AI and Agriculture: Using AI To Provide Quick and Effective Advice to Farmers with Dr William Derban

[00:00:00] You're listening to the Unlocking Africa Podcast.

[00:00:30] We use AI properly in Africa. If we don't allow it to pass us by, we can see a lot of growth using AI on the continent. Stay tuned as we bring you inspiring people who are unlocking Africa's economic potential. You're listening to the Unlocking Africa Podcast with your host, Terser Adamu.

[00:00:54] Welcome to the Unlocking Africa Podcast, where we find inspirational people who are doing inspirational things to unlock Africa's economic potential. Today, we have Dr William Derban, who is head of digital programs and partnerships at the Digital Innovation Group at Opportunity International,

[00:01:15] an organization that uses financial services, technology and training to create opportunities for underserved communities and those living in poverty. Welcome, welcome, welcome to the podcast. Dr William, how are you? Thank you very much, Terser, and greetings to all your listeners. I'm very well, thank you. Fantastic. How's your week going so far?

[00:01:41] So far, it's good. It's nice and sunny. And today is also Ghana's Independence Day. So happy independence to all the Ghanians out there. Happy Independence Day. Indeed, indeed. Thank you for joining us. I know you've listened to the podcast before and we've spoken before. As usual, I like to get straight into the conversation. I was hoping you could give us just a brief introduction into who you are and Opportunity International.

[00:02:11] Thank you very much, Tessa. So as you introduce, I look after programs and partnerships in our Digital Innovations Group and what we do within the group to see how we can use technology to solve problems around poverty reduction. I am a financial inclusion expert. I've spent over 25 years of my career working in banks and financial services and also with telcos and recently with NGOs.

[00:02:39] And my whole career has been around how we can use technology around financial services, how to provide financial services to the poor. So that's what I've been doing for the past 25 years. So if we look specifically at the innovations and technologies that you've worked on at Opportunity International, can you give us an overview of some of those, whether it's the initiatives or the activities you've been involved in?

[00:03:05] Yes, thank you very much, Tessa. So Opportunity International, as you pointed out, is over 50 years old and innovation in its broadest terms and digital as well has been part of our DNA for quite a long time. So when we started earlier on, we started lending money to micro enterprises in Latin America. In Africa, we founded institutions, banks to lend to the unbanked at a time that the banks didn't do that.

[00:03:32] We started micro insurance as well. And when we found that insurance is an issue and we've done agency banking with all our current banks, making sure that, you know, the basic infrastructure or digital infrastructure to save the poor is in place. But recently, over the past couple of years, we've really been looking at how we can use new technology.

[00:03:56] And by that, I mean really around MGen AI and how we can use that to really change how we deliver services to the poor. And my team that I work for have been doing some serious deep dives to understand our client base and create what we call personas.

[00:04:18] So we understand the lives of the people that we seek to serve, because by doing that, we can design solutions that can work best for them. So we have solutions around agriculture and education, and we are expanding into micro enterprise. So the bulk of my work has been around how do we make sure that we get the right personas really understand their needs

[00:04:45] and ensure that we are designing solutions using new technology to improve their lives. As you noted there, the initiatives or the technology is designed with the end user in mind. I know one of the main end users, as you mentioned, are those in agriculture and farming. So how are these innovations impacting smallholder farmers in the communities that you're serving? Thank you very much. That's a really good question.

[00:05:13] Now, you know, especially in Africa, we have between 60 to 70 percent of Africans, especially sub-Saharan Africa, engaged in smallholder farming. So that's a lot of people. And so agriculture has become one of the very main areas in which we are creating our personas. And we do have one persona that we call Faustina. So Faustina is a smallholder farmer.

[00:05:38] And when you go to opportunity and you talk about Faustina, everybody understands who Faustina is, the sort of challenges she's facing and, you know, how she lives her life. Because one of the challenges that we face when we are designing solutions, and that's not just an opportunity, I think generally, is that we kind of generalize. So you say SME. Well, who is an SME? Or even say smallholder farmer. Well, what does that actually mean? That's a whole range of people. You know, somebody from one acre to two acres.

[00:06:10] So we've managed to narrow down to almost like a persona that this is the sort of person that, or type of person that we really want to find solutions for and then escalate from there. So we have solutions as writing for the farm, for the smallholder farmer, for Faustina. I'm happy to elaborate more on the solutions. Yeah, so I guess one of the solutions you created is the Ulangizi AI app.

