AI on the Farm: How Daisy Is Reimagining Dairy Operations

Reese Clark:

Hello everyone, and welcome to the Milk Theory Podcast. We're at the A DPI conference here in Chicago, and I have met a new friend. I'm super excited to have this conversation with him. I'd like to introduce you to Andrew Moore, chief Executive Officer of Daisy. Welcome to Milk Theory. We're super happy to have you.

Andrew Moore:

Thank you. It's been a pleasure.

Reese Clark:

So, honest question, one of the reasons why I'm super excited to have this dialogue with you is because building an AI company high level question, is AI going to kill us all?

Andrew Moore:

I'm thinking who's watching this?

Reese Clark:

What do you think?

Andrew Moore:

No, my answer is no. First of all, no. I think it's definitely going to change everything, but I think we're still going to be human at the end of the day.

Reese Clark:

Very good. Okay.

Andrew Moore:

Yeah. So I know I'm not like if you look at the movies or whatever like that, I don't see it getting to that stage, but I see there being some fun stuff that I can help with.

Reese Clark:

Good. Okay. Very good. So when did you start tapping into the world of AI? When it started to become mainstream or whatever?

Andrew Moore:

Yeah, so obviously I grew up in a very technological generation. I don't have, I'm not 60, I don't have gray hairs falling down, right. So I'm not the typical food processor I should say. But yeah, I got exposed with it really, really young. I just got naturally curious as to what AI actually is. I have my theories of what it's actually going to do, and even by just plugging in things to it and using it in everyday life, you really see the power of it.

Reese Clark:

Yeah, absolutely.

Andrew Moore:

Yeah. From a farming perspective, I think it's going to go quite interesting. I'm really excited for what it's going to do actually at a farm level. Yeah.

Reese Clark:

When you're talking about the ag world or dairy, I got to be honest, I've been here at this conference for a few days, and you're the first person I spoke with that actually mentioned AI and the future and how it's going to impact. Tell me a little bit more about, we'll get into Daisy, but just from a super high theoretical level, how do you think AI is going to impact the dairy industry specifically?

Andrew Moore:

Yeah. Okay. So there's severe volatility in dairy, right? Markets just crashed. Cows love consistency, but sometimes the market doesn't perform like that. Okay? People in these food processors are older than other industries are using very static Excel tools, and I think AI is really going to help with the volatility aspect. It's going to help with the decision making. It's not going to actually do the work, but it's going to help you think about what you should do. And that's why it's just the application of that in terms of where the starting point is. I think that's where a lot of the dairy processors actually can't grasp because we actually try and not say AI. We actually try and just be as simple as possible.

Reese Clark:

Yes. Yeah. Do you think one day the term AI might actually just kind be on the back burner and the rest of it's just technology, everything else that we integrate with?

Andrew Moore:

Exactly. I mean, it's going to be so integrated that we don't know the difference between it.

Reese Clark:

Yeah. So I mean, chief executive officer at your age, I mean, that's very impressive. So tell me a little bit about your journey and your motivation to start something like Daisy. We'll get into the knickknacks about what it's about, but where do you get your entrepreneurship motivation from?

Andrew Moore:

Yeah, so I grew up in a 600 cow dairy farm so milking a lot of cows at home. And as the eldest son with two younger sisters, it was either a farm for the rest of my life or get interested in something else. So that's why I naturally gravitated towards the next stage of the farm, which is what happens to our milk when it leaves. And that just, I had this hypothesis that food processing in general was just all automated, right? Yeah. You knew exactly what's coming in, you knew exactly what's going to be sold, what products do you need to make, is just a given. And then you start talking to food processors, you start lifting up the hood and looking underneath, and you see just Excel spreadsheets. You see people shouting at each other. You see all of these different things, and I'm like, wow, the difference we can make here is just insane.

Reese Clark:

Yes.

Andrew Moore:

That's actually how I came about it in the first place.

Reese Clark:

I could vouch for that. I mean, just being in the dairy processing field, Excel is like, I mean, everybody, they're swimming in different files and charts. It is a mess. I'm not a fan, even as a marketer, I can't get in it, but it is crazy. It's everywhere. Now, where does Daisy come into the picture to support some of these stressors that are already on our industry? How could it make our lives more streamlined, easier, and more efficient?

Andrew Moore:

So look, I think production planning is the big one. What do you make? When do you it? When shit hits the fan, what customers do you commit to? And I think that's really where we see the value because it's at that decision-making stage. There's so many different things you can do. And I think what we're trying to do is really, really simplify the supply coming in, simplify what your demand actually looks like, and just say, Hey, based on X, Y, and Z and what's happening in the plant, what's happening all around, here's what you should do. And yeah, we've just basically broken it down step by step. So it's simple. And it's integratable with the processor's existing processes?

