Software development is an act of empathy | Kilimo

Leading the Shift—a podcast from Microsoft Azure 
Episode 11: Jairo Trad | Kilimo 
 [MUSIC]
Jairo Trad (00:00):
We spend a lot of time developing technology but not talking with farmers, so product development is an exercise of empathy. It's about being in the shoes of the customer.
Susan Etlinger (00:16):
Welcome to Leading the Shift, a podcast from Microsoft Azure where we hear from the people at the forefront of data, AI, and cloud technologies. I'm your host, Susan Etlinger.
(00:26):
Thanks so much for listening over the past several months. We've really appreciated it and please keep letting us know what you think. Today we're going to talk about water, and specifically the intersection between water and technology. So let me just set a little bit of context. 60% of the human body is made up of water. About 70% of the Earth's surface is covered in water. And of that 70%, 96.5% is oceans, saltwater. And so that little remaining extra bit, that is what we use for everything else. That's the freshwater. And 70% of that is used for agriculture, 20% for industrial use, and the remaining 10% is how we brush our teeth and take our showers and make our food and all the things that we do. So really, really interesting and important conversation.
(01:17):
And the incredible innovation in data AI and cloud technologies is really helping us better understand how we use our water and recommend ways to help us use it more effectively. So, today we're going to talk about Argentina and we're going to talk about farmers and agriculture in Argentina. And if you don't know, Argentina and many other countries in South America and in other places around the world have been undergoing severe, severe drought. And so there's a company called Kilimo, founded by Jairo Trad, who is our guest today, and they are helping farmers become better stewards of their water, optimize their water production or their water use, and essentially balance that beautiful connection between crop production and taking care of our water supply. So, anyway, it's going to be a great conversation. Jairo, welcome to Leading the Shift.
Jairo Trad (02:12):
Hi, Susan. Very happy to be here with you sharing our story and sharing how we're leading the shift.
Susan Etlinger (02:19):
That's amazing. So one of the unique opportunities I think we have in the current moment is to look at how technology can help us better solve some of humanity's greatest challenges, things that we didn't have the data or algorithms or AI to solve in the past. And this is a life or death issue obviously, but the second is that this also seems to be kind of a perfect application of data. And so can you talk a little bit about how you view this opportunity?
Jairo Trad (02:46):
Yes. So when we started Kilimo, we had the sense that there was a big opportunity in solving a hard problem in agriculture. The hard problem is how farmers use water. The initial and typical approach to that problem was let's put a sensor on the farm, let's put a device in mail for a farm and measure things and be more efficient by measuring directly. That's awesome. The problem is that that also means that you have to leave a small device in the middle of nowhere. If it breaks, you have to go and maintain it. So the IoT approach was a hard approach to implement that scale, but we early on understood that there was a gray shift in the availability of satellite data. Launching satellites to the space has been getting cheaper and cheaper and cheaper by orders of magnitude.
(03:47):
And by leveraging that new data, we were able to develop a tool that provides irrigation advice to farmers only using data, a combination of satellite data and weather data. And that tool combines also AI to provide clear advice to farmers, you have to irrigate half an inch in the next week and that has opened for us a huge opportunity around the globe to help farmers be more efficient.
Susan Etlinger (04:13):
Clearly it took a lot to get you to that point, and so I'm curious to hear a little bit, first of all, about your own journey, and then also sort of how you founded Kilimo, how the company came to be and all of that.
Jairo Trad (04:25):
Yeah, sure. So I'm originally from the countryside of Cordoba in a valley that's called Traslasierra Valley. I was born and raised in a small town called Luyaba. It's a 700 people town. So starting there, as a valley, it was a very dry area, it rains very little there. So I always had a connection with agriculture with the drought and the fact that water is everything, but that didn't click it deeply with me. So I ended up studying computer engineering. Don't asked me why because at some point in my life I got in love with computers and technology and I understood that technology was a superpower to change things.
