HOST
Australia
LOCATION
21
ATTENDEES
Finance & IT
INDUSTRY
AUSTRALIA, JULY to AUGUST 2026 — In partnership with The Ortus Club, GitLab hosted a two-city executive series across Australia to explore the future of agentic software delivery with the engineering leaders shaping how it will be adopted, governed, and scaled.
Titled around the launch of GitLab’s Duo Agent Platform and the newly announced GitLab Orbit, the series brought together senior tech leaders from Australia’s largest banks, telcos, retailers, and government agencies for a day of technical exploration, customer stories, and facilitated peer discussion.
Set against the changing landscape of AI-assisted development, the mornings united engineering leadership with the vendors, model providers, and peer practitioners best positioned to act on what comes next.
THE BRIEF
Taking the Agentic Roadmap to Australian Engineering Leaders
GitLab partnered with The Ortus Club to host an executive series in Sydney and Melbourne designed to bring AI engineers responsible for adopting and governing AI-assisted development into the same room.
GitLab’s Duo Agent Platform and the newly announced GitLab Orbit represent a significant bet on agentic software delivery, and the ANZ market needed a way to deliver four distinct modes of content in a single morning. A technical keynote with a partner appearance from Anthropic. A customer story. A peer panel across organisations at different stages of adoption. A facilitated roundtable that gave senior tech leaders in the room a chance to discuss their own experiences.
By pairing GitLab’s technical roadmap with the peer-level environment that The Ortus Club specialises in building, the mornings gave Australia’s AI & tech leaders a rare space to share their real questions on governance, adoption, and the future of AI in the software delivery lifecycle.
THE SETTING
A Morning at Sofitel Darling Harbour
The event in Sydney took place at Sofitel Darling Harbour, a waterfront venue chosen for its combination of technical facilities and executive setting. The atmosphere was purposeful and refined without being distant. Guests arrived to natural light streaming across the harbour, a refreshing breakfast spread, and the quiet buzz as their peers networked.
The half-day format ran under five hours, structured so the technical content, peer discussion, and the closing session had ample time.
Interested to find out how the event was shaped and the work that went behind it?
THE GUESTLIST
Platform Owners From Australia’s Largest Institutions
The evening brought together a distinguished group of platform and engineering leaders, including Chief Technology Officers, Field CTOs, VPs of Engineering, Staff Solutions Architects, and Associate Directors of Developer Platform teams.
They represented top organisations such as:
- Westpac
- Macquarie Group
- Optus
- Qantas
- Commonwealth Bank
- Woolworths
- ANZ
- Ticketec Entertainment Group
- Cochlear
- Pronto Software
The diversity of expertise, from financial services to telecommunications to retail to healthcare, sparked cross-sector discussions on how governance frameworks, adoption strategies, and productivity measurement differ across the ANZ enterprise landscape.
THE SPEAKERS
Vendor, Partner, Customer, and Peer
The speaking roster with GitLab was structured so that no single voice carried the morning. Craig Nielsen, VP of APJ at GitLab, and Andrew Haschka, Field CTO, opened with the keynote on intelligent orchestration and the Duo Agent Platform, including the first look at GitLab Orbit. Rob Williams, Staff Solutions Architect at GitLab, handled the live demonstration.
Natalie Mead, Head of Applied AI for APAC at Anthropic, joined the keynote to discuss the partnership rebuilding source control for the agentic era. Bringing the model provider into the room rather than referencing them on a slide changed the register of the questions that followed.
Chris Dickinson, Chief Technology Officer at Pronto Software, delivered the customer spotlight. The panel on agentic AI across the SDLC brought together Tane Oakes, CTO at Ticketek Entertainment Group, Sam Jarman, Engineering Manager at Cochlear, and Santhosh Thiyagarajan, Associate Director of Developer Platform at Optus — three organisations at genuinely different stages of adoption.
