Watch Tom's chat with ART's Adam Fitzhenry
As Australia’s second largest superannuation fund, ART is no stranger to operating at scale. The fund manages $330 billion for its more than 2.4 million members.
Inside the business, its 3.2k employees would flock to Adam’s People team with question after question, but more often than not, the questions would be the same.
The People team needed a safe, reliable way to answer these questions at scale. That way, the business could get the information it needs fast, and Adam’s team would be free to focus on the strategic, higher level work only they can do.
Listen on Apple Podcasts, Spotify or YouTube, or read our recap below.
Before Josef, Adam says, “I could pull a list of all the questions we’d get, and we’d have around 500 different inquiries a day. 400 of those would be the same thing.”
“There were lots of repetitive questions coming through, so we looked at using AI and Josef in particular and asked: How could AI help here?”
Adam explains, “We had the content already available. We’d just direct people to that content or write back to them and tell them the answer.”
His team thought, “How can we use AI to take that extra step away, save our team time, and create a frictionless experience for our leaders?”
“Everyone wants things to be easy,” Adam says. “Nobody wants to go to work and be de-skilled. If I can search on Instagram or Facebook at home, why can’t I have that same experience when I come into a workplace?”
Led in partnership with Legal, the People team weighed up what solutions could help them and ultimately landed on Josef.
The team’s first set of AI Q&A tools, designed to answer questions about popular People policies, were launched to the business and the feedback rolled in.
“We’ve got 15 minutes being saved” per question, Adam says. “And that’s just through a number of tools that people are self-reporting.”
“We hadn’t anticipated significant savings like that,” he adds. “But the fact that people are sharing that with us is fantastic.”
“The ability to know that there's a level of governance around the answers, and that we could report on that with transparency, was really important to us at ART.”– Adam Fitzhenry, General Manager, People Services and Enablement, Australian Retirement Trust
Adam and the team had several criteria their AI tools needed to check to safely get them out the door.
1. Accurate answers they can trust
“The last thing I want is someone raising an issue from interfacing with the AI and giving them a completely wrong answer,” he says.
“We just can’t have them head off and give incorrect advice on an employment matter, for example.”
“The accuracy and quality of what the model’s producing is so important.”
2. Good governance and control
“We were looking for a tool that enabled us to apply levels of governance that you don’t get from something like ChatGPT or Copilot,” Adam says.
“The ability to know that there’s a level of governance around the answers, and that we could report on that with transparency, was really important to us at ART.”
3. Easy access
“The tools really needed to be integrated into our daily workflows on our intranet as that’s where the majority of people get their daily news and the latest resources.”
He adds, “By integrating seamlessly with what the team already do, we’ll be able to gradually shift behaviour over time.”
Since launching the first set of tools, Adam’s already looking to what’ll come next.
“It’s what underpins AI that’s going to help generate more value for us,” he says. “How you choose to use this new infrastructure to create value is up to you and the use cases you pick.”
One use case waiting in the wings? Q&A tools integrated with live business knowledge that guide managers on when’s the right time to hire.
Adam says, “The ability to interact with our knowledge in a way that results in complete business outcomes relatively quickly… it’s really exciting!”
That’s Adam’s advice for other teams looking to plug their existing documentation into new AI tools.
He reminds peers: “If you don’t input clean, good quality information and knowledge, you’re going to find that the output’s going to be just as bad as what you’ve put in the beginning.”
ART’s approach to AI hasn’t focused solely on speed (although a boost always helps!).
It’s been one that centers on:
By increasing access to knowledge and reducing the amount of questions People has to field, Adam and his team are helping everyone in the business work that bit smarter.
“I just want to get all the low-level stuff out of the way,” he says. That way, his team can focus on, as he says, “creating frictionless experiences for everyone at ART who are here everyday just trying to make sure our members have the best possible retirement.”
Explore how you can empower the business with self-service AI Q&A, contract and workflow tools on Josef.
We'll be in touch soon to arrange a time to speak.