[00:06:40] I was wondering if you could give us some detail regarding how it works, and maybe some of the impacts it's actually had on the farmers. Yeah, so you've pronounced it well, it's Ulangizi. So Ulangizi is a chichewa word, which means sort of guidance and advice. Yes, but to explain how it works, maybe I can take a few minutes just to explain how we got to that point and, you know, how we started to understand the needs of the customers.

[00:07:10] So a group of us went to Malawi, which is one of our main markets. And Malawi, specifically also because it's one of the poorest countries on the continent. So we went there and we spoke to our clients who are farmers. So we found people like Faustina, and we spoke to them trying to understand what their challenges were. Now, speaking to them, at least putting my banker's hat on, I thought if you ask them what their challenges, they would say, well, it's finance, you know, as many people would say.

[00:07:39] But no, actually, they did say it was information. Because they've had a cyclone that had come in the past and the cyclone had destroyed a lot of the air crops. So a simple question like what to plant, when to plant, you know, all these are not simple anymore. They're not as straightforward anymore. And so information really came out strong. So our engineers put their heads together. And then we asked a question, if somebody could provide you accurate questions quickly, would that be useful? And they said yes.

[00:08:09] And where do they get that from? They get that from their extension officers, you know, who come there, you know, once in a long while, you know. So it means that getting information, and even for the extension officers, the information has to be ready. You know, it's in a big book. So how do they, when they go into the field, they have to carry the big book with them? And that's also a challenge. So our engineers got together and they decided to create this app, you know, Ilongizi. That's how it was born.

[00:08:37] But what we did is that we took the good agricultural guide. Because the problem we are trying to solve here is for the smallholder farmer for Sina to be able to get information accurately, but also quickly. So especially for the accurate side, we had to go to the government and then get the authoritative information from their good agricultural guide. So that's their sort of their Bible for agriculture in Malawi.

[00:09:04] So we put that into the bot through what we call a RAC model, which is retrieval augmented generation model, so that the bot would only answer questions from this one source. Because one of the challenges that you face with these kinds of bots is that, you know, if it goes onto the web, it will give you all sorts of information. And then you have what we call hallucinations. So the answers that it gives may not be that accurate.

[00:09:31] And agriculture is a very important part of any economy, especially in Africa. You've got 60 to 70 percent of the people engaged in smallholder farming. Then information that you give them is really critical and it should be guarded. So we took that information to make sure that it's accurate. Then we also put it on WhatsApp because we realized that, you know, if you put it on a laptop, how are they going to access it? They don't have those three.

[00:09:59] So we put it on WhatsApp because we realized that the extension offices, you know, all of them have got smartphones. Maybe not the farmers directly, but at least the extension offices. And we also have our own farmer support agents who go and train the farmers. So they have got smartphones. So we said, well, why don't we give it to them? You know, and then they, when they engage with the farmers, they can explain things to the farmers.

[00:10:24] So the solution really is as simple as the extension office will go to a farmer. A farmer will ask a question. They will type it into the app or they can actually speak to the app, you know, and to speak to the app. And the app through the chat GPT will go into the good agricultural guide and provide an accurate answer and even give you the source, which page in the book.

[00:10:49] If you ever can open the, find the book, you can go there to tell you exactly where it's from so that it's accurate. We went a bit further and put it into Chichewa. So it's not just in English. It's also in the local language. So actually the farmer can take the phone and actually speak the question, actually say the question, and then it will come up with the answer. And the answer will be translated into Chichewa for them. On top of that, we also put in the possibility to take pictures.

[00:11:18] So now when you go into the farm and they see a plant, they can take a picture of the plant and it will tell them, you know, what type of plant it is, but also if it's diseased and what sort of disease it has and what, you know, what the remediation will be. So it's a very nice package, you know, and it does what it says on the tin. It provides an accurate answer and it's also very quick. You don't have to wait in seconds. You've got, you've gotten an answer and you can trust the answer because it comes from a source. So we finished our pilot.

[00:11:47] We had about 150 farmers using it. It's successful. We are now in the process of scaling it. And the impact, as you mentioned, at least from the pilot was really good. The farmers really enjoyed it. I was fascinated at the way they engaged with the technology. So they were not interested, oh, this is Gen AI. I mean, they don't know what Gen AI is. But the point is that I need, I have a question. I can type it here or I can ask, see if I'm talking to a friend, you know.

