Catalog & Cocktails: The Honest, No-BS Data Podcast
Oct 1, 2026 · 52 min · 10 segments
Agents will build data infrastructure at every layer and become its primary consumers. Tristan Handy, President and Co-founder of Fivetran / dbt Labs joins Juan and Tim where we unpack the big shift…
Tristan HandyGuest
Juan SequedaHostTim GasperHost
Okay, I have, I have kind of two high-level statements on this, and then each of them is their own topic that, you know, we will all collectively be figuring out for the next, you know, several years.
Um, the first one is that, uh, I think that agents are going to, uh, be responsible for building, uh, data infrastructure, uh, at, at basically every layer, um, increasingly.
And then separately, uh, agents are going to become the primary consumers of the data that the data stack produces.
Um, and so the, the first one is a statement about how data infrastructure evolves, um, and what capabilities it needs and what APIs it needs, and how we as data practitioners are going to change our jobs, um, the dominance of evals and how, how evals are gonna work, and, uh, how agents are going to kind of on a, on an ongoing basis monitor our, our infrastructure.
The second one is more associated with, um, you know, right now basically every company that I talk to either has built or is trying to build, um, an agentic data analyst.
Uh, it's a, it's a very obvious- Uh, application of, uh, agents to structured data.
Um, you know, the, the core problem being, uh, my data team never, never has enough time to answer my questions, so what if we had infinite data analysts? And, and, like, that's important and good, but it falls into the category of, like, let's take the old world process and bring it onto the internet, you know? Like, we, we were using paper, now we're using computers.
Uh, and, and so that's typically the first type of innovation, but I think that the...
It's not that humans will ask agents to do stuff, and agents will do it and report back.
It is that, um, agents will... long-running agents will have their own, uh, goals that they will have been tasked with, and they will, uh, in the, in the, uh, furtherance of those goals, they will, uh, use the data platform as a source of truth.
Or p- p- retail organization, plan out the next 10 locations that we're going to build out or, you know...
And whatever kind of action or ultimate end result you're trying to drive, as opposed to things that are much further up the, the action supply chain, right?
Okay, I have, I have kind of two high-level statements on this, and then each of them is their own topic that, you know, we will all collectively be figuring out for the next, you know, several years.
Um, the first one is that, uh, I think that agents are going to, uh, be responsible for building, uh, data infrastructure, uh, at, at basically every layer, um, increasingly.
And then separately, uh, agents are going to become the primary consumers of the data that the data stack produces.
Um, and so the, the first one is a statement about how data infrastructure evolves, um, and what capabilities it needs and what APIs it needs, and how we as data practitioners are going to change our jobs, um, the dominance of evals and how, how evals are gonna work, and, uh, how agents are going to kind of on a, on an ongoing basis monitor our, our infrastructure.
The second one is more associated with, um, you know, right now basically every company that I talk to either has built or is trying to build, um, an agentic data analyst.
Uh, it's a, it's a very obvious- Uh, application of, uh, agents to structured data.
Um, you know, the, the core problem being, uh, my data team never, never has enough time to answer my questions, so what if we had infinite data analysts? And, and, like, that's important and good, but it falls into the category of, like, let's take the old world process and bring it onto the internet, you know? Like, we, we were using paper, now we're using computers.
Uh, and, and so that's typically the first type of innovation, but I think that the...
It's not that humans will ask agents to do stuff, and agents will do it and report back.
It is that, um, agents will... long-running agents will have their own, uh, goals that they will have been tasked with, and they will, uh, in the, in the, uh, furtherance of those goals, they will, uh, use the data platform as a source of truth.
Or p- p- retail organization, plan out the next 10 locations that we're going to build out or, you know...
And whatever kind of action or ultimate end result you're trying to drive, as opposed to things that are much further up the, the action supply chain, right?
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