Reese Clark:

Yes. Okay. So your platform, I saw a preview of it, and I would love to actually take some screenshots of it and display it over, but it's essentially a very user-friendly dashboard that you could load various forms of data within, and then you have an AI support agent or just chat that you could actually derive that, get that data more synthesized. I'm going to ask some backend questions and then I'll ask questions about the user experience. We know our biggest AI companies in the world right now. You think in an open ai, gemini cloud, are you guys running on something like that, or is it an amalgamation of a few of those?

Andrew Moore:

Yes. We actually build our own models similar to how OpenAI and Google and Gemini, I think the processors don't like their data being shared with these types of companies.

Reese Clark:

Yeah. Yes, exactly.

Andrew Moore:

So we need to be mindful of that. And we actually ringfence all of our data together so that each processor has their own backend,

Reese Clark:

Essentially. That is a huge concern, especially in this industry, is just data privacy and things like that. And when you say red fence, for somebody who's never heard of that term, what does that actually mean?

Andrew Moore:

Yes, so Ringfence

Reese Clark:

Ring fence?

Andrew Moore:

Yeah. So it's basically all the models are built for a specific processor. So every processor is unique, every processor, we basically siloed their data so that no data gets shared with any other external models, any external sources, and it's just completely independent and standalone.

Reese Clark:

Oh, interesting. Okay. Now you're getting these models. Tell me about where this data comes from. How do you make sure the models are good? I am going to botch the saying, but I've heard if you put bad data in a model, you'll get bad data on the other side. Yeah, exactly. And so how is that being vetted? Tell me a little bit about that process. I know you've got a team behind you. It's not just like AI continues to probably collect that data. There's some human critical thinking behind this. How does that get loaded into the backend of your system?

Andrew Moore:

So we have data scientists. We have machine learning engineers. So the options for processors, okay, you could either do this yourself internally, you could upgrade your spreadsheets, you could, I dunno, maybe figure out some other systems. You could try and poach engineers from Google, OpenAI, Microsoft, which would they be a cultural fit within food processors? Not sure. Or you could work with us where we're like a team of young, hungry people who are rooted in the farm itself, but also are coming from high levels of technical ability where we actually just intertwine ourselves within the process. So yeah, typically, to answer your question on the testing side, we run, we are taking data sources from over like 50,000 different variables. And we typically have a month long onboarding process that's like the first two weeks are building the models, and the second two weeks are actually testing. So we're back testing to make sure that our accuracies are actually in the extreme high percentiles, and that there's no anomalies within the data.

Reese Clark:

Okay. So is it accurate to say then this platform is not necessarily, I mean it's turnkey once they're in it, but when you are customizing each solution, tailoring it to the needs of the specific customer? Or is it more universalized?

Andrew Moore:

It is more universalized from what we do. There's two things that we need to consider from a security perspective that all the models need, the data can't be shared. And two, from a process perspective, some processors might do valorization before master scheduling, and there's different layers of command that they go through within the plant, which you're very familiar with.

So yeah, we don't want to go in and intrude and say that unless you do it our way, you can't produce anything. We have to be integratable with every processor's process. And yeah, that's a meeting internally that we have with the team to figure that out.

Reese Clark:

Okay. That's great. That's it. Okay. So yeah, you have that whole onboarding component. Okay. Okay. Now tell me a little bit about the front end. I was talking to one of your colleagues. Tell me the importance of a beautifully aesthetically pleasing, but also easy to navigate kind of user UX experience. There's a lot of AI tools. I've noticed that people are trying to piece different components together to get a project done based on a certain tailored AI tools. You're bringing it all together. What's the importance of a good user experience when it comes to just navigating a dashboard, like what you guys provide?

Andrew Moore:

Yeah, so I'd say one simplicity, right? Yeah. I mean, the problem today is that there's excels everywhere. There's so much data hidden within the ERPs. No one knows what's going on. Any meeting that has to happen, everyone has to do this preparation work, right?

Reese Clark:

Yes.

Andrew Moore:

And that just takes all of your time and you can't focus on what you're meant to do, which is produce high value products. So what we do is we a bring all the data together, and how we present it is, yeah, it's very visual, it's very aesthetic, but we only present a couple of key different things, right? And even from a forecasting perspective, we don't just supply the tool that you play around with the forecast. We literally take in all of your data and we give you the forecast.

