You already built your AI stack. It was just for analytics.

Ryan Green: Hello, everyone, and welcome back to Data Cloud Now. We're currently in San Francisco, where I'm now delighted to be joined by George Fraser, CEO and co-founder of Fivetran plus dbt Labs, along with Tristan Handy, president and co-founder of Fivetran plus dbt Labs. Gentlemen, such a pleasure to sit down with you here at Snowflake Summit.

George Fraser: Thanks for having us. Great to be with you again, Ryan.

Ryan Green: Looking forward to diving in. I have to say it's not often that you get the opportunity to sit down with two founders and CEOs in a world where AI remains the top priority for just about every leader. How are your platforms enabling the AI revolution and what unique challenges do you solve in the AI data pipeline together? George, why don't we start with you?

George Fraser: Well, if there's one message that I am delivering this week to every customer, it is that the same data stack that you've been building for the last 10 years to support analytics, the same data stack that Tristan and I and our respective companies have been helping you build is the right data stack for AI contexts. The same problems that you have to solve of curation and permissions, defining data domains, integrating all your data in one place, those are the same problems that you have to solve in the context of AI. The right way to connect your AI agents to your company's data is to connect them to the data platform that you already have and that you've been investing in for the last 10 years. Well said, George. Tristan, from your side,

Tristan Handy: I think that the conversation that I keep having is where we've been a company that's very focused on best practice for a long time. We've taught an entire industry of people how to do this work in a cloud-native era. the benefits to using best practices and to using industry standard tools has never been higher. When you map out your organization's data using Fivetran and dbt, by default, you get a map of how data flows through your organization and now that context can be automatically loaded into agents. Keeping things out of very bifurcated heterogeneous workflows like standardizing on a single best practice is more valuable than ever.

Ryan Green: It's wonderful to hear, gentlemen. And I want to build on this. How are 5G and DBT working very much together as one to unlock customer value across the board? And what examples come to mind? Tristan, let's go back to you.

Tristan Handy: Yeah. So like George mentioned, we have been working together for a very long time. I mean, I remember calls with George and Taylor 10 full years ago when they were working out of some WeWork or WeWork clone. In truth, before even coming together, we had 1,500 shared customers, and I think that that actually under-represents because there's 100,000 teams using dbt in the world, and I think it's actually somewhat unusual for a Fivetran customer to not use dbt in some way, shape, or form. We've been delivering value to customers for as long as a modern data stack has been a thing. Traditionally, that has always been moving data from left to right, ingesting it at the source and transforming it and making it ready for analytics of any type. But George, do you want to talk about any use cases?

George Fraser: There are so many customers that use us together to solve problems and examples that come to mind of customers who really have gone wall to wall with Fivetran and dbt. One example is Zendesk. They use it across the business to integrate all types of data. They've done everything from establishing a 360 view of their customer by integrating data from all the different tools that have data about their customers to helping their finance team close the books. by integrating financial data into their data warehouse. Another great example, OpenAI, who set up Fivetran and dbt when they were a little research lab and we just scaled with them. And boy, did we scale. So every kind of company you can imagine. I mean, we're talking about technology companies here, but we have lots of consumer packaged goods companies, manufacturers, you name it, use Fivetran and dbt together. It's really the powerhouse of how to do things with data when you want things to be simple.

Ryan Green: I absolutely love that, the powerhouse. Gentlemen, how are both engineers and business users benefiting from stuff like CoWork and stuff like CoCo and innovations to drive business outcomes? George, why don't we start with you?

George Fraser: Well, I'm a huge user of AI on top of data. I routinely use AI agents to query our own data warehouse. I've been telling a story this week from last week in an exec team meeting with me and Tristan and the rest of the combined exec team when we got together for the first time. We were talking about the adoption of data lakes at Fivetran, and I had it build a report that involved not only traditional analytical queries, querying counts of customers, percentage of revenue, things of that nature. But it also then went and read the transcripts of the gone calls. One of my favorite data sources, which is in our data warehouse, the AI agent went and read these transcripts for the top 20 customers and reconstructed the story of why they adopted this product capability. It's like having an infinite supply of analysts that work at hyper speed when you get it set up correctly, and it's easier than you think. to set up this workflow if you bet on the data platform that you already have.

Ryan Green: Wonderful to hear George and Tristan from your seat.

Tristan Handy: I love that Snowflake has native AI features because it is actually There's actually this weird situation where tokens are hard to come by. Now, that's a weird thing to say, but large language models use token counts in and out to generate their intelligence. And you can't just generally use your laptops, compute for that. And so you actually have to go on a search for like, where's the supercomputer that I'm going to use to get tokens? And sometimes it is actually hard to find a token provider that has access to your organization's data. And so the fact that Snowflake can bundle both of those up in one and give you access to both like the supercomputer in the cloud that you need to generate tokens at scale and have access to your data is actually like quite a hard problem that every customer here needs to solve.

Ryan Green: It's wonderful to hear, gentlemen. Now I want to look towards what's next as you both look ahead to say, hey, the next three to five years, how do you see the data landscape evolving and what role will your platforms play in that future? Tristan, back to you.

Tristan Handy: So I think in some ways the answer is we're going to continue to play the role that we've always played, which is to say sourcing data and making meaning out of that data. uh the the interesting part of this now though is that uh rather than humans being the primary consumers of this data i think we're looking at agents being the primary consumers of data and if you look around us right now basically every single vendor here has the word agent in in its copy somewhere um what i've been exploring this week is When people say agents, do they know what that means yet? What does that mean to them? And I think that really there is not a shared understanding of like what are the true use cases that most enterprises are going to have for agents? Everybody has this sense like it's going to be big and coding agents are already big. But we are over the next three to five years going to figure out what that really looks like and we're going to deploy it at scale. Well said. And George?

George Fraser: I think everything Tristan just said is exactly right. And the only thing I'll add is that there is a huge amount of uncertainty over the next few years. No one knows what data platforms exactly are going to look like over the next few years. And that's why one of the big ideas we've been advocating for is interoperability. Why you need to make sure that you're making investments that not only work today, but that make you future proof. Things like Iceberg, which Snowflake has heavily invested in, are the way that you make sure that the workflow that's going to be so important that hasn't even been invented yet is going to be possible on your data platform in a couple of years.

Ryan Green: Well, gentlemen, I'm looking forward to this next chapter. Congratulations again. Such a pleasure sitting down with both of you. Thanks a lot. Thank you. For the audience watching, I'm Ryan Green with Data Cloud Now. Thank you so much for tuning in. We'll see you next time.