The Australian AI builder scene is small enough that you keep running into the same names online. OpenClaw is one of the few places those names actually show up in person. We flew up to Sydney for the meet last week — first time attending, and worth the trip.
The format is informal. A venue, a few dozen people, no keynotes. Engineers, founders, agency owners — most of them building something with AI, most of them in the messy middle of figuring out what actually works at a commercial level. The conversations are better for it. Nobody is performing. Everyone is comparing notes on real problems.
The people in the room
A few of the conversations that stuck with us:
What the room was talking about
A few themes came up independently in enough conversations that they felt like real signals rather than a single person's hot take.
Evals are underrated and mostly absent. Most people in the room were building agents without systematic evaluation. Not because they don't care — because it's genuinely hard to build eval pipelines that test the things that matter in a business context. This is a real gap in the tooling.
The ops/IT split is getting awkward. As AI automation moves into actual business workflows — CRMs, ERPs, internal tooling — it increasingly touches infrastructure that IT owns. Nobody has a clean handoff story for this yet. Several conversations circled around the same tension: the business wants to move fast, IT wants to assess risk, and neither side has great vocabulary for what these systems actually are.
Voice is having a moment. More people building voice agents than we expected. The quality of the underlying models has crossed a threshold in the last six months where voice is genuinely deployable in commercial contexts — not just demo-grade. We're seeing the same thing in our own builds. Bodyline wouldn't have been possible eighteen months ago at the price point we hit.
Voice has crossed the quality threshold. Not just demo-grade anymore — commercially deployable, at SMB pricing.
Worth attending again
OpenClaw is informal and unstructured, which is either its best or worst feature depending on what you're looking for. If you want panels and structured content, it's not the right event. If you want to spend three hours talking to builders who are actually shipping things and comparing notes on what's working — it's excellent.
We'll be back for the next one. If you're building in this space and you're in Sydney or willing to travel, it's worth looking up.
If you were at the meet and we didn't connect — reach out. Always keen to compare notes with people who are doing this for real.