Reflections from Snowflake World Tour 2026

A day of contrasts, and a lot of AI

Data Domain sponsored Snowflake World Tour 2026 at the New Zealand International Convention Centre in Auckland on 3 September. This year's theme was "Making AI Real for Business," and the day lived up to it: solid sessions, a busy expo floor, and no shortage of opinions on where AI is actually headed.

Catching Lourence's session at the Theatrette

One session worth calling out was Lourence Kruger's talk at the Partner Expo Theatrette, "Navigating the Map: Building Route Optimization and Spatial Layers in Snowflake." Lourence, our Snowflake Data Superhero, skipped the theory and went straight to live demonstrations, building spatial layers and calculating optimised routes directly inside the AI Data Cloud. It was a good reminder that a lot of the value sitting in location data goes untouched simply because teams don't realise Snowflake can do this natively.

The agentic enterprise needs a floor to stand on

The idea that came up again and again, across sessions and conversations, was the agentic enterprise: AI agents can only be as good as the data foundations they rely on. Trusted, governed and consistent data isn't a "nice to have" anymore. Without clear definitions and strong foundations, organisations risk spending more time and tokens for less useful results, no matter how capable the model is. It was good to see the conversation moving past "what can AI do?" and into "what do we need to get right underneath it for AI to actually deliver value?"

Maturity is uneven, and that's normal

Walking the floor, the range of AI maturity across businesses was hard to miss. The tools themselves are widely used in most organisations, but turning that into a coherent, company-wide strategy is another matter, even where presenters put their best picture forward. There's room in most organisations to step back and think ten times bigger, rather than using AI to automate one isolated step at a time.

That gap shows up inside companies too, not just between them. People with the same access to the same tools move at very different speeds: some keep pushing forward, others need to be dragged along and show little proactiveness or fresh thinking. Training only closes that gap so far. What seemed to help more was encouragement to experiment, not another how-to page.

The views in the room ranged just as widely, from genuine excitement about what the next model release will bring, to a quieter worry that AI will replace people altogether. Both reactions are understandable, and both are worth listening to.

The migration mountain is now a molehill

One of the clearest practical takeaways for organisations still sitting on legacy platforms was Snowflake AIM. For teams weighing up a move from Teradata, Oracle, Redshift or similar onto Snowflake, AIM turns what used to be a multi-year, high-risk rewrite into a guided, largely automated process, with code conversion, testing, data migration and validation all running under one shared migration state. We've covered how AIM works in more depth in a separate post, but the short version is this: the mountain most data leaders picture when they think "migration" is a lot smaller than it used to be.

The speed of that change is real. A proof of concept that used to take weeks to stand up, test and, if it was wrong, tear down again now takes hours. When the feedback loop is that tight, failing fast and cheap stops being a slogan and becomes the actual way people work.

What hasn't changed is where good AI enablement starts. It still begins with the data fundamentals: a sound model, clear ownership, the right governance. AIM will convert and validate code faster than any team could manage by hand, but it has no opinion on whether what you're moving is worth keeping in that shape. That judgement call still belongs to the architects and data leaders in the room, not the tool.

Worth the day

Between Lourence's session, the conversations on the floor, and a live look at what AIM can do, Snowflake World Tour Auckland was a good marker of where AI adoption sits in New Zealand right now: further along than the slides suggest, and for anyone still planning a migration onto Snowflake, closer than it's ever been.

If any of this sounds familiar, whether it's the data foundations under your AI agents or a legacy migration you've been circling for a while, we're happy to compare notes. Get in touch, or read more about our Data Foundation practice.