. That meant giving employees the opportunity to use AI in their work, to see what it could do for our productivity and to drive new, exciting results for our customers.

It was a big undertaking, but the move to embrace AI can be replicated in every business, regardless of size. So we’ve built this guide in two halves. First, the honest story of how we did it and then a practical blueprint you can act on, with tips to use along the way.

Employment Hero was built on a simple belief that technology should make employment easier and more valuable for everyone. Staying close to what technology makes possible has never been optional for us. So when AI began to change how work gets done, our leadership moved quickly to work out how to adopt it well.

We knew we had to figure out how to use it to gain a significant advantage; to create the best product and offer the best experience to our customers possible. That is no small challenge for a business with more than 350,000 customers globally.

Employment Hero has stayed close to new technology throughout its history, so moving early on AI fit how the company has always worked.

Under CEO Ben Thompson’s leadership, this was made explicit. “We Are AI First” was added to our company values. That meant that embracing AI became a company-wide expectation directly from the CEO, instead of an HR initiative.

As Liam explains: It wasn’t just a comms exercise. It was actually about, how do we change people’s performance? How do we change the expectations that people work within? And it made it much harder for people to avoid, quite frankly.

From there, we began regular communication and training that encouraged the use of AI. The cadence was consistent and deliberate. Our internal comms channels were updated at least once a week with tips and ideas and there were regular company-wide presentations. Everything tied back to the same framing.

When we wrote our AI framework, we made a deliberate choice which was to start with what people were allowed to do. It was a clear signal that the team was encouraged to use AI, paired with firm parameters that keep everything safe. That difference in framing is what made it something people felt comfortable to pick up and follow.

For our team, the policy had to account for the sensitivity of the data we handle on behalf of over 350,000 businesses globally. Getting the balance right between permission to experiment and risk management wasn’t optional. It required HR, information security, AI enablement and others to work together very closely from the beginning.

The outcome was a policy that HR, legal and security were confident in from a risk management, compliance and enablement front. When HR presents AI policy as something that exists to protect the business from its own employees, people feel policed. However, when HR presents it as something that exists to give employees the confidence to learn and grow, people feel supported.

Now, people know what they can do, what requires a conversation first and what sits outside the boundaries. They have clear guidelines on which tools they can use for what and what kind of data is permitted to be used. That is exactly what makes confident adoption possible.

Part of making AI adoption real is being deliberate about which tools your business standardises and why.

For Employment Hero that meant making decisions based on our existing infrastructure and our data privacy requirements. Gemini, sitting within our existing Google Workspace environment, became the primary tool across the business for exactly those reasons.

For businesses thinking about their own tool choices, the most important consideration is that it fits with your existing environment. The best AI tool for your business is the one your team will use, that sits within the systems you already have and that your policy can govern.

Not sure where to start? Here are some popular AI tools you can look into.

The businesses that hesitate on AI policy longest are often the ones most worried about losing control. However, when AI is already being used in most businesses without oversight, introducing a policy provides security. You need to know that they’re using it with your knowledge and within a framework that protects them and the business.

The problem here was the ratio of time spent on admin versus expertise. Before AI, the work that required that expertise, like reading a market, assessing real fit and building a relationship with a candidate, was getting crowded out by everything else the role demanded.

The first tool the team brought in was the Employment Hero Recruitment Agent. It changed the top of the funnel by scoring candidates so recruiters could prioritise who to review first, without skipping anyone.

To be clear, we will always review every single application. We never disqualify one without having read it, says Kate. But it certainly enables us to prioritise the order in which we’re reviewing applications, because we’ve now got a scoring framework.

The change that followed ran deeper than screening. When a new role opens, the talent partner now sits down with the hiring manager to capture a thorough brief before anything gets posted.

It’s almost like a role intake process where the talent partner sits down virtually with the hiring manager and really gets a thorough brief of what the problem is we’re trying to solve for by hiring that person, says Kate.

