Published on the 26/06/2025 | Written by Heather Wright

Plugging in a literacy program

AGL Energy is powering up for a data and AI future by ‘plugging in’ its people with an ambitious data and AI literacy program which is already reaping returns.

Shaun Code, AGL chief data and analytics officer, says the 187-year-old energy provider says data and AI is going to be ‘absolutely pivotal’ for the company to achieve its ‘bold’ plans – balancing energy affordability, reliability and security, meeting customer requirements while while decarbonising and helping customers decarbonise – and still operating effectively and efficiently.

“We scared people a little with some of the things that have gone dramatically wrong in applying AI,”

Coal-fired power stations power around 80 percent of Australia’s current power supply but that’s changing, in what Code calls ‘a really complicated way’.

“When we have a few major generators, they’re reasonably easy to run. But those few generators are being replaced by hundreds of wind turbines with multiple grid scale solar systems, with millions of solar systems on people’s houses and potentially millions of batteries within those houses.”

Add in vehicles – and the potential for them to cooperate with the grid, and it’s a ‘crazy challenge to orchestrate that market’. Maintaining smaller assets spread much more widely will also bring big scheduling and data problems.

Transforming workforce skills

But if data and AI will be absolutely pivotal’, workforce skills transformation will be absolutely critical to surviving and thriving in the data and AI future.

AGL’s solution has been a data and AI literacy program, which has been underway for around six months.

Senior leaders were the first target for the program, with a series of workshops to get them involved. Code admits providing the skills needed was just one part of the aim of the workshops – they were also aimed at getting senior leadership endorsement and sponsorship that would make driving the initiative throughout the wider business that much easier.

Around 50 senior leaders were taken through a series of three progressive workshops.

“Just booking them into the workshops was pretty hard, let me tell you,” Code says wryly.

The first workshop provide data and AI fundamentals, including clarifying what exactly AI was – and wasn’t – along with the basics of responsible AI and what it meant for AGL.

Feedback would quickly show that the leaders wanted more action from the workshops, and the second and third workshops were rejigged to be more interactive.

Workshop two looked at possibilities, ideating, and prioritising the big bets and no regret opportunities. Experts were brought in to talk about what other industries and companies were doing to provide an idea of ‘the art of the possible’ and get leaders looking at what they should be doing. There was also a focus on value management, and understanding how to assess, realise and monitor value.

Workshop three is the one Code is most fired up about. It scored a 100 percent positive feedback rating, and looked at how to take the ideas people had come up with in the previous workshop through to reality. Participants travelled through a series of four ‘stations’ building out their ideas in a workbook which would ultimately provide the basis for a business case.

Code is hoping FY26 will see ‘a lot of work for us’ as leaders take their workshopped projects into FY26 planning – the workshops were timed to allow projects to be included in FY26 plans.

The first of the three stations was a value station, which discussed forming a business case and the value drivers that could be unlocked.

“We know about efficiency and productivity, but there are other factors you may be unlocking,” Code says. Risk reduction was one big opportunity for the business.

A data and AI capability station saw leaders getting hands on with rapid prototyping with Microsoft Copilot, with experts on hand to help them create a working model.

“It was probably a little bit flaky, but it showed how easy it is to use AI to create AI,” Code says. “If you haven’t used things like Microsoft Copilot to create AI agents, give it a crack!” he adds.

A responsible AI station followed.

“We scared people a little with some of the things that have gone dramatically wrong in applying AI,” Code notes. Think Robodebt and its ilk.

“At the same time they’re filling in their workbooks, thinking about what responsible AI means in their particular initiatives.”

The final ‘workforce’ station saw leaders getting more theoretical with discussions around the impact on employees of AI joining the workforce and how to ensure adoption and engagement.

Those workbooks were a key part of the project’s success, Code believes. While they may sound ‘a bit basic’ Code says they’ve been ‘quite powerful’ and demand has run high for the template.

The senior leadership program helped raise excitement across the business, with FOMO kicking in for many others keen to get involved.

“Underpinning it, you also need people who do the real work trained, because the leaders can talk all sorts of things and come up with all sorts of ideas, but people have to make them real.”

So underpinning the leaders program was an online training platform focused on 150 business leaders, and providing self-paced online training through Udacity. The preconfigured content was curated and adjusted to suit AGL, the different ‘personas’ within the business, and the tools available within AGL so teams completing practical exercises on the platform could go and use the same tools in their real-world work.

“One of the things we were really careful about doing was understanding training needs for different personas across the organisation by talking to their leaders, the individuals and then tailoring learning journeys based on that,” Code says.

Preset learning journeys were established, based on a user’s role. The reward for completion? Access to even more training – in this case enabling access to any course on the platform for one year.

Measuring success

Code admits measuring success can be hard. He says AGL still hasn’t nailed it.

What it did do, was establish at the very start of the program where its people were and where they wanted them to be, and established five key success measures:

• Having a common language across the organisation for data and AI
• Empowering leaders to be confident to lead around data and AI
• Increasing use of data and AI tooling within AGL. “We are literally counting how many people are logging in and active in our data platforms and can see that is increasing.”
• Increasing awareness of responsible data and AI and data governance. Code admits this is a hard one to measure, but for a company with 4.5 million customer services – and all the PII data associated with that – it’s a critical component.
• Uplifting AGL’s data and AI engineering capabilities. “We didn’t want to forget the techies. We wanted to make sure the people building the platforms are not being left out and their skills are continuing to grow.”

Getting results

Survey results post the leadership workshops shows a big jump in confidence among participants. Prior to the training they rated themselves just 4.9 out of 10 – ‘that’s a fail’, Code says bluntly. By the end, they rated 7.5 out of 10 – in old school report terminology, that would be a B+, he says.

And when it came to how ready they were to progress their data and AI use case – a tangible outcome for the training – leaders rated themselves eight out of 10.

But it’s a story from the finance team that most excites Code.

He recounts how one finance leader – ‘a go-getter who loves tech – was so enthused after the workshops that on the train on the way home he logged in to the system and created his first AI agent. The agent helps answer questions around finance policies. The finance team member admitted it was still a bit shaky and needed some help. Code set him up with other more experienced people to learn more on training models and optimising searching and retrieval.

The person has since created a finance working group for AI (while it’s not a tech-lead group, there is, Code notes, a tech person on the team ‘to keep them in line a bit’).

A few weeks ago, the finance leader ran a session for finance people and anyone else who wanted to join, showcasing some of the things the team was working on, their ideas and to talk about how AI tools could be used. Around 100 people attended.

“I can hardly get a finance person to come to my meeting, let alone 100 turning up,” Code notes. “Clearly this has created its own subculture, starting their own working groups.

“It makes me proud.”