AI Deployment at Retail Speed - with Larissa Schneider of Unframe

**Daniel Faggella** (0:12)
Welcome everyone to the Emerge AI in Business Podcast. Today's guest is Larissa Schneider, co-founder and COO at Unframe AI. Unframe builds secure AI native software that integrates directly with the company's existing systems, letting enterprises deploy custom AI solutions on their own data quickly. Larissa discusses how modular reusable AI components can compress integration timelines from months to weeks, and what enterprise data should prioritize in their first 90 days if they're serious about deploying AI at scale. Today's episode is sponsored by Unframe AI. According to Edison Research, 79% of Americans aged 12 and older listen to online audio monthly. That's an estimated 228 million people. For busy executive audiences, podcasts offer a rare opportunity to capture 20-plus minutes of attention with VP-level leaders in America's largest enterprises. Emerge reaches millions of those listeners every year. Learn about how recipes drive pipeline for other AI brands. Download our media kit at go.emerge.com/partner. That's go.emerge.com/p-a-r-t-m-e-r.
Now the conversation with Larissa.
Larissa, welcome to the show.

**Larissa Schneider** (1:52)
Thanks for having me today.

**Daniel Faggella** (1:53)
Yeah, I'm excited about our conversation. A lot of real and candid things that we need to talk about today that I think loads of leaders don't really want to talk about out loud. But something that we've been treading on in the back end in preparation for this conversation and in our general research is that in this space, it feels like everyone's talking about, oh, we've got a pilot running. And they're talking about it as if that's the finish line. But the reality sounds like that's actually just the starting point, right? Nobody really explains why so many of those pilots just stall. Not because the model is bad or most of the time, it's not because the model is bad, but because something else gets in the way once real production, the stakes start to show up. So I want to know, where do you see that friction hit hardest? What is the moment a promising pilot starts losing steam on its way to production?

**Larissa Schneider** (2:41)
That's a great question and actually something that we see in the market with our customers every single day. We work with a lot of large enterprises and I would say like, we don't have a technology problem. Like you said, the models are not the problem. There's so much tech out there. There's also a lot of noise in the market. Like we probably know all of the AI buzzwords at this stage. And so naturally, everyone's employees and all departments and all levels of the organizations, they have seen the shift on how ChatGPT like a couple of years ago started changing our personal lives. And now they're like, okay, I'm coming back to the office and there, I copy stuff from one system into another. I download a file, attach it to an email to then have a colleague remind me in two weeks to upload something. You know those types of legacy large enterprise workflows. And that's the opposite of something being AI native. So what do we see? These people are like, hey, I've seen there is an AI tool for this. I have seen a vibe coding platform and I can build something myself. So you kind of have these little islands of a lot of projects, a lot of trials, a lot of POCs running in all parts of the organization that naturally happened, which is great. We love people that are eager to try new technology, to innovate and change their work schedule. But how does that all fit together? That's the question we're asking. Because in order to have reliable, governable, secure AI that is really moving the needle for the business, everything needs to play in tandem. Everything needs to work together. And so that is often where we start and say, hey, what is your AI strategy? How are you seeing it? Do you have a core AI program where everything works together? You can see that it is giving the quality outputs that you are after and not ending up building and buying and vibe coding and have a lot of different separate channels that then make it really, really hard to see the output quality, but also the ROI that you're hoping to get from AI that you trust and that you put into full deployment and full production across all parts of your business.

**Daniel Faggella** (4:50)
That makes sense. I'm thinking that I've written about the technology is most of the time not the problem. We're throwing around the words like it could be data, it could be people. Who is the real culprit here? Is it mostly data being the problem or is it more a people thing or something else entirely?