(05:16):
So I decided to study computer engineering, but as why I was wrapping up my university journey, I saw an opportunity to start a company and a startup in the agriculture space. Again, no reason in particular just because I understand that agriculture was hard and complex, and somehow I ended up approaching the problem of the intersection between water and agriculture and that's where everything fit together. So I started one company, we ended up closing that company because we failed, and that was very good too for us to learn many things that we didn't have as a background from university. And then-
Susan Etlinger (06:01):
So actually, give me an example. What things did you learn that you are applying now?
Jairo Trad (06:06):
Wow. So one of the main learnings was that we spend a lot of time developing technology but not talking with farmers. So in the end, product development in general is an exercise of empathy. It's about being in the shoes of the other person, of the customer. The problem is that farmers are a very distant type of customer. So when you see every successful technology, you're going to see that the demographic of that technology in at least the initial demographic was a very homogeneous demographic. So everybody was sort of the same. Farmers are all different. Agriculture is if you move a hundred kilometers north, agriculture is different than agriculture a hundred kilometers south.
(06:54):
So in the first company we didn't spend enough time talking with farmers, so in Kilimo we spent the first six months doing nothing but talking with farmers. And then we came back to the office and then we spent some time doing actual coding, developing the first version of the tool. And somehow, eight years later that tool is still live and farmers are using it. And it's very, very similar to the initial development we did with a lot more technology on the backend, but very similar because we took the time to understand how unique and different are the needs of these customers that we weren't able to connect with the first company.
Susan Etlinger (07:38):
Yeah, it's so interesting that you say that because this theme is coming up over and over in the conversations that we're having and frankly whether they're social impact related or human impact related or whether they're commercially focused. And so we've seen this in financial services, we've seen this in healthcare and rural communities around the world in the global south that need to really get close to the customer. And one of the people I interviewed early on, Perry Hewitt from Data.org, had said that earlier in her career, like you, she had sort of underestimated the importance of trust and proximity in co-creating sort of sociotechnical tools for customers, whoever those customers or patients or consumers or citizens might be. So it's really interesting that you say that. I'm curious, what was that process like of sort of getting to know the farmers, building trust with them, and sort of envisioning these projects?
Jairo Trad (08:36):
That process was a long one because it took us several months. Second, we didn't have a deep connection with industrial agriculture in Argentina. We didn't have neither family or friends in the space. So it was a game of just attending events of the topic and trying to meet a friend of a friend. The good thing about Argentina and many Latin American countries is that agriculture is a huge chunk of the GDP, so it's everywhere. Every time you scratch a little bit, you find somebody who is connected to agriculture. So we were able to find the farmers, but it took us a while to just keep asking until we found them. We have the exact number.
(09:16):
In the truck of my founder, we ended up traveling in over 50,000 kilometers around Argentina, going to the events and things over those first six months just to meet people. But it was great and was the only tool that enabled us to build a product that was needed and was clear for the farm. Even when at the same time we were developing the product, the technology was just being developed. So we were doing this in 2014, 2015, and the satellites were using right now we launch on 2015, 2016. So we were sort of [inaudible 00:09:53]. At the same time the technology was being developed, the backend technology we need, we were doing this slow and very thorough product development part of the business that this more than half of the problem of course.
Susan Etlinger (10:09):
So how did they initially respond to this idea of putting sensors on their farms or even the idea of data collection at all?
Jairo Trad (10:18):
Well, what you see is that agriculture is plagued, but another big problem that is loneliness. In the US for example, there's a high rate of suicide between farmers. It's a lonely profession. You are alone in the middle of nowhere. And this is not a figurative nowhere. This is a literal nowhere. Middle of the nearest town is sometimes 30, 50 kilometers, so you're away. And farmers are alone and they are extremely open to get you in that farm. Initially and still we always find certain level of distrust from them, but that's 100% okay. First, we are saying to them that a bunch of young people are going to tell them what to do in their farm, something that they have been doing for 50 years or more sometimes.
