Latest / GrocerTalk / Fruitist’s Jim Trahanas on Applying AI Across the Entire Value Chain | Groceryshop 2026

Transcript

Fruitist’s Jim Trahanas on Applying AI Across the Entire Value Chain | Groceryshop 2026 · GrocerTalk

• 0:00
  Hello, everyone.
• 0:00
  This is omnitalk Retail.
• 0:02
  I am Ben Miller, host of grocertalk, and I'm coming to you live from Las Vegas from the Vuzion podcast booth.
• 0:09
  That's number 913 at grocery shop.
• 0:12
  And joining me now is Jim Trehanas.
• 0:15
  Jim is Chief Technology Officer, a futurist.
• 0:18
  Welcome to grocertalk, Jim.
• 0:19
  Thank you for having me, Ben.
• 0:20
  Okay, let's do a really quick intro.
• 0:22
  I'm sure that most of our listeners will know your blueberries, but they might not know the scale of the company and they certainly won't know what you do.
• 0:30
  So give us an intro.
• 0:31
  Great.
• 0:31
  So, Jim Trehanas, I'm the cto.
• 0:33
  As you mentioned, Frutus is a global provider of premium snacking in the form of berries and cherries.
• 0:40
  We grow our berries in seven countries.
• 0:42
  Globally, we distribute about 12,500 stores in the U.S.
• 0:47
  even bigger footprint if you include Europe, Middle east and Asia.
• 0:50
  12,500 stores.
• 0:53
  You're going to find us on the shelf 52 weeks a year in our retailers.
• 0:56
  We have distribution going nonstop.
• 0:59
  We have growers across the world, as I mentioned, which allows to cover it.
• 1:04
  And we see ourselves as changing the game and what snacking looks like for folks at home.
• 1:10
  So premium, healthy and nutritious snacking.
• 1:13
  Okay, that's really interesting, this whole changing the game thing, this is the thing we're going to dive into, because what many of our listeners might not know is that Fast Company has just named you as one of the
• 1:23
  world's most innovative companies, which, that's quite something and not necessarily what you would associate with primary production and food production.
• 1:31
  So I'm pretty sure as CTO you've had a big hand in this.
• 1:35
  So go on, tell me, what have you been doing?
• 1:37
  How have you been so successful?
• 1:39
  Yeah, yeah, good question.
• 1:40
  I mean, a question I often get alongside that is why is a.
• 1:43
  Why is a food, a CPG company have a CTO in this space in berry production?
• 1:49
  So everything for us starts with a forecast, right?
• 1:53
  Our ability to see what's going to happen in the future and predict our operations, et cetera, is a very important pillar in our execution.
• 2:02
  So we use data and advanced analytics and AI models to be able to predict outcomes in the future in the form of volumes and quality, so that we ensure that every berry you get at the grocery shelf, store
• 2:14
  or the grocery store on the shelf is of the same high caliber and high quality.
• 2:18
  Okay, let's start breaking that down.
• 2:22
  So tell me, what does that involve?
• 2:24
  So you say we Use AI.
• 2:25
  Okay, great.
• 2:26
  I think everybody has been through the booth so far this week said we use AI.
• 2:30
  So what do you actually do?
• 2:31
  What sort of AI are we talking ML?
• 2:33
  Talking LLMs?
• 2:34
  Like what are you doing and how does it help you to get that quality premium product onto the shelf?
• 2:40
  Yeah, yeah, so.
• 2:41
  So it starts at the bush level.
• 2:42
  It starts at data collection at the bush level.
• 2:44
  Okay.
• 2:45
  So we have massive farms, 2,500 hectares in Peru, and we have to be able to predict, as I said, to do that you need the data underpinning it.
• 2:52
  You'd have to have foresight and observations.
• 2:53
  In the field, we deploy computer vision models, we deploy advanced analytics, machine learning models for forecasting, using all the information and weather information that we collect from third parties to be able to understand how quality evolves with all
• 3:08
  of that information over time.
• 3:10
  So we have deployed remote sensing, meaning satellite imagery.
• 3:14
  We have computer vision in the field, we have optical sorting in the pack house.
• 3:18
  We have a variety of different types of ML models that are being used, from informing pick and harvest to just understanding how commercial allocation should go.
• 3:27
  It's absolutely.
• 3:28
  I mean, I'm trying to figure out.
• 3:30
  You say it like it's really easy.
• 3:33
  We apply computer vision to a 25,000 hectare farm in Peru.