Reese Clark:

Oh, okay. So it's based off of the data of the actual customer, and then they're just okay.

Andrew Moore:

Yes.

Reese Clark:

Actually they're getting that output,

Andrew Moore:

And there's no high estimate, low estimate, mid estimate. It's like one number

Reese Clark:

Wow.

Andrew Moore:

On a graph. And this is taking the decision, making stress out of it. And this is actually what some of our customers across Europe and the US actually want. They don't want five different options. They just want the most accurate one.

Reese Clark:

Okay. Yeah,

Andrew Moore:

That's what I would say.

Reese Clark:

Okay. Tell me about, I guess, a customer use case that you guys have. I think you have something listed on your website. If you can mention customers, if you can, or can keep it vague, but tell me, give me, just walk me through a story of there was a problem or we all have issues of what you actually solve for that customer and with your platform, Daisy.

Andrew Moore:

Yeah, yeah, absolutely. So yeah, I think a well-known processor would be like Kerry Dairy Ireland.

Reese Clark:

Oh yeah, we know that. Yeah, for sure.

Andrew Moore:

So yeah, I think basically they had a challenge around supply forecasting. So yeah, I think in Ireland especially, we have a seasonal cabling pattern where all of our cows ca it once, and based on the weather and the market and how many animals there is, it's really hard to understand how much actually milk some of these processors will receive.

Reese Clark:

Oh, okay.

Andrew Moore:

So what we did is we actually have a climate scientist on our team, and we actually model satellite data, weather data, market data, and their internal historical milk volumes. So instead of Kerry, for example, having to prepare all this work on the forecast themselves and constantly trying to update it and bring in new data and everything like that, we just have essentially all their data and all they need to do is just speak the number to their commercial team, to their ops team, to supply team, and yeah, that's essentially taking the work out of their forecasting. And yeah, we've almost a hundred extra speed of forecasting.

Reese Clark:

That's amazing. Okay.

Andrew Moore:

So this is some of the things that AI can do that I'm pretty excited about.

Reese Clark:

Holy cow. So I mean, you guys are clearly off to the races. You've got your launchpad set up. What's your vision for this company and new, I mean, I imagine you're thinking what can this platform do next? What are you looking into? I mean, AI's moving so fast, so maybe let's say six months to a year, two years, what are you thinking for what's next for Daisy?

Andrew Moore:

Yeah, so we're starting with dairy, and dairy is so complex. So if we can crack dairy, we can crack a lot of other different industries, right?

Reese Clark:

For sure.

Andrew Moore:

I mean, look, the thing is wide open from a competitive standpoint in terms of what we see. So I think in terms of where the vision is, I think everyone is actually on quite a similar page in terms of what this could do from not a lights out plant, but decision making autopilot that can essentially speak the language of what everything needs to happen to go in place.

Reese Clark:

Oh, okay.

Andrew Moore:

So that's where it's going. Yeah, we're starting with dairy, but grain, fish, beef, pork, poultry, all of these different industries have the same supply and demand issue.

Reese Clark:

Yeah. Why just stick with dairy in a sense? Exactly. Yeah. Obviously you want to master it, but yeah, of course. And obviously

Andrew Moore:

Dairy is the best, right?

Reese Clark:

Yeah, of course it is. Thank you for starting with us first. We appreciate it. Yeah, we definitely need that support. Why the name Daisy? Just curious. Yeah, tell me about that.

Andrew Moore:

So Daisy is a cow on our farm at home.

Reese Clark:

Okay.

Andrew Moore:

She turned seven this year and she had her fifth calf, and she's currently producing about 6,000 liters at about 5% butter fat and 4% protein. And she was our first pet cow, basically. So yeah, she's pretty special to us.

Reese Clark:

Absolutely. Yeah. And that name will live on forever. That's fantastic. Where can people learn more about your company, you, Daisy? Where would you encourage 'em to go find more information? Yeah,

Andrew Moore:

So it'd say Daisy.Inc. That's our website. So yeah, we have a lot of information there. Oh yeah, you'll find us on LinkedIn. Reach out to ourselves and yeah, we'd be happy to have a chat.

Reese Clark:

Awesome. Wonderful. Thank you. Thank you so much for working your way into the Milk Theory podcast. We're so happy to speak with you. You're clearly on the cutting edge of what's next for the dairy industry and egg in general. This is so needed. It's nice to see young blood in our field, and we truly wish you the absolute best. This has been really inspirational to listen to. So thanks for joining us here.

Andrew Moore:

Thank you for having me.

Reese Clark:

Wonderful. That's all for today on Milk Theory. We'll see you next time.