Where does AI fit into the future of payroll?

The recently held Australian Payroll Summit explores how employers can achieve AI actualisation, build frameworks for everyday AI use, and keep safe governance at the centre of operations.

In a session presented by Benchmark 365 chief executive James Vickery, titled “From AI curiosity to everyday practice”, a five-step framework was outlined for practical AI outcomes: security, standardisation, training, skill sharing and scaling.

Four years into mainstream professional generative AI use and two years into agentic AI, Vickery noted that organisations are often underutilising the technology.

He said: “I’ve yet to meet an organisation that I’ve seen that’s getting significant traction with AI. There’s still a dispute internally about which tools to use. There’s still people that are using their own tools instead of company tools.”

As AI appears in organisations without planning, Vickery said governance also becomes a significant issue, noting recent cyber security data branches related to employee prompts – the risks of which are compounded when sensitive information such as payroll data comes into consideration.

He said: “It’s important to check the permissions of all of the data that you have. So if someone has access to internal payroll information, you need to check that not everybody has access to that information. If you have access to it, then the AI has access to it, and then the AI ingests it.

“It’s important to set proper permissions … you can’t put the genie back in the bottle.”

To begin, Vickery urged employers to ensure companies have an AI use policy in place.

An acceptable policy “will have things like which tool do we use at the company, which tools we’re allowed to put information into, which tools we are not allowed to put information into. It will have things in there like use of personal information and which tools are appropriate, if at all, to put information in, and it will have some other things in there as well”, he said.

Reflecting on his experience cleaning up “technology sprawl” for businesses, Vickery said that AI retains the hallmarks of shadow IT, whereby workers would pick and choose technology as they went along and expect it to streamline functions.

When it comes to standardisation, Vickery noted the importance of understanding AI as software, and like any other software, picking one tool and investing in capability. Once a tool is picked, he said, organisations can build their workforce’s capability, starting with training.

Despite the fact that rolling out enterprise versions of AI software is a big investment for companies to make, Vickery said training is not often backing up the hefty price tag – with many simply using AI as a glorified search engine.

In addition, training is often both simple and the strongest signifier to productivity.

Vickery said: “It’s as simple as, how do I use this tool? What’s the best way to prompt the tool? What’s the best approach to upload a picture? What’s the best way to upload a file? How do I do deep research?

“A lot of these things are foreign to the average user.”

In comparison, workers equipped with base capabilities will often refine their skills, with some developing to levels such as agentic AI that can operate more autonomously or having a greater capacity to share their knowledge with others.

Explaining the value of skills sharing, Vickery said: “If you have a win using AI, you find an efficiency. Maybe it’s a way that you calculate something or deal with a complex issue. Wouldn’t it be great if everybody in the team had that?”

Vickery explained that the natural flow-on effect of skills sharing is scaling.

He said: “Once we get everybody on the same platform, once we get everybody using the same tool, once we get everybody trained on that platform, and once we start to share skills, custom prompts, even just prompt sharing in general, that’s when things move very rapidly.

“When we actually use agentic AI, and when we share these skills with one another, we go from an individual to an organisation.”

Want to see more stories from trusted news sources?
Make HR Leader a preferred news source on Google.