• 3:37
  Like how, how a really specialized team that are interested in solving this problem, doctorates in environmental science, doctorates in computer vision, folks that are quite talented alongside best in class partners.
• 3:50
  Right.
• 3:50
  So it's not like we're building cameras from scratch on our own.
• 3:53
  We're finding the latest and greatest in industry and pulling that in to operate alongside us and really creating our technology on those two pillars.
• 4:02
  Jim, as well as the computer vision.
• 4:04
  You also mentioned satellite imagery.
• 4:06
  How are you using satellite imagery if you just.
• 4:09
  Yeah, remote.
• 4:10
  We call remote sensing with satellite imagery.
• 4:12
  So there's really two use cases for it.
• 4:14
  We use it for our own field intelligence and just kind of mapping what we call NDVI to monitor crop health and performance.
• 4:21
  But we're also looking at a broader application of satellite imagery alongside other data to be able to get a beat on where the market's going to understand movement for the country.
• 4:32
  Where are volumes going to be moving in the next 12 months.
• 4:36
  We do this by analyzing a lot of data coming from satellites, but looking at polygons across all of Peru and then being able to predict using other information when those volumes might be coming at what level.
• 4:49
  So we're working on this, we're piloting this year and we're seeing some really good results so far.
• 4:53
  So you're mapping an entire country's crop to help understand where the flow of product is going to come?
• 4:59
  Yes, yes.
• 4:59
  I got a dynamic duo team of this guy by the name of Brandon and Roberto.
• 5:03
  Fantastic gentleman, very bright, and they have built quite a capability for us.
• 5:08
  Fascinating.
• 5:09
  So talk to me about the building blocks you need.
• 5:10
  So you talked about the cameras, but obviously the data is a big part of that.
• 5:14
  How do you organize that?
• 5:15
  How do you build the building blocks that we need?
• 5:17
  Yeah.
• 5:18
  So to our founders credit and our CEOs credit, we're sitting on about 10 years, a decade's worth of information already that's been previously collected.
• 5:26
  There was a part of the DNA, I think, of the founding team that said we have to meticulously track everything that's happening at this company because we need to learn from it year over year.
• 5:34
  So that is really the gold pot.
• 5:37
  Right.
• 5:38
  That's the input into all of these models.
• 5:40
  Without that data sitting there and existing and being ready for us to analyze and run analytics against, the AI is a commodity.
• 5:48
  The data is really what's the moat for us, the special input.
• 5:52
  This sounds like a lot of investment.
• 5:54
  So how are you able as an organization?
• 5:57
  Where's the return come from?
• 6:01
  There isn't a huge margin in many, many fresh categories.
• 6:04
  So how do you justify what sounds like an incredible and complex organization?
• 6:09
  Yeah.
• 6:10
  So I mean, there's three pillars that were three levers that we're pulling.
• 6:13
  Right.
• 6:13
  It's an increase in yield, it's an increase in quality.
• 6:16
  Right.
• 6:16
  And ultimately the best possible relationship that we can have with our retailers because we're providing them the forecast that they need to be reliable for the consumers and that trickles into pricing.
• 6:27
  Okay.
• 6:27
  Because the forecast is at the very heart of everything.
• 6:29
  It's at the heart of everything.
• 6:31
  Okay.
• 6:31
  And so when you get these forecasts, is it.
• 6:33
  Are you directing the farmers to say what to pick, when to pick, how to pick, how, how does that connection run?
• 6:42
  Decisions start with our models.
• 6:44
  So our models are the starting point for making operational decisions in the field.
• 6:49
  So what we do is we say we're collecting all the information.
• 6:52
  Right.
• 6:53
  We're using our satellite imagery, we're using our computer vision of the field operating autonomous drones that are looking at Barry level information on our plants.
• 7:01
  We run all that through our models.
• 7:02
  Those models determine what's the optimal moment of pick for this particular field.
• 7:08
  And that's where we deploy labor and pick.
• 7:10
  Incredible.
• 7:11
  So it starts there.
• 7:12
  Do you want to talk about your customers?
• 7:14
  So that is the retailers that you work with, you talked about the sheer number of stores that you're in, but all the work you've done, all the investment that you've made to create this premium product, how does that
• 7:24
  relate to what you can do on shelf?
• 7:26
  What the merchandising and the value that you can bring to your customers and their shoppers?
• 7:31
  Yeah, look, in this category, I think customers, customers being the end consumer, the consumer really wants a consistent product.