[00:12:16] So in a way, it doesn't matter who is at the end of it. You know, I'm still getting this in my language. I can hear it come back to me in Tishewa, which was fairly accurate. So we think that it has great potential in going forward. Interesting. So the Oolongizi app was focused on Malawi. You also launched the farmer AI pilots in Kenya and Ghana.

[00:12:40] What are some of the key differences that you've seen and how AI is being applied in these different regions? Yeah, thank you very much. So in these two countries, we looked at different things. So in Malawi, we worked with our own government extension officers, but also with our own farmer support agents. You know, so we are scaling it amongst our programs. But in Kenya, we had the opportunity to partner with Digifarm.

[00:13:07] So Digifarm is an agri-tech company owned by Safaricom. And Safaricom being the largest selco in Kenya. That would really give us reach and scale because, you know, everybody uses Safaricom services in Kenya. And Digifarm themselves have a large platform. So Digifarm provides financial services, mainly digital financial services to farmers.

[00:13:30] And the idea is that if we are able to improve the capacity of the farmers to earn more and to grow more through providing information, through an advisory board, then they would grow more, they would earn more, they would be able to apply for loans, they would be able to repay for loans because they are earning more. So that's sort of the general theory of change. And there we started off with the potato value chain. And it's still early days. We just launched this last month.

[00:13:59] But we are seeing some uptake of the farmers. So they are blasting SMSs to the farmers, telling them about it. And we are seeing a gradual pickup of the product there. In Ghana, we partnered with the Development Bank of Ghana, which provides loans to local banks. And then Gersar, Gersar provides guarantees and also has content. OK, so they provided us content from value chain in Ghana to we started with a rise value chain.

[00:14:27] In Malawi, we didn't have any particular value chain. We just opened it up. But in Ghana, we really wanted to look at the rise value chain to see how that would work if you narrow it down to a value chain. We would expand going forward. So we've done that. We launched that again last month. And we are beginning to see people using it, the rice farmers using it. So the question there again, the questions that you ask would only be around rice farming until we expand it.

[00:14:56] So these three projects, they are all looking at different value chains, but also looking at different forms of partnerships. Because one of the key things that we need if you are going to scale these solutions is partnerships. No one can do it alone. But we need to find the right sort of partners that we can work with in different situations if we have to grow and scale these solutions. So that was the reason why we have these different partnerships. And we are still looking for more partners to work with.

[00:15:24] So would you say that partnerships are a deliberate strategy to ensure that these solutions are accessible to farmers? Oh, yes. Partnerships are key. And I think for Africa, we need more collaboration. You know, there are lots of things that we work in silos, you know, but we really need to come together. At the end of the day, we are dealing with similar situations. It's amazing when you travel across a continent, you see very similar situations.

[00:15:52] You know, the same type of smallholder farmers planting very similar crops, having the same challenges. So partnerships is key. But secondly, we need to bring different stakeholders to the party. You know, it's not just about, you know, just dealing with farmer support agents. Indeed, with technology. Now we are working with people like Microsoft. You know, we've talked to Google. And these are not your traditional agriculture organizations.

[00:16:18] But they have solutions and expertise that can help in these kinds of solutions. So collaborations and partnerships are key. I mean, governments need to be part of the picture. We cannot go to a country and talk about large scale development in all areas such as agric or health or education without talking to government and making sure that it's aligned with their government plan. So, yes, you're right. Partnerships, collaborations are really key.

[00:16:48] So partnerships are key to accessibility. But how do you overcome the challenge of giving access to people, farmers who have limited digital literacy? Yeah, you phrased a really good point. Digital literacy is really important. But, you know, if you find solutions that really meet the needs of people, it makes digital literacy or providing digital literacy easier and better.

[00:17:16] Remember the days where mobile phones had huge manuals. It came with huge manuals. You had to read. Now you go to the rural area and everybody has a phone. And there is no manual. It comes with nothing. It's just the phone in the hand of the person. That's it. And then people use it. You know, people will get their children who are educated to teach them. So, you know, the use of the device itself, you know, if people can understand the need.

[00:17:42] And that's why the human-centered design and really getting to understand the challenge you are trying to solve is important. Because if the need is there, you are catching the person's need, the person is more likely to use that device. But digital literacy goes beyond just the use of the device. That's just one thing. I think the big issues there around data security, you know, and protecting yourself because now you're opening yourself to a lot of information.