The job description is the easy part. What’s harder is working out whether the person sitting across from you is going to thrive in an environment that looked different six months ago (and will likely look different again by the time they’ve finished onboarding). Experience and track record won’t always tell you how someone behaves when the ground moves.

The quality that’s moved highest up the list is learning agility.

It’s a quality worth screening for deliberately, because the bar is rising quickly. 30% of Australian employers already list AI skills as entry-level hiring criteria, so the candidates who arrive fluent are only going to get harder to compete for.

In practice, that shows up less in what someone has done and more in how they talk about it. Does the candidate describe solving problems by figuring things out or do they wait to be told? When something didn’t go to plan, what did they do next? Those questions tend to tell you more than a list of software on a CV.

Employment Hero went a step further and built something formal. Every applicant now completes a case study during the hiring process.

What makes it useful isn’t just the end result.

Not every business needs to build a formal case study, but the principle translates. Ask candidates how they currently use AI in their work, not to screen out people who say they don’t, but to get a feel for their relationship with new tools and whether they’re curious or just saying the right things.

Be upfront early about what the role involves day to day, because people who will do well in that environment tend to respond well to hearing about it and the ones who won’t are better off knowing before they start.

The candidates arriving fluent in AI are going to keep getting harder to compete for. The businesses that will be in the best position aren’t necessarily the ones that find those candidates first. They’re the ones that have already built a process that develops that fluency once someone’s through the door.

That starts with who you hire and what you signal to them about what the role actually involves from day one.

Our journey found one of the biggest misconceptions about AI in marketing is that it’s coming for the strategist’s job. In fact, our AI Paradox research found 45% of Australian workers feel AI is making them replaceable. However, our experience on the ground tells a different story.

Employment Hero’s Marketing Director, Tasman Page, says there is a lot of industry discussion around AI replacing marketing roles, from strategist to creative, but the Employment Hero experience is that AI is actually helping them deliver faster, more insightful work.

In other words, what disappears is the waiting, the bottlenecks between having an idea and getting it out the door. Once those bottlenecks disappear, the whole flow of work changes, starting with how creative gets tested.

The benefit isn’t just speed, but what the increase in speed does for the people on the team. Work that once had to go out the door to an agency can now be picked up by team members who always had some of the skills but never had the runway.

That experience isn’t unique to one team either. Far from replacing people, AI is giving them more room to grow.

Small marketing teams don’t have time to test every new AI tool that comes out. Most only need to know three things.

1. What to use first
2. What it replaces
3. Whether it’s worth the setup

We’ve mapped out what you could do with Claude in the order that makes sense to adopt them, starting with what gets you the most value for the least setup.

When we started embedding AI into how the team works, we wanted to hear how it landed. We asked people across the team to talk to us through their own experience.

Here’s the response to using AI and what’s changed most for them day to day.

Alicia Marr, Head of Global Content Marketing, shared that when Employment Hero made its AI-first announcement, she was super excited about it. For her, the conversation around AI in the workplace had always carried unnecessary stigma and the announcement was a chance to reframe what was possible, not just in how fast work gets done, but in what becomes possible at all.

The work is still yours. You’re just spending your time on judgement and direction instead of execution. Part of what makes that sustainable is being open about it and Alicia made sure to let people know when work was completed with Claude or AI.

In Alicia’s role, content approvals was one of the most time-consuming parts of her workflow, requiring careful manual attention at every step, explaining:

Neil, Head of Sales NZ, says that “it’s been a really pleasant surprise to find out just how useful AI has been.”

What’s changed isn’t a small efficiency gain. It’s meant that Neil and his team now have a different way of working entirely. Before, getting a new view of data they needed meant going through the data team and waiting on their queue. “We’d send a request off to the data team to say, can you show us some insights into this? And we’d have to wait weeks to get a new panel put up in a dashboard.” That whole cycle, the request, the wait, the back and forth, has effectively disappeared. “Now it’s literally a 15-minute job to build a new view of the data.”

That hugely changes what the role looks like day to day. Instead of waiting on answers, Neil is generating them in real time, which means decisions don’t have to wait either. “Being able to make decisions, being able to change and pivot really fast has just been the biggest impact.” The work moved from requesting insight to producing it directly and that’s a fundamentally different job than the one they started with.

