DGX Monthly - Three gold nuggets from ADG 2026

DGX Monthly – July

Hi everyone,

Chad Barendse here, co-founder of DGX, and welcome to the August newsletter.

ADG 2026 is done. Sold out again, 210 people through the door, 29 speakers, our biggest event to date and one of the largest data governance gatherings in the country.

Early on we asked for a show of hands from anyone who’d been with us last year. A few went up. The rest of the room was new, it was great to be able to share the ADG with the Melbourne DG community. The Rydges gave us a proper space for it, I still can’t believe the size of the screen!

Thank you for turning up. Genuinely. An event like this only works because the community shows up for it.

Tim opened the day by telling everyone to go hunt for a couple of gold nuggets and take them home.

Here are the three I left with.

→ Build capability, not policy.

Phil from Canva ran my favourite session of the day. His argument was simple. Our job isn’t to write the policy. It’s to build the tools and the capability that let teams do the right thing without ever having to read one. I’ve written before that most governance programs produce artifacts nobody touches. Phil’s answer is better than anything I’ve come up with. Stop producing the artifact. Start producing the thing the team actually picks up and uses. A policy asks people to remember. A tool means they don’t have to.

→ AI governance failures are data governance failures.

Every AI problem we heard about on the day traced back to something upstream. Bad data. No lineage. No ownership. The AI just made the problem visible faster. Subhrajit Banik put it best on the panel: AI didn’t create these problems, it exposed the cracks in quality, classification and access that everyone already knew were there. Laura O’Meara’s version was that metadata, lineage and semantic layers were always the foundation, and AI has simply shown us the cost of not having them. What’s changed is the tolerance. You used to survive a manual data quality check done quarterly. Put an agent on top of that same data and the error doesn’t sit there quietly, it propagates everywhere at machine speed. Same gap, no slack left in the system.

→ Governance has to be engineered, not written.

Policy documents don’t scale. Embedded controls do. A governance rule shouldn’t live in a PDF, it should be expressed as a machine-executable constraint, because that’s the only way it stops being aspirational and becomes enforced. Kelly Bayer Rosmarin’s line was the sharpest version of why: launch approval is point in time, but risk is not. You can’t approve an agent once and audit it a year later.

Which means governance at AI scale can’t be manual. You can’t have a steward reviewing every feature that goes into a model, or a committee approving every metadata entry in a catalogue that ingests automatically. It has to be automated, embedded and event-driven. That requires people who understand both the governance intent and the technical architecture well enough to build the systems that enforce it, and it requires governance teams to sit with data engineering as a co-designer of the control rather than as a client requesting a service. Most teams aren’t structured for that today. It’s a role that didn’t exist in most organisations five years ago.

I’m working through the sessions properly at the moment, so the full write-up of the day lands next week.

Thanks for reading

Meme of the Month

Credit: Chad Barendse

AI & Data Governance Summit 2026

The AI and Data Governance Summit 2026 has officially wrapped. 210+ attendees. 29 speakers. A sold-out room at Rydges Melbourne. ADG Summit 2026 is officially in the books, and it was our biggest one yet.

The takeaway from the day was clear: data governance is booming. There’s more executive attention on it than ever before. The opportunity now is to meet that attention with substance: governance that’s engineered, AI-enabled, and built to keep delivering value long after the summit ends.

Speakers

Thank you to our 29 speakers who gave their time and expertise across a full day of keynotes, panels, and fireside chats. A special thank you to Ole Olesen-Bagneux, Chief Evangelist at Actian, who flew in from Denmark to deliver our International Headline Keynote.

Sessions covered everything from AI security and governing agents, to human-centred governance, to practical case studies from teams standing up AI governance inside real organisations, practitioners across Canva, UniSuper, Transurban, KPMG Australia, AustralianSuper, GrainCorp, Hewlett Packard Enterprise, Bupa, Seek, Standard Chartered, Vanguard, Accenture, RACV, and more.

Sponsors

None of this happens without the organisations backing it:

→ Actian — Headline Sponsor
→ Experian and Xinference — Gold Sponsors
→ ISACA Melbourne Chapter — Industry Partner

Thank you to each of them for their support of Australia’s largest independent Data & AI Governance conference.

Award Winners

Congratulations to our 2026 ADG Award winners:

🏆 AI & Data Governance Leader of the Year — Phillip Ward, Canva
🏆 Rising Star in Data Governance — Brad Hill, Aware Super

Both incredibly well deserved, and a strong reminder of the calibre of people doing this work across Australia right now. Thank you also to all of our finalists for their leadership and contribution to the community.

The Community

And most of all, thank you to the Australian data governance community who showed up, engaged, and made the day what it was. These events only work because you keep coming back.

What’s Next

The ADG Summit is just one part of how we’re building Australia’s leading community for AI and Data Governance. Keep an eye out for upcoming webinars, Data Governance Connect meetups, and executive roundtables — plus planning is already underway for ADG Summit 2027.

Full session highlights coming soon.