(11:13):
With satellites, not even installing anything on the farm, just with data, we're going to tell them from the sky what to do next. So that promise, when you see it from the point of view of the farmer, and even from the point of view of me initially, it's pretty similar to magic. So when you're telling that customer you're not going to make magic, and that is always tricky and it needs a level of distrust. And also agriculture is a risky profession. So farmers on average lose one of five crops because you get hell, you get something happening to a crop, you lose one of five. And you have one crop per year, sometimes two, sometimes. So you get very few chances at success and risk is present all the time. So they're very open to get you into the farm, but they're very thorough when evaluating new technology and making changes, and there's a reason for that and we understand that that's part of the game.
Susan Etlinger (12:16):
Okay, so I understand the trust piece because you are young people, you're coming, you're showing them all this technology and making promises, and I don't know, I can't remember who said it, but at some point somebody said at the most advanced level, AI is indistinguishable from magic. And I think we've all been living through that, especially the last few years. It just feels like magic and then you start to understand what it is and that it doesn't feel like magic anymore and the next thing feels like magic. But more than that, more sort of trusting the technology, I also thought as you were speaking maybe was there any sense that by participating in this project that they would have some kind of collective understanding that there would be some way to build community among farmers digitally? Was that part of the equation at all?
Jairo Trad (13:06):
We never factored that in as a fundamental part of our business. We never tried to build a social network of farmers. That was not the goal, but then that happening, so we build something that's called the Irrigation Academies, where we train farmers, we do that online and in person. We have trained 120,000 farmers across Latin America with the academy. At some point at the height of the pandemic, when everybody was at home, we used to do one webinar every two days, 180 webinars in a year with 300 farmers from Latin America on each webinar. So we understood that one of the key blockers of adoption in ag tech, so adoption in technology for agriculture is really, really low. In the US using all with farmers, very incentivized and with a lot of support to adopt technology, adoption of irrigation management practices like ours is lower than 10%. That was the last USDA survey. so-
Susan Etlinger (14:16):
What do you attribute that to?
Jairo Trad (14:21):
We see there are two main pillars. The first one is education, so learning and sharing, and because education lowers the risk. If you know that your neighbor is doing something that works, that lowers the risk for you. So that's key, that community education is key. And the second one is the one who has propelled our company or technology is incentives. What happens with water is that water for agriculture is essentially free and there is a great reason for that. The reason is we want food, we want to produce food and we try to keep water the main resource needed to produce food as cheap as possible, but that's in general a terrible incentive to conserve water. Why would you conserve something that is free? And everybody's not only giving that research for free, but also provide a strong incentives to produce more.
(15:16):
For example, in the US, there are many areas where you have use it or lose laws. If you don't irrigate all the water you have allocated, you lose it in the next year. That's a clear incentive to use more water. So policy is going against efficiency, economic incentives are going against water efficiency. So at some point we understood we had to break that, that was working against us, and maybe everybody else in the space, and we managed to find other stakeholders that were sharing the resource that were very eager to pay for somebody else to be in a good water stewardship. And that's what we did, we connected companies like Microsoft for example, that have a water-positive goal with farmers that shared the resource in watersheds across the globe. And by that we were able to provide incentives to farmers to be more efficient, and that's what has been doing in the last few years.
Susan Etlinger (16:11):
So let's switch gears, because I think it's so important to understand the problem, but let's also talk about what the projects look like. What happens on day one? How do you actually bring this technology to farmers? Obviously, they're starting to get satellite data. Just walk us through what the technology looks like?