[00:18:11] So there we are creating videos. So before you can use the app, you need to consent because we are collecting some amount of information from you. We are collecting your mobile number, for example, which is an identifiable data point. So we've got to make sure that they consent. And what we've done is that we've created videos, cartoon videos, just to break down these very complex situations. Because, you know, PII, data privacy, data security, these are big topics and they are very heavy.

[00:18:41] And to try and explain them to people is not easy. I don't understand everything about it. So we've sort of created videos that make it very simple for people to understand why should you protect your data? What is your data? What does it actually mean? And if I put data somewhere, where is it actually going? So we are really trying to educate people on the implications of using a digital service so that you can identify fraud when it comes to you.

[00:19:10] When somebody puts fraud, you can see the information that you are getting. You understand where it is coming from and the limitations to that. So it is a process. And that's another area that we need collaboration because it's not just about our agri-solution. Digital literacy is about how you, you know, what the implications of digital to your life. And that would be in the finance. Other people are providing them mobile money. Somebody else is providing them another service over digital, you know.

[00:19:36] So they need to be very digitally literate across many things and at a very basic level so that they can identify when there is a negative threat to them. Thank you for sharing that. You mentioned something quite interesting there, which was the increased access to devices such as smartphones, which were often seen as products or items for the wealthy.

[00:20:02] So is this one of the main vehicles used to ensure that the benefits of the solutions that you're providing reach underserved communities? Yes. So at the moment, our solutions are based on the smartphone, which is, you know, there are still more smartphones, but there's still a lot of people who don't have them and cannot afford them. So that's one of the big challenges that we face using smartphones.

[00:20:31] And there are two ways in which we are going around it. So the first is working through what we call a human in the loop, which are like our farmer support agents. So if the farmer themselves may not have a smartphone, you can give the solution to the extension officer or the farmer support agents or the human in the loop who would have a smartphone and then can use that to serve the end user or the farmer or whoever doesn't have a smartphone.

[00:20:59] So that's one way to get around it for now until they can afford a smartphone. Because for people who are poor, a smartphone is a heavy investment. So we've seen cases in farms where they buy, when they harvest and they get a lot of money, they buy a smartphone. In the lean season, they sell the smartphone and get money because, you know, because they see the smartphone and I need money. Well, I'm just using this to make calls. I might as well just sell it and then, you know, have money.

[00:21:27] But you and I will never sell our smartphones in times of need because we need it to work. You know, not just for calls and for fun. We need it for service. And that's what we really want them to appreciate. That's why we need to create more use cases for them so that when they buy a device, it's not just for calls and just to, you know, have it because I now have money. I want something nice. But I'm using it in my job. As a farmer, this item is actually helping me improve my livelihood.

[00:21:56] Therefore, I would keep the investment. But we take time to get there. So a human in the loop helps that. And then secondly, the technology is also growing. So now we are exploring ways in which even if you have a feature phone, which most of them would have, you can dial an access code and you will be able to still access the AI services. So that's another solution that we are working on. And, you know, once that is made available, then, you know, if you don't have a smartphone, it doesn't matter.

[00:22:26] You can dial a code and then it still takes you to the same service on the smartphone. But it's still something that we all need to collaborate and work on. Thank you for sharing that. We touched on earlier in terms of the partnerships and scaling and access. Yes. Outside of this, what would you say are some of the main barriers to scaling AI-powered agricultural solutions across the continent? Yeah. One is language.

[00:22:56] So most of these smallholder farmers, as you can imagine, they all wouldn't speak English. And, you know, that in itself is not their problem. But these, the models that we use for our AI-s are mainly in the large languages like English, French and others. And so getting these models to speak the local languages, you know, it's improving, but that's still a challenge. And also the nuances in the language, you know, so I come from Ghana. So you take Akan.

[00:23:25] Akan is a very broad language. Now within Akan, you have Fanti, you have Ikwapim, you have other, you have Chi, you have other languages there. So it means that if you want a farmer in one part of Ghana, even speaking Akan, it's not quite the same dialect as somebody else where. So how do we get these languages? Because if you leave them in English, then it defeats the purpose. It means they're uneducated farmers who cannot access these services. So that's one big challenge.

