Transcript
A (0:00)
Foreign.
B (0:05)
Welcome to reshaping workflows with Dell Pro Precision and Nvidia, where innovation meets real world impact in high performance computing.
A (0:19)
Logan, live from GTC 2026. And I'm actually sitting down for an interview, which is fantastic because I have not sat down for one in a while. So I'm here with Craig Chris, also known as Jerry from Whale. So Chris, let's get started. Tell us a little bit about you, your role at Whale and what Whale does.
C (0:37)
Sure. Happy to be here. My name's Chris. I'm the VP of Sales country credit for Whale. We've been around for about eight years. This is year number one for the US expansion. What we do is basically AI for store operations and revenue growth.
A (0:53)
Store. So store operations for revenue growth. Define what that means.
C (0:58)
Essentially we focus on how to ultimately increase revenue for any retail settings. Right? So think revenue leaders are interested in things like how many people walked in the store, how much time are people spending in certain location in the store? Are we losing money because people are leaving because they're spending too much time in line or for auditing purposes, Right. Like are people making burritos or coffee the way that they're supposed to be making? Are people wasting food? So those type of things we could do with our devices and what we call a VLM visual language model, vision language models on the back end that does all the, all the smart work, computing back end and essentially gives revenue leaders these type of numbers and analytics so they can make better decisions.
A (1:42)
Got a follow up question. So pretty familiar with, you know, VLMs, right? So let's say we take this little camera, we'll use a grocery store example, right? And let's say people are loitering too long in the beer section because you know, they want beer, whatever it is. Right. So tell me a little bit. How does Whale. Is Whale fine tuning this model? Like what additional context are you buying of something that's off the shelf that then. And then where does that go? Where do you kind of fit in? Like the, I'm not going to say revenue pipeline, right? But where do you fit in kind of that, that context chain of the information that you're collecting, where does it fit in? Like the ERP or whatever systems that they're using within a grocery store?
C (2:18)
Yeah, that's a great question. So in the example that you've given, right, like different revenue leaders have different KPIs and different things that they really care about. So one of the customers that we work with is LVMH and Obviously the things that they sell are more, more high end. They're selling, you know, 20, $30,000 products. Right. So what they may essentially care about are things like are my product displayed correctly or you know, are there people that I'm that are not intended to. Right, the people, potential buyers, Right. So a big vertical is also for us is automotive. Right. So if I walk into a car dealership and I'm walking around the car for very long, right. Why, why am I not being attended to? I'm a potential buyer that, you know, could lifetime customer, lifetime customer value could be about half a million dollars. Right. So these are the things that revenue leaders that manage essentially 2,000 stores, they really care about. So how many people or what's the time that people are not attended to across 2,000 stores? In the case that you mentioned, safety and security, loss prevention is also a big thing, Right. Like we can do dwell time analysis. So if somebody is spending a lot of time around an area where typically if you're picking up beer should probably be no more than 30 seconds, but we could create an alert where if somebody's at a section for more than a certain amount of time, we probably want to have a set of eyes on this person. And so it's really different for different environments. But the beauty about VLM is you can prompt it to do visual search and the models are already pretty, pretty good, right. CV's been around for more than a decade now. So things like, you know, I don't want to say facial recognition, but we could do people tracking, dual time analysis, demographic information. Those type of models are out of the box, ready to go. But it's also very easy to custom build some models with data available. Typically with any store that has some type of CCTV setup, it's pretty easy to configure that.