Jairo Trad (16:32):
On day one with the farmer, we work with them to understand the boundaries of the field, their goals. We have a small interview with them to understand where and when, how much water they have available, how much water they can apply, if they have any restriction in general, because again, agriculture is very unique to every area. And then with that, we just draw the field, it's literally drawing the boundaries of the field. And then our technology start collecting a combination of satellite data and weather data. From satellite data, we combine five public satellites that are available for everybody and 150 private satellites that we buy data from. And then for weather stations we do the same, so we have a crawler that is crawling every publicly available weather station around the world, focus of course in areas where we have operations and then we buy data from some private providers and we combine that data.
(17:40):
What we do is something that is, and where we have the bulk of our technology and our knowledge is in doing a lot of machine learning and AI to sensor fusion, that is mixing different sensors from different types because we had a weather station from the local government, we had a weather station for a local co-op. And those two are different brands with different sensors, or combining the data requires a lot of processing and adjusting and technology. And the same with satellites. Satellite data has a high noise to signal ratio. Yeah, it's very, very noisy. It's getting better and better, but still very noisy. So what we do is build a magic satellite that mixes everything together and provide a very smooth monitoring on what's going on with the crop.
(18:33):
With those two data sources, we can know how much water that crop uses on every given day, and we can do that from the past and do it from yesterday. If you know how much water your crop is using, you can know how much water you need to apply, and that's what we do, we provide advice by knowing how much water each crop is using.
Susan Etlinger (18:56):
So you're doing, from what I hear, a vast amount of data preparation in order to be able to send clear signal and clear insights back to the farmers. Did you create a benchmark? How did you figure out not just what they are using but what they should be using? And do they opt in their data so that you can... How does that work?
Jairo Trad (19:18):
No, we model a balance of inputs and outputs and you always try to... There are clear benchmarks of where you should be in the balance at each part of the cycle. So in the area where you have, for example, you are going maize, in the moment where you have the seed production of the corn, you need to be at 70% water on the water balance, Seven, zero, yeah? So we try to keep-
Susan Etlinger (19:52):
A stupid question though, stupid question for you, Jairo. How do you know it's 70%? Where did that come from?
Jairo Trad (19:58):
Because we keep a balance. We keep the inputs, the farmers report us how much water they irrigate. If they have an automatic way of reporting that, we gather that, and we have some satellites that can estimate how much our irrigation he's applying. We know how much water he's putting there, we know how much water is getting out from the evapotranspiration. That's the technical parameter we measure with satellite wear data so we know how much you're getting, how much is getting out. And by keeping track of those two, you have a balance, you have where you are in the field in that specific moment and we try to keep the farmer below certain threshold that is preset at the beginning of the season.
Susan Etlinger (20:41):
And so how did the farmers receive the data? Do they have an app that they use?
Jairo Trad (20:45):
They have an app, they have a mobile app that works offline because connectivity is not available everywhere. It's not as harsh as it used to be. Things are changing, you have satellite internet now that works very well, so it's changing but still.
Susan Etlinger (21:01):
Yeah.
Jairo Trad (21:02):
And then you have a web app they can access with more data. But the key thing here, and thanks for that question, is that we aim to provide beyond charts and analytics, and this happens all the time in ag tech, you have a lot of charts and analytics. We aim to provide them with clear advice, you have to irrigate half an inch next week because we understood that that was key for the farmer to act. They're very busy and they're also in harsh condition, most of our farmers are using the app in the middle of the sun, sweaty, maybe with the gloves on because they're driving a tractor or something else. So they are in a condition where analytics don't provide a lot of value, but clear, readable, direct advice of what to do next do provide value for them, so that's what we do on there.
(21:58):
We also use GenAI and NLP because we've attended what to do next. You have to irrigate half an inch next week and that's easy for them to digest, and they also can adjust how that advice count. Some farmers prefer to understand things in inches, some farmers prefer to understand in our words, "You have to irrigate two hours tomorrow." So adjusting that, it's not that hard from a natural language processing point of view. And that's what we do.