[00:23:53] How do we get all these languages that people speak across the content? How can we get them into models that, you know, chat GPTs and these AI models can learn and produce solutions for? So that's a big one. We've done that for Chichewa and Swahili for our farmers, but there is still more work to be done. The second I would say is content. So I spoke about the authoritative content. And what we have to guard against is hallucinations.

[00:24:21] You don't want people just to go in the chatbot and get information from anywhere because there could be biases, you know. So a lot of the information are Western based, you know, if you go on the net. So if the farmer is somewhere in Malawi and you just leave them up onto the web, the chances that they will get some biases is high. So how do we get authoritative content for farmers across, you know, across Africa?

[00:24:50] And how do we ensure that that content is also updated? You know, so it's not that it was done five, six years ago and nobody put in the money to update it. You know, so how do we ensure that this content is updated? It's, you know, it's correct information. We've talked about the changing climate. How is that impacting the information that we have? Because otherwise you'd be giving them old information, which doesn't help. So that's another challenge that we face.

[00:25:16] And any, all the patterns that we work with, we make sure that, you know, the content is up to date. Then we also have implementation. So how do we implement these solutions? We need to sensitize people still because the farmers, they've got a way of doing things. If they have a challenge, who do they ask? They ask their fellow farmer next door. Where did he get his information from? You know, so that's where they naturally go.

[00:25:46] So now you're going to tell them that, oh, this is a phone and this is how I get my information. So there is some sensitization there. Even though it meets a need, we still need to do some sensitization. So I think that that's another big challenge. And those are expensive. If you go and do campaigns or you do market activations, talking to people about it, that process can be expensive. In Kenya, we are working with Safaricom, which helps Digifarm.

[00:26:14] So they are able to do SMS plus because they are telco. So that's helping. And that's also allowing us to push the message there. And yeah, we've talked about the smartphone issue and internet issues, which are infrastructure issues. But hopefully those would also improve. You highlighted language and also content bias as barriers.

[00:26:34] So what is the process or approach that you use to balance technological innovation with a human-centered approach to the work that you do? So the human-centered approach is core to the process, you know, in many ways.

[00:26:52] Because without understanding the person that you are trying to solve for and the real challenges that you are facing would end up just producing, you know, nice toys, nice solutions that we talk about. And it's nice to talk about. It's fascinating, but nobody will use it. So the human-centered design is so key. In fact, in Opportunity, we have what we call the Idea Day.

[00:27:20] We just had one a couple of weeks ago where we bring all our staff together, all our staff together across the world. So some virtually, some in our offices. And we have this day where we get people to just dream about solutions for specific personas. So it's very focused. So we see this is Faustina, who is a smallholder farmer. This is Jacqueline, who is a teacher. This is Moses, that is a headmaster.

[00:27:47] This is Miriam, who is a micro-entrepreneur. We do extensive work to understand their life, you know, understand the way they live their life and what their challenges are. And then most of our staff actually work in the field anyway. So they understand some of these challenges. So what they get to do is to spend their day thinking about what a solution might look like on some of the challenges that they themselves would have also seen by going into the field.

[00:28:15] So we are bringing our expertise in the field to come and start thinking about solutions. And all these solutions that we had. So last year we did this, we had about 200 plus ideas. We brought it all the way down and selected the top three ideas, which we are developing now. We've done the same for this year. So that and that's what kicks the process. So that the solutions that we are getting, they are real solutions.

[00:28:43] They meet the needs of the people. And then once we select them, then we go into the field and we start talking to people and we start building it out. And that's how we managed to even build out Ilongizi in the way that it is, you know. And then once we have looked at the challenge and we've understood what the challenge is, then we can see how technology can solve that problem. Because if it's a technology, we don't want to force fit, you know.

[00:29:09] If technology can't solve that problem, then the technology that we know now cannot solve it, then that's fine. You know, we wouldn't force fit. But we would just make sure that the technology that exists now can actually solve this problem that we have seen. Then we will build it. So it is a balance. But I would say that the driving force is that human-centered design. And that to me is so vital. Because otherwise we will be letting people down.

[00:29:39] We'll create lots of solutions. But you see that the usage will be very poor. We always say that the technology exists, you know, but only that it's not evenly distributed. And that's a quote by William Gibbons that we use quite a lot. The technologies, they've been very interesting technologies for years out there. But somehow the underserved and the poor have been left out.

