For Australia’s logistics operators questioning if AI can ever deliver the goods, Steve Anderton, Brennan’s Head of Digital, explains to MHD Supply Chain how the road to success doesn’t need a moonshot.

Artificial intelligence (AI) continues to dominate strategy discussions across the logistics and supply chain sector. Yet despite the excitement, and an ever-expanding set of both theoretical and actual use cases, many organisations are still grappling with how to extract business value from AI.

That’s the challenge Brennan, Australia’s largest independently owned and operated technology systems integrator, has been tackling head-on through a national series of Data & AI Breakfasts.

With a track record of helping businesses align technology with strategic outcomes, Brennan designed these forums to move the AI conversation beyond hype and into practical action.

Hosted by Steve Anderton, Brennan’s Head of Digital Solutions, the events brought together business and technology leaders from across the country.

The goal was simple: To connect and share with peers, learn where they are on their AI journey, gain insight on the AI landscape from across the industry, understand some of the opportunities and challenges, and take away ideas, concepts, approaches and strategies to feed into their organisation’s AI journey.

Over the course of these sessions, a few themes came into focus – especially for logistics businesses seeking to move from proof-of-concept to production.

The gap between promise and production

One of the recurring questions at the forums was: If AI offers so much potential, why are so few projects making it into production?

Data from Australian research firm, ADAPT, was included in the breakfast presentations, offerings part of the answer. Despite high interest in AI – and strong intent to invest – only around 5 per cent of AI proof-of-concept initiatives actually reach production.

“There is no end of use cases in an organisation that could benefit from the application of AI. So why do only around 5 per cent of proof of concepts end up in production?,” the Brennan presentation posed. The answer? “Ultimately you have to align to business outcomes and value, then prove it.”

Half of the picture

“Right now, Generative AI – think Microsoft Copilot and ChatGPT – is gobbling up all the attention,” Steve says. “It’s been positioned as a major productivity enabler, and we’ve seen some compelling use cases, particularly in the legal industry, where the use of Copilot to summarise contracts, run conflict of interest checks, and draft documents can deliver real efficiency gains.”

Steve also said he noted a challenge. “Many CFOs told us they’re struggling to justify investments in Generative AI, simply because the ROI isn’t clear, and productivity gains are notoriously hard to measure. Unless you’re directly reducing headcount (which few organisations want to do), it’s tough to turn ‘giving people time back’ into a concrete business case.”

For the logistics industry, where margins are tight and efficiency is critical, that disconnect can be a roadblock. In contrast, AI embedded in core operational processes, like route planning, maintenance scheduling, and load optimisation can present a clearer picture.

The biggest barrier

As Steve explained, across the country, one message was repeated: AI is only as good as the data that powers it.

“Fragmented, siloed data remains a major barrier to scaling AI initiatives,” he says. “Without secure, scalable, and well-governed data architecture even the most promising AI projects are unlikely to deliver real value. We spoke to one customer who’s doing advanced work with AI, but even they admitted they’re struggling to scale because the underlying data infrastructure just isn’t where it needs to be.”

The presentations also echoed this question: “To truly unlock the value of AI, you need to hand it the keys to your most valuable asset: your data. So, the real question is, is your data ready?”

Brennan encourages organisations to “treat data as a product to unlock enterprise value,” defining a data product as “a valuable, domain-specific dataset that is self-contained, discoverable and easily consumed by teams to solve business problems or drive informed decisions”.

Another key takeaway from the events is the importance of micro innovation as a pragmatic, iterative approach to AI.

“When it comes to AI, you don’t need to bet the farm.”

“One of the most popular approaches we discussed was the concept of micro innovation – starting small, rapidly ideating, prototyping, and testing AI use cases to prove value quickly, without overcommitting resources upfront,” explains Steve.

But what does a typical micro innovation journey look like? For Brennan, it includes:

Rapid Ideation: Use cases, value identification and thinking.

Rapid Validation: Validation and prioritisation.

MVP and path to production: Build, refine, design, validate.

According to Brennan, this allows businesses to “rapidly prototype, test, and validate to determine proof of value. This approach ensures that you do not over invest in concepts before they are proven and can accelerate value realisation of the real business opportunities you identify”.

Stalling progress

The risks and concerns around AI were also a large part of the discussions.

“There’s also concern about the risks associated with Generative AI. From privacy and compliance issues to the phenomenon known as AI ‘hallucinations’ (when models confidently make up incorrect answers), the potential downsides are real – and top of mind for many of the leaders we spoke to,” Steve says.

“One customer, for example, had temporarily turned off access to generative AI tools across their organisation due to fears around brand damage and regulatory risk. They’d seen employees unknowingly inputting sensitive information into public AI tools, raising serious data governance and security concerns.”

While those concerns are seemingly valid, Brennan’s message is clear. “While it’s important to be aware of these risks, it’s equally important not to let fear stop you from moving forward. AI can deliver huge benefits – if you focus on the right use cases and put the right guardrails in place,” Steve added.

AI won’t take your job

The human side of AI adoption was another recurring theme during the presentation.

“There’s no question that AI will change jobs – but it won’t necessarily eliminate them,” Steve says. “Instead, we’re likely to see many roles evolve, with AI taking over repetitive, manual tasks and freeing people up to focus on higher-value, more strategic work.

“The bigger risk, in our view, isn’t that AI will take your job. It’s that someone who knows how to use AI might.”

“That’s why building AI literacy across your organisation is so important. By empowering your people to understand and use AI tools in their everyday work, you can ensure that they’re part of the AI revolution, not left behind by it.”

To embed change at scale, culture must be part of the equation.

As Dave Stevens, Brennan’s founder and Managing Director, put it, “Technology alone can’t deliver business outcomes. Success also depends on your people – and your culture. One of the biggest barriers to innovation is human adoption, even the best-designed tool or process will fail if employees don’t use it.”

Brennan’s forums showed that while AI is a complex field, the path to success doesn’t require a moonshot. Instead, it requires the right use case, good data, a clear outcome, and a willingness to start small and learn fast. For supply chain and logistics leaders, that might mean using AI to improve last-mile delivery accuracy, predict spare parts demand, or reduce paperwork through natural language processing.

Whatever the use case, Brennan’s message is consistent: cut through the noise, focus on value, and take a micro innovative approach to getting started.

This feature first appeared in the June 2025 issue of MHD Supply Chain.