Susan Etlinger (22:25):
Oh, that's fascinating. We were having a conversation with a gentleman from EUI about their neurodiversity programs and working with neurodiverse developers to reduce notification noise because it's so distracting. And it's distracting for everybody, but it's particularly distracting for neurodivergent people. And of course anyone who is in a situation in which focus is critical in which the environmental noise around you, and I don't mean literal noise, but just the sun, the heat, the sweat, just the actual physical activity and the need to keep track of all of these different things happening is so much that it really ramps down the kinds of notifications that you would want to send, and you must go through some kind of very delicate balance with that.
Jairo Trad (23:16):
That's awesome, Susan. Well, something that happened to us by chance is that the first template we used to develop the app was a template with very bright colors. Yeah, I can't say that it was a brilliant idea from us. It was just by chance, but it works so well because you're in the sun, you need bright colors.
Susan Etlinger (23:39):
You need the contrast.
Jairo Trad (23:41):
You need the contrast. And that was really good for us, for example. And those small details come back to what I was saying before, this exercise on empathy, on understanding the type of customer you're talking to, and something that you also brought up before, and I think that this is interesting in this new age of AI, is that there are many common problems that are already solved by technology. I don't know, buying something and get it delivered to your house, calling a car to go to somewhere else. So they're like a fixed type of technology work, but what happened with this enormous amount of problems that are extremely different and extremely personalized, and there is a huge opportunity for AI to adjust the interfaces, to adjust how you talk to the user, to adjust how the app works using that specific context, and agriculture is that type of problem.
Susan Etlinger (24:46):
Yeah, it seems like the perfect opportunity. And of course as the technology develops, maybe that even happens with less involvement from you or less involvement from them because the signals are coming in and you can think about how they're responding as a way to adjust what that process looks like. I think that's fascinating. So it kind of brings me to what's next. What do you envision as the next kind of stage of this technology? Where do you think it's going?
Jairo Trad (25:23):
We are growing a lot because we managed to solve the systemic barrier. The systemic barrier were incentives, so we're expanding. Right now we are operating in 15 watersheds in Latin America and the first one in the US together with Microsoft and other partners in Latin America. And we are seeing a strong traction because more and more companies and governments are starting to understand that if we don't correct how we perceive and treat the value of water, we're getting ourselves in big, big trouble. And this is of course exacerbated by the climate crisis that has a huge impact on the water resources, either to severe drought or flood that is only the other side of the coin for severe drought. So we think that it was a long journey and it is still a long journey, but we are at the right moment with the right solution, and especially with the right configuration and the solution to move the needle.
Susan Etlinger (26:30):
Yeah, I think you're making your way through that pie chart, where if I'm correct here, 70% of freshwater is used for agriculture, is that right? A little over 70%, and then about 20% for industrial purposes. And I think the remaining is for the rest of us to take our showers and drink our water and live our lives, right?
Jairo Trad (26:51):
Yes. And that's also, that's average numbers, average are always unfair. If you take developed countries, it's more like 85%. So if you don't find a way to realign incentives in food production, we don't have a chance. So it's not only about what Kilimo is doing, it's something that we have to solve. Kilimo is just exploring one vector and has opened a huge vector of opportunity and many other irrigation providers are going... I'm moving to the model we develop and discover. And there are many other players that understood that by connecting this industrial water user that can put a bigger value on water because they have a biggest amount of margins and returns, and also a biggest investment on, you pay a lot more to set up a data center than to set up a farm, so you have a biggest investment there. So this industrial stakeholders that have a much more risk on the line can help these other stakeholders that need better incentives to change the things the way they're doing things.
Susan Etlinger (28:10):
So it seems like this model, certainly as you say, it has application for other areas of agriculture or agriculture and other areas of the world. It has application industrially. I feel like it also just the model in terms of how you built the trust, how you came to the empathy-first approach to understanding the problem is something that's so applicable to other industries irrespective of what they make or do or save or care about. And I just wonder, as a leader, how big is Kilimo now? How many people do you have?