[00:30:03] Because nobody is thinking about how I take the solution that exists and apply it to a situation that a raw farmer somewhere in Africa is actually facing. Very true. I mean, all of the processes, technologies, solutions and innovations, they sound great. But there is obviously a cost. How do you ensure the financial sustainability of these digital programs, solutions that you are providing?

[00:30:31] That is a really, really good question. But also a big loaded one. So we are a charity. We are a non-governmental organization. So we get funding from various sources. So that's how we fund the building of these or developing of these solutions. But the idea is that once that we build them, we put them out there.

[00:30:55] Most of them are, you know, the codes and the technology, put them on open source so that other people can take the solution and then sort of grow it. But there is a cost for scale. And that's why partnerships are so key. So, for example, if you think about the agricultural solution that we are providing, if we are able to partner with, let's say, a government. Remember, the government also has extension offices.

[00:31:19] They have to provide information to the extension offices for them to go and help rural farmers. Now, what they might do is that they will have to bring all the farmers together and their extension offices together, train them. There's a cost to all that training. They have to make sure that any information they have is disseminated to all these extension offices. And they will do that in books or, you know, they will have to provide written material.

[00:31:44] In fact, one of the farmers that we spoke to was saying that extension offices, sorry, what they said was back in the day, without this solution, they would have to carry so many tube heavy documents into the field so that if a farmer asks them a question and they don't know or they are not sure, they can open this file and then search for the answer. That takes a lot of time. They have to carry this book. They get tired. You know, if they have to go into the rural area on a bicycle, can you imagine carrying these books?

[00:32:14] I swear to even put them is an issue. They get rained on, they get destroyed, and it means they have to be reprinted. But now everything is on his phone, you know, which he's carrying in his pocket. So it's ever so easy. So there's a lot of costs that has been saved. So what we need to do in collaboration is to see how the cost savings that we are making can actually be translated to the payment of scaling of this, of this, these kinds of solutions.

[00:32:42] So that's how we are, that's how we are sort of positioning it. And that's why, again, partnerships and collaborations are important to see some of it to help develop solutions, but some of it to see how we scale it and where we position it. Who will pick up some of these costs? But for some of them, it makes sense for them to do so because it's helping them actually save money. So it may not be new money that is bringing to the table.

[00:33:08] It's just, you know, shifting their money from one place to the other. It's not taking away jobs because the extension offices will still need to be employed. So it's not replacing jobs as other people would think. It's just making them more efficient. As you mentioned, you are a charity and you also touched on the importance of partnerships. Can you talk us through, I guess, the role of corporate partners like Cisco and MasterCard

[00:33:37] in terms of enabling you to carry out this work and funding and supporting these initiatives? Yeah, our corporate funders have been really fantastic. You know, so people like Cisco, as you mentioned, and also the MasterCard Foundation, and there are many others out there. So they bet on our innovation. Okay, so they look at what we are doing. They believe in it. And then they give us the amount of money to be able to do this. And these are grants.

[00:34:06] So it's not like an investment that they are looking for their money back. And that's really key to have people who allow you to test these solutions, you know, and to try it out and to improve them. So there is no sort of failure here. You know, we continue innovating. Something doesn't work, we shift it. We change it. We modify it. So I think those kinds of funders are so important.

[00:34:32] And I think they've been, you know, we wouldn't have been where we are without our partners or our funders, our donors who give us these corporate grants, because they've allowed us to go this far. And they continue to support us. They continue to give us money to try new things, you know. So the Ilongizi, as we have now, there are new things that we are going to add going forward, like weather, you know. So how can they get the weather information?

[00:35:01] So the journey has only just started in many ways. And we do need such support to help us build and improve the use case going forward. Keeping on the theme of partners, you slightly touched on this earlier. I was wondering, how would you work with governments and local institutions to ensure AI adoption actually aligns with their national agricultural goals?

[00:35:27] Yeah, so in all our deployments, especially in Ghana and Malawi, we did have extensive engagement with the government, especially in Malawi where we started. We made sure that they understood an extension officer. Providing extension services is core to, I mean, every country that we've been to, that's really one of the core services of the Ministry of Agriculture. Because that's the only way you can make sure that, you know, the information that you have is being passed down to farmers.

