DGX Monthly – July
Hi everyone,
Chad Barendse here, co-founder of DGX, and welcome to the July newsletter.
We’re now 49 days out from ADG 2026, and planning has stepped up another gear. We’ve now locked in 25 speakers! It’s shaping up to be one of the biggest events in the data governance space this year. I’ll keep sharing names as they confirm, so keep an eye out.
But the thing I’ve been thinking about most this month is the semantic layer, and why it might be the most strategic thing a governance team can own right now.
Most governance programs produce artifacts nobody touches. Glossaries that go stale the week they’re published. Frameworks that sit in a SharePoint folder while the data keeps flowing ungoverned underneath. I’ve watched three teams build the same metric three different ways the month after everyone nodded at the policy.
The semantic layer changes that. It’s the shared translation layer between raw data and the people and tools consuming it, where “revenue” or “active customer” gets defined once and means the same thing in every dashboard, every model, and every AI prompt. Without it, “active customer” is whatever each analyst decided: one counts a login in 90 days, another counts a paid invoice, and the board gets three different numbers. With it, the definition stops being an opinion and becomes infrastructure.
Here’s why it lands on the governance team’s desk. A semantic layer is only as good as the metadata underneath it: definitions, lineage, certified metric logic, ownership. That’s the exact work a governance program is meant to do anyway. We already know what happens when an asset like this is left to rot, because we lived through it with the data catalogue. Enterprises bought one, populated it over months, launched it with a slide deck, then quietly stopped maintaining it. A semantic layer nobody maintains follows the same arc, except the failure is louder this time, because stale definitions feed live dashboards and AI answers.
AI helps here. It can suggest definitions, flag drift, draft glossary entries. But point an LLM at ungoverned, contradictory inputs and it produces ungoverned, contradictory outputs, just faster. What keeps a semantic layer alive is someone accountable reviewing definitions and retiring the wrong ones. That’s not a model capability. That’s what a governance team is good at.
I’m not going to pretend I’ve got this fully solved, but here’s where my head’s at.
I want a named owner on every certified metric, not “the data team” as a collective. I’m looking at connecting the business glossary straight into the layer, so the agreed definitions are the upstream input rather than a document that quietly drifts out of sync. And I reckon the whole thing lives or dies on review cadence. A definition nobody revisits is just a liability waiting to surface in an AI answer.
Thanks for reading
Meme of the Month
Credit: Chad Barendse
AI & Data Governance Summit 2026
Australia’s largest independent conference focused on Data and AI Governance returns for its second year.
Held on 20 August 2026 at Rydges Melbourne, ADG brings together data leaders, AI professionals, and governance practitioners from across the country to confront the issues actually landing on their desks, not the theory.
This is one of the strongest speaker lineups we’ve ever put together — and we’re not done yet.
25 speakers confirmed. Practitioners from Canva, Aigentsphere, Actian, UniSuper, Transurban, KPMG Australia, AustralianSuper, Cambiq, GrainCorp, Hewlett Packard Enterprise, Angle Auto Finance, SmartyGrants, and DAMA.
Real people. Real organisations. Real problems.
These are the conversations the industry needs to have, and ADG 2026 is where they’re happening.
This year’s themes are the three things every data governance practitioner is wrestling with right now:
→ The Future of Data Governance — turning governance into an enabler with policy as code, automation, and catalogues people actually use
→ Governing AI in the Real World — practical frameworks for accountable AI, plus the harder edge nobody’s solved yet: securing agents and MCPs
→ Human-Centred Governance — building governance teams that actually adopt instead of working around
One stage. One day. This is where the people doing the actual work show what they’re actually doing. The wins, the failures, the stuff that never makes it into a polished case study.
👉 Secure your ticket
We look forward to seeing you in Melbourne!
ADG AI & Data Governance 2026 Awards
The ADG AI & Data Governance Awards are back for 2026, and nominations are open now.
For the second year running, we’re recognising the individuals, teams, and organisations making a real difference to the Australian data and AI governance community.
This isn’t an award decided behind closed doors. It’s community-driven. Nominated by peers. Judged by practitioners who understand what this work actually takes.
🏆 AI & Data Governance Leader of the Year
🏆 Organisation / Team of the Year
🏆 Rising Star in Data Governance
Whether it’s someone who’s quietly built a governance program from scratch, mentored the next generation of stewards, or pushed for accountability when it wasn’t easy — these are the people worth celebrating.
Self-nominations are welcome. If you’ve done the work, put your hand up.
Winners will be announced live at ADG Summit on 20 August in Melbourne.
👉 Nominate Now
Nominations close 31 July 2026.
