The AI Co-Pilot Has Landed at the 2026 AfMA Summit

The 2026 Australasian Fleet Management Association (AfMA) Summit delivered no shortage of big ideas this year, but it was a keynote from Eden Shirley, Founder & Managing Director of FleetGuru, that left the room buzzing. Part live demo, part industry provocation, and part theatrical spectacle, Eden's presentation was a bold signal that the AI transformation of fleet management is no longer theoretical, it's already in the hands of clients.

Meet Mic: The AI Maintenance Co-Pilot

Eden took the stage and quickly made good on his promise to introduce the audience to "someone you might work with in the future." That someone was Mic, FleetGuru's AI maintenance co-pilot and he joined the keynote live, via microphone, in real time.

For anyone who has ever sat through a polished pre-recorded AI demo, this was something different. Eden was candid from the outset: "Today could go brilliantly well, or it could be very confusing." It went well, mostly and the moments of imperfection only made the technology feel more real.

Mic introduced himself as "the nerdy data guy on your fleet team," but the numbers behind him are anything but modest. Trained on fifteen years of AutoGuru and FleetGuru's vehicle servicing data, Mic has ingested 3.4 million online transactions, over $1.2 billion in services and repairs, 180,000 labour rate changes, and 2 billion part price changes from 12,800 approved fleet repairers. He can see approximately 490,000 unique vehicles on the platform and has analysed roughly 3.4 million quotes, of which about 69% were ultimately approved or completed as work orders.

Three Agents, One Stage

What Eden actually demonstrated wasn't a single AI tool, but three distinct agents working in concert. He unpacked these clearly in his closing remarks:

1. The Insights Agent

This agent is fully trained on the AutoGuru and FleetGuru databases and can surface pricing intelligence, track trends, and flag anomalies that would take a human analyst hours to find. During the demo, Eden, playing the role of a fleet manager for a fictitious company called FleetX, asked Mic to pull up everything flagged in his review queue.

Mic identified 12 items, leading with a brake pad part for a 2021 Iveco Daily, quoted at $191.40, compared with a peer benchmark of $162 across 860 similar jobs. He flagged a 2018 Hino 300 Series with a labour line running longer than the peer median. Across all 12 flagged items, the total exposure sat around $400 above benchmark, modest per booking, but meaningful at fleet scale over a year.

At Eden's instruction, Mic reached out to the service provider directly to clarify the pricing. No email drafted, no task created, just done.

2. The Booking Agent

Eden then asked Mic to surface all vehicles overdue for service. He returned a prioritised list, with a 2021 Toyota RAV4 sitting three weeks past due at the top. Eden's next instruction: "Pick the top three, reach out to the drivers, schedule those bookings in, and report back to me."

Mic got to work, and Eden pointed out to the audience that three actual calls were happening simultaneously in the background. It's the kind of task that consumes hours of a fleet controller's week, and here it was being handled in seconds.

3. The Multilingual Booking Agent

The third demonstration was, by Eden's own admission, the one he's most excited about. He invited a colleague, Kota Tsuda, GM of FleetGuru Japan, to the stage to simulate a mechanic calling in to provide a quote to a fleet management company in Japanese.

Mic handled the interaction seamlessly, switching into fluent Japanese to collect the vehicle details: a 2018 four-wheel drive Isuzu MU-X with 93,333 kilometres on the clock, needing a logbook service, fuel filter replacement, and wiper blades. He provided a total quote in Japanese, then confirmed the booking in FleetGuru.

The point Eden was making goes well beyond novelty. In Australia, the majority of repair order approvals are handled digitally. But in most countries around the world, phone-based approval is still the norm and most of those countries are non-English speaking. Mic can conduct those conversations in 77 languages and convert them into clean, structured data.

"This is what's happening with AI," Eden said. "The frontier is all about how structured and clean that data is."

The Chainsaw Analogy

Beyond the demo, Eden offered some grounded perspective on what AI actually means for the industry and for the people in it.

He shared a metaphor he'd heard recently: if you ran a logging company with ten workers using axes, and someone invented a chainsaw, would you let nine people go and hand one chainsaw to the last person standing? Or would you give all ten workers a chainsaw?

"AI is an amplifier," he said. "It has the ability to help us do and achieve so much more."

It's a point worth sitting with, particularly given that Eden acknowledged the sector being disrupted most aggressively right now is technology itself. Those building AI tools are on the front line of the transformation they're creating.

Real, Live, and in the Wild

One of the more striking moments of the keynote wasn't the demo itself, it was Eden's acknowledgement that this technology is already live with clients. He called out Summit Fleet Leasing and Management as an example of a company that has taken AI insights and built them into a production-ready product that clients are actively using, called AI Fleet Insights.

"Everything that you saw today was real," he said. "To see it in the wild and to hear that clients have got access to it, and that it's creating real insight and real value, that's just a fantastic thing."

What Fleet Managers Are Actually Worried About

Eden closed with a nod to a recent global survey of fleet managers, which found that the top two concerns are rising costs and increasing compliance complexity. His argument: AI is well-positioned to address both.

It's a compelling case. An agent that can benchmark repair quotes against millions of data points in real time, flag anomalies before approval, schedule overdue services autonomously, and handle multilingual repair authorisations at scale, that's not science fiction. That's what was on stage at the AfMA Summit this year.

The AI co-pilot has landed. The question for fleet managers now is how quickly they get in the seat.