The morning closed with Nadine Champion, martial arts expert and the speaker behind the TEDx Sydney talk on ten seconds of courage. Booking a non-technical closer for a technical audience is a risk. It worked because the subject matter was change and resilience, which is what everyone in the room was actually managing.
THE TOPIC
Intelligent Orchestration, Now with Context
The framing question was blunt: AI coding is easy, governing it at scale is the hard part. Writing code with a model is a solved problem for most of the organisations represented. Knowing what that code touched, who approved it, whether it met policy, and how to prove any of it to a regulator is not.
The roundtable sessions with GitLab ran under the heading of speed with control in leading agentic engineering at scale. Sub-topics explored across the morning:
- Governance and auditability of AI-generated code in regulated environments
- Where agentic AI delivers measurable value across the software delivery lifecycle
- Connecting AI across planning, development and deployment rather than at a single stage
- Security, privacy and regulatory constraints on assistant adoption
- Measuring engineering productivity and return on AI tooling investment
THE INSIGHTS
Governance is the Bottleneck, Not Capability
Ahead of the roundtable, The Ortus Club circulated a set of engagement questions designed to surface what guests actually wanted to discuss with GitLab. The responses shaped the conversation in a specific direction.
Governance outranked every other concern by a factor of three. Guests named governance, oversight and compliance as the single biggest challenge in scaling AI-assisted development. Some named security and code quality assurance. Nobody chose talent, cost or technical capability. The models are considered good enough; the control plane around them is not.
The room split on how much governance is enough. Seven favoured risk-based governance that varies by application. Four wanted enhanced oversight with dedicated AI controls. Three argued for governance identical to that applied to human-written code. This was the clearest disagreement in the data, and it is not a small one: it is the difference between treating agents as tools and treating them as staff.
Security concerns block adoption more than ROI doubt does. Nine of fifteen guests cited security, privacy or regulatory concerns as the factor currently limiting broader adoption of AI coding assistants. Three cited difficulty measuring business value. Yet when guests were asked what single question they would put to the room, ROI came back repeatedly — one asked simply where people are actually seeing business value.
The room was optimistic, but nobody expected a revolution. Eleven of fifteen guests expected agentic AI to significantly improve productivity over the next 12 to 24 months while human oversight remains essential. Two expected incremental change only, two said it was too early to tell.
KEY TAKEAWAYS
What the Room Agreed On, and What It Did Not
These are the key learnings from the events with GitLab:
Capability is no longer the constraint. Not a single guest identified model performance as the barrier to scaling AI-assisted development. The conversation has moved entirely to control, evidence and accountability.
Risk-based governance is the emerging default. Roughly half the room favoured varying oversight by application criticality rather than applying one standard across the estate, though a meaningful minority still wants parity with human-written code.
Regulated industries set the pace. With banks, insurers, health providers and government agencies in the room, security and regulatory constraints were not edge cases. They were the primary gate on adoption for the majority.
ROI measurement is an unsolved problem, and people know it. Guests asked each other where value is actually being realised. The absence of a confident answer was itself the finding.
Human oversight is assumed, not debated. The dominant forecast was significant productivity improvement with humans still in the loop. Autonomy without approval was raised as a question, never as an expectation.
THE RESULTS
A Room That Kept Talking Afterwards
For GitLab, the commercial value sat in the composition rather than the count. Four Westpac attendees across three functions, plus platform and engineering leadership from Macquarie Group, Optus, Qantas, Commonwealth Bank, Woolworths and ANZ, is an accurate map of the ANZ enterprise engineering market. The follow-up pipeline generated from the room was tracked against named account executives, with meetings set within two weeks of the event.
The Sydney edition also produced the benchmark for the tour. The survey data captured here gave GitLab a read on Australian sentiment that shaped how the Melbourne event was positioned two weeks later — and, as it turned out, revealed a meaningful difference between the two cities.
The morning closed the way any well-designed event usually does: with guests still talking as they gathered their things, the follow-up conversations already being booked and scheduled, and the sense that the questions raised in that room would keep relationships alive long after the venue doors closed.