How Enterprise Leaders Are Building Their Agentic AI Strategy - with Matt Renner of Google Cloud

**Daniel Faggella** (0:12)
This is Daniel Faggella, you're tuned in to the AI in Business Podcast. We've had on countless unicorns on the show. Lord knows if it's hundreds of companies worth a billion dollars or more. There's only so many companies worth over a trillion dollars, and Google Cloud is one of them. This week, we have Matt Renner, who's the Chief Revenue Officer for Google Cloud, one of the largest, most powerful companies in the entire world, clearly one of the most important players in AI infrastructure. And we speak directly with Matt Renner in this episode around what infrastructure means now compared to what it meant two years ago. When enterprise leaders are saying we're leveling up our infrastructure, what the heck does that mean? Google works with all the biggest companies in different industries. He makes that definition and then gives us a sense of what it looks like to make infrastructure nimble. If we're going to have a world where we have more digital employees than we have physical employees by hiring digital workers, spinning up all kinds of workflows that are vastly more agentic, in order to spark that up across the enterprise, what's the transformation that has to happen? There's patterns that go well and there's patterns that fail. Matt Renner talks about both and he proactively talks a lot about security. You'll notice in this interview, I didn't ask him about security, but he had a bunch to say and I think it's a bit of a portent of some of the considerations we're going to see in the agentic world coming up. So this was a lot of fun to be able to chat with Matt here. This episode, it was originally aired on the AI Infrastructure Podcast. Most of the episodes on that show are not aired in AI in Business, but because AI Infrastructure Podcast is so new, we're featuring some of our heavy-hitting episodes from that show on AI in Business, so that you, our AI in Business listeners, can know about the AI in Infrastructure Podcast. If you go on Apple or Spotify or anywhere you listen to podcasts, you type in AI in Infrastructure Podcast or AI Infrastructure Podcast Emerge, that's EMERJ, you'll find the show. But also you can follow this convenient link to go directly to Apple. We're going to get into Matt's episode. Let me just give you the link here.
emerj.com/inf and then the number one, that's INF like infrastructure. So emerj.com/inf and the number one, that's going to take you straight to Apple for the AI in Infrastructure Podcast. We are proud to launch this new show. We have a tremendous number of heavy hitters from multi-billion dollar, in some cases, trillion dollar companies on the AI in Infrastructure Show. We started off with big guests. So check out that show if you're interested in the bigger AI foundations, the enterprise of the future. Without further ado, let's fly in. This is Matt Renner, Chief Revenue Officer of Google Cloud, here on the AI in Business Podcast.
So Matt, welcome to the show.

**Matt Renner** (3:00)
Thanks, Dan. Glad to be here, and appreciate the time.

**Daniel Faggella** (3:02)
Yeah, totally. We've got a topic that, as I was just mentioning with you, has never been more salient with our crowd. The idea of infrastructure is relevant to people vastly outside of IT, as we're kind of preparing for whatever the agentic enterprise looks like, and I think everybody's figuring that out in real time. I want to start with what you see as the hurdles in legacy enterprise around getting ready for leveraging AI at scale, having data silos broken down, feeling like we can leverage agentic enterprise in different parts of the enterprise, et cetera. Different people are at different maturity levels. But when you look at the hurdles that are most common for companies that have been around 50, 100 years or plus, what do you see?

**Matt Renner** (3:37)
Yeah, look, I think there's a lot of challenges that we've seen across the enterprise, without question. But some of the old things that probably were always lining up, even the last few iterations of technology come to play, and the big one's data that comes into play. When we got into this last three years, working with AI, AI just hit the world by storm, and companies took a series of approaches with varying degrees of success on how they were going to go after it. I would probably say there's all kinds of numbers out there as to what the failure rate were for the original POCs, the original science projects, and the things that went on.
But the majority did fail, and failure being defined as it didn't end up in a production system that was driving value.