Jairo Trad (28:47):
70 people or so.
Susan Etlinger (28:50):
70 people. Okay.
Jairo Trad (28:51):
35 cities and towns.
Susan Etlinger (28:54):
Wow, that's amazing. What a ratio.
Jairo Trad (28:56):
Remote.
Susan Etlinger (28:57):
Yeah, you guys are spread thin. Okay. But in terms of as a leader, what have you learned? Because you must have started this, you started this, you've learned a lot. What would you share with others?
Jairo Trad (29:12):
I think that we have talk about how technology sometimes is magic, meaning like a superpower for something that it's a tool that can move the needle. What I would share is that I prefer to tackle problems. And I think that we should be very mindful of using this superpower. I mean not judgment for the FinTech, whatever, but we must use this super power and this common people magic that is available to move the needle. And that's our internal definition of success, is not about raising money or an exit, internal definition of success is this company is going to truly move the needle, but truly in a meaningful, measurable, global way. And somehow by just keeping that as the north parts of every effort we were doing, we got there.
(30:19):
We think that we're right now truly moving the needle and in the areas where we're not moving the needle enough, we are doing the steps that we need to do to be there. So it took us a while, it took us eight years to get product market fit only to find the right type of stakeholders engage together. But if the outcome of that, the long effort is that you're solving a planetary problem in a scalable way, then eight years is nothing.
Susan Etlinger (30:49):
I love that advice or that passion about if we have this technology, it's a privilege. It's a privilege that we have these powerful technologies that use that opportunity to solve something meaningful, and meaning it's so culturally constructed, it's so different, like place to place, area to area. So whatever's meaningful for you, I think that's incredible. And I really appreciate the way that you've approached it because it's such a great example of why it does sometimes take a while for product market fit and we think about the pace of these technologies moving so, so quickly. But really a lot of the point of this podcast too is to sort of stop and reflect and think about what kinds of decisions should we be making, what kinds of questions should we be answering? So thank you so, so much. It's been a pleasure to talk to you.
Jairo Trad (31:45):
Thank you, Susan. I am very happy for this conversation. It was really great and I'm looking forward to hearing more about what you guys are doing here at Leading the Shift.
Susan Etlinger (31:55):
Thank you. Okay, I've had a little time to process the conversation with Jiro and I keep coming back to this Albert Einstein quote. You've probably heard it. "If I had an hour to solve a problem, I would spend the first 55 minutes thinking about the problem and the last five minutes thinking of solutions." And I've always thought, "Okay, sure, you're Albert Einstein, you get to do that. But what about the rest of us?" And then I'm at Jiro. So can you imagine this? Can you imagine getting into a truck and spending six months driving over 50,000 kilometers to meet your customers, hear about their needs and earn their trust?
(32:32):
And that's exactly what Kilimo did, even though it can have been either psychologically or physically comfortable because above all, they believe that product development is an exercise in empathy. And that empathy plus the superpower of technology is what they're using to solve one of the world's most consequential problems. And so I hope this conversation with Jiro inspires you no matter what it is that you're building to do two things, combine your superpower with your common people magic to move the needle. Thanks for listening.
(33:09):
We hope you've enjoyed this episode. And as I'm sure you know, new podcasts live and die on engagement. So please like, comment, share, tell your friends, and let us know what you'd like to hear about. We're listening. If you'd like to learn more about how people and organizations are innovating with Microsoft Azure, visit azure.microsoft.com. The Leading the Shift podcast is a place for experts to share their insights and opinions. As students of the future of technology, Microsoft values inputs from a diverse set of voices. That said, the opinions and findings of our guests are their own, and they may not necessarily reflect Microsoft's own research role positions. Leading the Shift is a production of Microsoft Azure. I'm your host, Susan Etlinger. Our executive producers are Haley Meehan, Aaron Russell, Erik Williams, and me. Production support provided by the incredible team at Indigo Slate.
[END]