Data Community Virtual Conference 2026 Debrief. What stuck with you?

[RhysHanscombe]

That’s a wrap on the Data Community Virtual Conference 2026! Thank you to everyone who joined us across the two days.

A few themes came up again and again in the sessions and the Q&A: where Data Vault fits as AI reshapes modelling workflows, how much (or how little) of the data architect role is going to change in the next few years, and the ongoing tension between self-service access and governed, consistent business definitions at scale.

We had people dialling in from over a dozen countries and every corner of the data world, and the questions you asked were some of the best we’ve had at a Data Community event.

We’d love to keep the conversation alive right here on the forum. So what did you think? And what are the key points you took away from the talks?

What was the one idea, quote, or session that’s still rattling around in your head a few days later?

Did anything change your mind, or confirm something you’d already suspected about where data modelling is heading?

If we did this again, what’s the one topic you’d want a deeper dive on next time?

If you couldn’t make a session live, ask away here, there’s a good chance someone in this thread was in the room (virtually) and can fill you in.

Thanks again for making this community what it is.

[wow0609]

One interesting point that was made I am still thinking about:

With AI becoming as prevalent as it is, the Prompts used are now closer to code or a script than just directions for the AI… and should be treated as such. This makes me think that prompts to the AI should be saved/managed just like a python script, stored in a repository, versioned, etc.

But this leads to another question: AI, and especially different AI versions, will interpret a prompt differently… so the outcome may be very different from one run to another… whereas a python script and expected output is static.

Very interesting and thought-provoking points.