• 7:38
  Right.
• 7:38
  Like they want consistency and quality and they're after shelf life and acidity and sweetness and all these other quality characteristics that we measure and we forecast out using all our technology.
• 7:47
  Right.
• 7:48
  So if we're able to provide that level of service to our retailers in which they can feel confident putting our product on their shelves, they're going to, it's going to be gone as fast as possible.
• 7:57
  They're going to keep building a relationship with us.
• 7:59
  Right.
• 8:00
  So we're going to keep investing in the technology that's going to be able to build that level of trust and that, that flow, that pipeline of highest quality caliber product that you could put on your shelves.
• 8:11
  One of the big themes of the show so far this week has been around the growth of E Com, but particularly around same day delivery, rapid delivery kind of items.
• 8:24
  I'm guessing that kind of is a pro for you in a fresh category.
• 8:28
  Tell me how, what have you seen impacting your category on this?
• 8:32
  It is, I think we've all had the experience at this point.
• 8:35
  The majority of consumers have had the experience of ordering something online and it could be fresh produce and you're basically having to succumb to whoever's fulfilling that order via that platform, whatever their tastes and qualities or convenience is.
• 8:48
  And when they pick their produce or their products off of the shelves.
• 8:52
  Having a stable supply of high quality premium products for those executors to go pull from means that the roulette game of having a variety of types of tomatoes that might come to you or berries or fresh produce, that
• 9:06
  kind of gets erased.
• 9:07
  Right.
• 9:08
  Because you're always going to have a consistency no matter what they pick from the shelf.
• 9:11
  That's our products.
• 9:12
  By the time it gets to you, it's going to have the same quality regardless if you picked it from the store shelf or if somebody else did it for you.
• 9:18
  Yeah.
• 9:19
  Okay, so that's right then.
• 9:20
  So you need to drive the brand awareness to know that you get the consistency of that product.
• 9:23
  Correct.
• 9:24
  And how do you drive innovation from an end consumer perspective?
• 9:28
  Because I can understand all the input you put into the Back end.
• 9:33
  But a blueberries.
• 9:33
  A blueberry or isn't it?
• 9:35
  Go on.
• 9:36
  It is and it isn't.
• 9:37
  Right.
• 9:38
  It's not a commodity.
• 9:38
  We're not producing commodities.
• 9:39
  Okay, perfect.
• 9:40
  We're producing what we like to call healthy, nutritious snacks.
• 9:43
  And we're doing so in various formats as well.
• 9:45
  Right.
• 9:45
  So different consumers have different needs.
• 9:47
  We have types of products that we sell in different arenas and different markets.
• 9:51
  But we're really looking to cater to the experience that the consumer wants.
• 9:54
  Right.
• 9:54
  Some people want snack cups to be able to quickly grab a single portion or a couple portions on the go.
• 9:59
  They don't want large containers sitting in their fridge for long periods of time.
• 10:02
  So we're catering to those needs.
• 10:03
  And our marketing team and sales team are doing a wonderful job of working with our retailers and our operations teams to figure out what works best for consumers.
• 10:11
  So there's a question I've got to ask you.
• 10:13
  You mentioned your founders, but how does a kind of a very affordable through a primary production company become and maintain to be so innovative?
• 10:22
  Yeah.
• 10:23
  Give specialists really interesting problems to solve and they will come and do it for you.
• 10:29
  Those folks that we have built, we built a team of about 100 years of agriculture technology experience in the first 90 days of my tenure with the company in this seat.
• 10:39
  They have propelled us forward in such a way that is incredible.
• 10:42
  So over the last year, we moved from sitting on a foundation of 10 years of data to having advanced analytics.
• 10:48
  All these models that we mentioned really being part of the heartbeat of the company and driving decision making every week, every day.
• 10:57
  So it's really built on them.
• 10:58
  It's really built on them.
• 10:59
  Fantastic.
• 10:59
  Okay, Jim, we've had so many conversations and people often come by the booth and say it's great all this AI thing.
• 11:07
  Can you tell me how it's really working?
• 11:08
  So thank you for coming by and sharing an example of.
• 11:12
  Of rich data and a rich example of how it's actually happening.
• 11:15
  So really appreciate it.
• 11:16
  I'm never going to look at a blueberry in the same way again, so thank you.
• 11:19
  Hopefully you can taste the difference, I'm sure.
• 11:21
  Thanks to the team at Fusion and Grocery Shop for making this coverage possible.
• 11:25
  And stay tuned for more.