Every marketing leader I talk to is under pressure to "do something with AI." Most are stuck in the same place: a few enthusiastic power users, a series of pilots and a nagging sense that the organization isn't actually changing. We felt it too. So a year ago we made a decision: Snowflake’s global marketing team would become Customer Zero for AI, the first and most demanding users of the platform Snowflake sells to its customers.

We didn't want a center of excellence or another tool rollout; we wanted AI woven into how a 600-person global marketing organization works every day. Here's what we learned and the playbook other marketing leaders can borrow.

Movement not mandate

The instinct is to mandate adoption from the top, but we went the other way. We made adoption peer-led, with a simple premise: marketers teaching marketers how to build real applications and workflows that solve everyday challenges.

The difference showed up in the data. After one of our North America events, weekly active users of Snowflake’s internal coding agent jumped 15% and stayed there four weeks later. That's the part most adoption programs miss: the spike is easy but the staying power is everything. Adoption stuck because it was social, not imposed.

According to an internal survey of our marketing team this year, 93% now use AI tools daily and most use them weekly as a core part of their workflow. We're not at our stated 100% goal yet but we're close enough that AI is now simply part of how the team operates on a daily basis. 

Build on your own platform, don't just query it

The biggest shift was building with AI instead of prompting it. When you put capable tools in marketers' hands, they stop asking AI questions and start shipping things.

The team has built production tools on our own platform: an inbound-email engine that triages thousands of replies a day; an account-based marketing app that delivers personalized intelligence to every account executive; a competitive-analysis workflow that turns a week of deck-building into minutes. One designer rebuilt a creative production pipeline that took a full week per batch and now runs in about 20 minutes. The line between "marketer" and "builder" is disappearing and that's the change that compounds.

Build a community of AI champions

That kind of sustained adoption doesn’t happen on its own. It’s driven by two groups we call the Snowflake “AI Accelerator.”

The Steering Committee focuses on “big bets,” secures resources and holds the organization accountable for impact. The larger AI Peer Committee leads education and evangelism, identifies use cases, promotes adoption and runs our “AI Day” events.

Giving enthusiastic peers low-pressure space to brainstorm, build and share has improved participation and fueled meaningful innovation.

Scale globally from day one

It's tempting to perfect this in your headquarters market and export it later. Don't. Our most striking results came from outside North America. In EMEA, an AI-powered localization workflow helped content in one market generate more than 27x the leads year over year from a small set of assets by activating the right content in near-real time to catch a demand wave.

In APJ, the team built tools to translate Japanese, Korean and Chinese company names and fix account matching. Treat AI as a global capability from the start and the wins show up everywhere.

Measure it and be honest about the hard parts

Here's the part most AI success stories leave out: speed creates its own problems. As AI raises the floor on output, some of the team spend more time reviewing AI-generated work, not less. A few have been candid that "speed without quality is not a win." They're right.

That tension is the real frontier. The next phase is governance, review standards and knowing when to trust AI output. And this isn't just Snowflake’s experience. Our 5th annual Modern Marketing Data Stack report finds the whole industry moving the same way: toward AI that acts, with governance as the thing that makes it trustworthy.

The takeaway for marketing leaders

Three things matter more than the tool you pick:

1. People first: Make AI adoption a peer-led movement, not a directive. Momentum sticks; mandates don't.
2. Technology your team can build with, not just query: The value is in what people create, not the responses they prompt.
3. Process and governance from day one: Celebrate the speed, but invest early in quality and trust.

We became Customer Zero because the best way to lead on AI is to live it first. A year in, AI hasn't replaced a single marketer on the team. It's raised the ceiling on what every one of them can do.

To learn more about Snowflake’s AI Peer Committee, please visit our blog.