[00:35:57] You know, so the extension services are always, it's an expensive way of doing it, but it's one thing that all of them want. And that's where we align. So we spend a lot of time talking to them. In fact, in Malawi, the head of the extension services for the Lelonga area was really part of the pilot. He was really monitoring and making sure that the information that is given out. So when somebody asks a question on the bot, the information that is given is actually accurate

[00:36:26] and actually alliance with what they want, you know, what they are thinking. So they have been really, really involved. The same in Ghana. We just went to see the Ministry of Finance, the Ministry of Agriculture, and they were also very keen to see that. So working with governments is very key, you know, and in all our deployments, they have always been keen to see what we are doing, keen to join the pilot,

[00:36:55] and keen to help us to monitor it. So I would say that, you know, as I said before, agriculture, especially, and culture, health, education, these are core areas for any government. And it has to be also protected. So you can't just allow anybody to come in and start providing just any information to farmers because the information is wrong. That's your crop gone. Yes. You know, 70% of the people involved providing almost 80% of the food that we eat,

[00:37:24] you know, little bits of food. So you can't really allow just any information. So we always make sure that they are online. And we encourage more governments to actually take a stake in some of these works, not just for us, but, you know, anybody that's engaged in these sectors. If we move from, I guess, the alignment with current and upcoming government goals and look at upcoming AI projects at Opportunity International,

[00:37:51] are there any upcoming projects that you're excited about? Yes. So as I said, we did our idea days. We had tons and tons of ideas that have come that we are going to work on. But one of the others that I'm really excited about is not in our greatest education, where we have an education program where we have a similar app that works with teachers to help them do lesson plans.

[00:38:18] You know, so if you've ever taught, you'll see that doing lesson plans is a really big challenge, you know, to do a lesson plan for every lesson. And also to make sure that that lesson plan aligns with the needs of your pupils in the school. So we also, again, went to the government in Ghana. We looked at their curriculum and we are in the process of putting it into a bot so that the teachers can easily go and prepare a lesson plan from that.

[00:38:47] So you can say that I'm a teacher. I mean, I'm teaching class four or year four. I want a lesson plan to teach ratios. And these are the capabilities of the, these are the number of people in my class. These are the capabilities of the people. Answer a questionnaire and then you click and it will give you a full lesson plan. Based on the national curriculum. So then again, not just going anywhere and picking any information, but based on that. So you can see how transformational that could be for a teacher in a rural area, for example,

[00:39:17] or any teacher who needs to prepare lesson plans quickly. So that's one solution that I'm quite excited about. And we are about to put that together and to pilot that. I'm assuming that your upcoming projects align with trends that you're seeing out on the market, on the ground. Are there any emerging trends in AI that are helping, say, address challenges in rural communities that you're currently excited about?

[00:39:46] Yeah, I think one, the languages are getting better. But video creation, you know, is really, for me, it's amazing how, one, you can capture pictures, you know, and analyze pictures and images. For me, that's really key. I think that's really going to be transformational because you can imagine that for our great, but you can also imagine that for all sorts of other sectors, the fact that you can take an image of something and it can analyze and analyze it for you. Now, video creation soon could be interesting.

[00:40:15] So, you know, especially as we are doing training, you know, we create a lot of, back in, you know, back in the day without all this AI, we would have to go and create videos at an expense that is, you know, all that is, and you are in production, so you would know. But now with AI, we can actually customize it to a particular set of people, you know,

[00:40:38] so we can actually use that to improve training delivery because we can create content a lot faster, a lot cheaper, you know, and we can customize it. So these, for me, these are the trends that are coming up, but I find that, you know, it could be really interesting, can really change the way we deliver services to the underserved. Thank you for that. If we move from current trends and look ahead, where do you see AI use and its accessibility

[00:41:07] in Africa's underserved communities in the next five years? Yeah. So in the next five years, I think that one in agric, I think it would make a lot of, a huge impact in the agric sector, even beyond what we are talking about in terms of, you know, agriculture and advisory, you know, because now AI can analyze a lot of data very quickly,

[00:41:36] very simple forms. So all the data that we collect, we can actually analyze them very quickly and come up with solutions. So, you know, whether it's disease, whether it's weather, you know, all these things, I believe that, you know, and also in risk, you know, and so I think that, that in the future, well, I'm hoping that it would have a huge, significant impact in the agric sector.