For Daniel Brady, GTM Solutions Lead, AI has had the biggest impact on his work in project documentation and communication.

This time-saving has meant that he can focus more on moving the project forward, instead of the admin that comes with managing a project.

Behind the headlines about AI and jobs sits a very human question. If AI can do parts of my job, what happens to me? The worry behind this is real and in Australia, it runs hotter than in any other market we surveyed. 45% of workers are concerned that AI is making them more replaceable. It’s worth meeting that fear directly, because the data doesn’t bear it out.

Instead, what happens when AI is embedded thoughtfully across a business is something different. The tasks that get automated are almost always the ones people liked least. The repetitive, time-consuming, admin-heavy work that filled hours without adding much. The time that opens up goes back into the work people were hired to do in the first place. Businesses with AI at the core are actually growing entry-level headcount at more than double the rate of lower adopters.

Inside Employment Hero, we’ve seen this play out across the business. In our creative function, AI has changed the shape of what the team works on, removing the mechanical parts of content production and freeing up time for the judgement that moves the work forward. In sales, admin that required hours of manual work has been restructured around AI tools, with the team now focused on analysis and improvement instead of data handling.

Our research supports what we experienced internally. 75% of Australian employees say AI has improved their productivity, while 74% say it’s improved the quality of their work. The work nobody really wanted to do is disappearing and what’s left is the work people signed up for.

Our journey with AI is far from over, but the early lessons are worth sharing. With everything we’ve learnt so far, we’ve built a practical starting framework for Australian businesses ready to move from intention to action.

Work through four phases of a company-wide rollout. These phases are drawn directly from what we did when embracing AI use at Employment Hero.

Phase 1: Get leadership aligned

This doesn’t have to be a CEO. In a smaller business it is simply the owner, or whoever is in charge, but it has to be someone with the authority to set expectations. They also should have the visibility to model the behaviour they’re asking for. At Employment Hero, our CEO announced the new AI-first value. That signal alone changed the conditions people were operating in, well before a single training session had run.

Phase 2: Build the guidelines first

Clear guidelines around AI use give your people permission to move. Without it, you risk having two groups:

1. The employees already using AI tools at work without telling anyone
2. The employees too cautious to start because they’re not sure what’s allowed

Neither group is where you want them.

Your guidelines primarily need to answer three questions:

1. What can people use AI for without asking
2. What requires a conversation before proceeding
3. What is off limits

Write it in plain language and have your HR function own it, alongside whoever manages your data security.

Phase 3: Train in segments, not just company-wide

The instinct is to train everyone at once but in practice, a company-wide rollout without a clear order produces uneven results. If you start company-wide, start by clearly laying the foundations of AI knowledge, like we did with our AI bootcamp.

Then, move to segments of the business. Start with the roles where AI has the most immediate, visible impact. Through their work, you’ll be able to demonstrate a clear before and after within a few weeks. Use those early adopters as internal proof points. When a finance team member shares how a process that used to take three hours now takes twenty minutes, that’s more persuasive than any training deck.

Phase 4: Measure the results

As with any business initiative, you need metrics to know if it’s working. The metrics that matter are the ones that track behaviour like time saved on specific processes, employee confidence scores before and after training, as well as the number of workflows that have changed because of AI.

Not every phase will run smoothly and that’s fine. What’s important is that you keep the momentum going. Employment Hero’s early weeks in AI adoption had their own challenges and the businesses that move fastest through hurdles are the ones that name them early.

AI should raise the value of every person on your team, not reduce the need for them. It’s the thinking behind Employment Hero’s whole platform, which we built to take the repetitive work off your team’s plate and hand back the human parts of the job. That includes the judgement, the relationships and the thinking.

Everything in this guide came from doing it ourselves first. The policy framework, the training pathways and the mistakes. We’re not describing a methodology we developed for customers, but what it actually took to move 1,800 people from AI-curious to AI-confident.