[00:42:00] Education, as well, as we said, you know, lesson plans, but we also have another solution that is looking at our school management as a whole. So I'm expecting to see a lot of impact in that area as well in the next five years in that area. And obviously health already, we are seeing a lot of work being done in AI, in health, being able to analyze diseases and health issues, especially in

[00:42:25] countries where we don't have a huge, you know, have a lot of doctors and nurses, you know, and that AI can help play a role. We'll do everything, but at least it will help us. So I believe that if we use AI properly in Africa, if we don't allow it to pass us by, we can see a lot of growth using AI on the

[00:42:49] continent in the next five years. And if we look closer to home and revisit this conversation, say in five to 10 years from now, what do you hope to have achieved with the AI driven development work you're doing? Yeah. So for me, one is increasing productivity. It's the first thing we are really hoping that, you know, in all these sectors, we would see an increase in productivity. But we are also

[00:43:17] hoping, betting that Africans, you know, especially, I mean, Africa is the youngest, we have the youngest population in the world next 10, 15 years who have a lot of young people. So we really need to take a leading role in the knowledge aspect. Yeah. So we might have, we might have the tools, but how are we also applying ourselves? So that's why we are so keen to get our staff in our own small way, to get

[00:43:41] our staff to understand how to use AI, you know, to provide use cases. It's not just using it for me, you know, and, and using it was, how can I understand this technology to be able to solve problems for other people? Because that's the way we need to think if we are going to make a difference. So we need to upscale our people. We need to get people to understand this technology. You know,

[00:44:06] we need to be developers. So not just users, but actually using this technology to improve the lives of people. Because, you know, for me, change can only come from within, you know, especially on the continent. Nobody is going to come in and change things for us. As Africans, if we don't grab this technology and start thinking about how we can use this technology to solve solutions in our own backyard, which we understand, then nobody will do that because

[00:44:35] nobody understands our backyard more than us. So that's really what, that's the change that I'm hoping to see, especially the young people to start thinking about how we can use this to solve problems, not for themselves, to create videos and just put it on TikTok, but how can I do this to like solutions, solutions that I'm seeing around me? So to me, that's the real change. Because if that

[00:45:00] happens, then you will see that the technology will be applied to a lot of the challenges that we are facing. And there are many, you know, there are many challenges, are there big, small. And if we have people who understand how to use the technology, then they will build these solutions. They will build these apps. They will build, it doesn't build from anywhere else. They can build it locally because they understand the situation. So I think for me in the next, you know, few years, I'm hoping that the talent,

[00:45:30] you know, would embrace this technology, but with the mindset that this can actually help improve things, you know, and I'm hoping that that would really be a game changer for us. And I always say that, you know, we had the oil boom. Africa had oil, you know, but somehow it didn't, it didn't turn into the rapid economic growth that we expected it to be. AI is being touted as a new

[00:45:56] thing and everybody's investing in it. Everybody thinks it is the next big thing, which I believe. But then where does Africa fit in this? Are we just going to provide the minerals to support the devices? Or are we actually going to take a big chunk of the development and the knowledge to use it to improve our own livelihoods? Quote of the week. As people, we often have quotes, mantras, African proverbs, or affirmations that keep us going when times are challenging or when times

[00:46:25] are good. Do you have one that you can share with us today? Yeah, yeah. I think as I said in the beginning, because today is Ghana's independence, 68 years. Yes. And as you know, Dr. Kwame Nkoma, who was the first person, he was all about the strength of the unity of Africa. So my proverb, Ghanian proverb is, which means that when a single tree faces the wind, it breaks, you know,

[00:46:55] and there's strength in unity and there's standing alone makes one vulnerable. And I think for me, that speaks to what we are trying to do in terms of what we've talked about, collaboration and partnership, but it also talks about what I feel the continent needs. Perfect, perfect, perfect. Thank you for sharing that. And thank you for joining us on the podcast today. It's been an incredible conversation. I've enjoyed hearing your views on how AI can be used to

[00:47:24] drive real meaningful change on the continent for smallholder farmers and people in underserved communities. And also looking forward to seeing how those innovations continue to grow and impact those who use them. So thank you, Dr. Durbin. Thank you very much, Tessa. Thank you to all your listeners as well. Thank you to everyone who has listened and stayed tuned to the podcast. If you've

[00:47:52] enjoyed this episode, please subscribe, share or tell a friend about it. You can also rate, review us in Apple podcast or wherever you download your podcast. Thank you and see you next week for the Unlocking Africa podcast.