OpenIDEO is an open innovation platform. Join our global community to solve big challenges for social good. Sign Up / Login or Learn more –big data designed for small farmers

Using the latest techniques in Big Data we are building the world's first market matching chatbot

Photo of Georgia Barrie
44 24

Written by


When it comes to harvest time many farmers start looking for a buyer for their produce. It can take weeks - farmers message their friends and family to ask if they know anyone, they visit the local market and talk to brokers. Unfortunately all too often produce is sold at below market prices or not sold at all. At our mission is to create the most simple and scalable way to connect farmers and buyers. That's why we built the world's first market matching chatbot. Messaging our bot is just like messaging a friend. The bot can understand what the farmer is selling, where and their harvest date. It can interpret spelling errors, speak in multiple languages and help guide the farmer through the process. However, unlike a friend our bot searches a database of thousands of buyers to find the best selling options for each individual farmer. In less than a minute farmers can go from a simple "hello" or "jambo" to receiving the phone numbers of trusted buyers looking for their crop in their area. All of this we can provide for free to farmers as our business model charges fees only to large buyers. We have been overwhelmed by the level of demand for this service. We are still in Beta, testing and refining our product with an initial group of users. However, through word of mouth alone, hundreds of farmers have contacted us asking to try our service. This means we could reach thousands of farmers not just in the first year but the first month.


We are creating a mass-market product with the opportunity for significant scale. Our product is built to work through the major messaging platforms such as WhatsApp and Facebook which are already pervasive in Kenya. Although smartphone ownership among farmers is currently below 50% this figure is growing at a phenomenal rate. We believe in getting out in front of the market, not designing for technology that is already on its way out.


The solution will be implemented in Kenya initially due to its high smartphone and mobile financial service usage. However, the solution could be rapidly implemented in new markets not just across Africa but across developing markets worldwide.


  • Yes


  • I’ve worked in a sector related to my idea for over a year


  • Yes, for more than one year.

TELL US MORE ABOUT YOU! is a social enterprise founded by Adam Wills and Georgia Barrie who were excited to apply the latest tech to some of the oldest problems in ag. Through this journey we've met some fantastic farmers such as Noah Nasiali who is now working with us to help design the product and marketing.


For the past year we have lived and worked out of East Africa with the aim of designing the most simple, scalable solution to market access (we've started a blog to collate some of our research here: We've designed and tested dozens of ideas: from SMS and USSD based solutions, to apps and websites. Each time we've taken our idea to farmers, carefully listened to their feedback, packed up our things and gone back to the drawing board. It's been a real labour of love! This IDEO Challenge comes at a perfect time for us. In March this year we finally landed on an idea that we thought had real potential. The solution was built on feedback from hundreds of farmers and you can see their thoughts in the affinity analysis we created above. Through the Beneficiary Feedback and Improve stages we further tested the idea, built a working prototype and shared it with our initial pilot group. We're now excited to finally be taking a prototype to full launch. By analysing user feedback and usage data we will continue to hone the user experience as we scale up the product and we would be delighted to have the support of the IDEO team on this journey.


Many marketplace services have been designed to work over SMS or USSD but these services suffer from poor user experiences. Farmers struggle to remember short codes or how to follow multiple non-intuitive steps. Services like this have only worked at scale when they've had vast agent networks to train and support users. This makes products expensive to use and slow to scale up. By contrast, services such as WhatsApp have seen explosive growth in Kenya without a single person on the ground. We want to follow in their footsteps in 3 ways: 1. A simple process. The value proposition "match" is simple to understand and the entire process takes less than a minute from start to finish 2. An intuitive interface. Chatbots are the most intuitive interface out there - talking to is no more complicated than talking to a friend 3. Virality. WhatsApp's rapid growth was driven through word of mouth. We are designing simple ways and incentives for people to refer their friends and family. Our test user group was created purely through word of mouth and spans the breadth of the country. This would never have been possible if we relied on agents to sign up farmers.


Our small team of tech and farming experts is based in Kenya and manages everything from data analysis to programming the bot in-house. We want to be as agile and lean as possible so we don't believe in outsourcing any of our product development. To boost our customer acquisition rate we are currently negotiating partnerships with some of the largest inputs organisations in Kenya to use their distribution channels (e.g. posters in agrovet shops) to raise awareness of


Beneficiary feedback showed that building our service as an app would see barriers to usage. This is due to cost of data, tech literacy required to download and use apps and lack of memory on low-cost smartphones. By contrast it is very easy to engage farmers over messaging services such as WhatsApp as most smartphone-enabled farmers are already using these services. Chatbots are expected to be the next revolution in mobile UI and are incredibly well suited to a market connection service. We therefore decided to build the service as a chatbot over a platforms such as WhatsApp or Telegram. Beneficiary feedback showed that ratings are very important to both farmers and buyers but they need more than a simple star rating. We are kicking off a project to investigate the methods used by best practice services and to conduct detailed interviews with farmers and buyers to understand the right rating metrics. First impressions are incredibly important. We will work alongside farmer advisors to get the content and tone of the bot conversation right to ensure that first time users are able to immediately understand and trust our service.


There are lots of questions still to answer and we plan to build our solution in an agile way, constantly refining and improving it. We are lucky to be working with some fantastic digital farmers as well as a huge amount of data to help us hone our product. We have 3 big questions for the next phase: ○ Awareness - how can we create virality by designing easy ways to share our bot through social media and incentivising referrals? ○ Understanding - how can we analyse usage data from the test group to develop the most simple and compelling bot conversation? ○ Trust - what can we learn from established online marketplaces about designing the perfect ratings system for farmers and buyers?


This idea is emerging at a critical point in time: ○ Smartphones are finally reaching the mass market with 1 in 2 Kenyan adults using mobile data every month. The shift away from basic phones opens up a whole world of tech solutions to market access ○ Farming Facebook groups are taking off at a rapid pace in Kenya. The vast amount of data contained in these groups have helped us build a critical mass of buyers and sellers ○ New developments in areas such as data mining, natural language processing and bot technology means our idea wouldn’t have been technologically possible even a year ago. This service really is the first of its kind for small-scale agriculture


We see as the ultimate farming virtual assistant. Our simple bot interface together with our huge database of who is farming what and where opens up so many opportunities such as: ○ Connecting farmers to loans and insurance based on their selling history and ratings ○ Advising farmers on market prices and what to grow when based on supply/demand analysis ○ Creation of virtual farmer groups to aggregate produce across farmers selling the same crop at the same time in the same area


  • More than a year


  • Within 50 km of where our team does most of its work


  • We didn't have an operating budget

We set up as a small team. After spending years advising commercial & development organisations on mobile technology solutions we saw big unrealised opportunities for mobile solutions to have an impact in agriculture. We also know that this is a crowded space: too much wheel re-invention is happening and important lessons from pilot projects is being lost.

To address this, for the past year we have lived and worked out of East Africa to really understand the opportunities for technology in agriculture. We have seen the potential for truly innovative solutions to the problem of market access first hand. We have designed, developed and tested dozens of prototypes; used the very latest tools in big data and human-centred design to develop innovative solutions; and – most importantly – spent all our time in-market experimenting and getting real product feedback from farmers and buyers. We are building on what has been learnt from and we are responding directly to services that farmers themselves are asking for - not deciding what they ought to want.

Based on our on-the-ground research we have developed three core design principles:

  • We don’t believe in SMS-first solutions for farmers: 60% of all handsets sold in Kenya last year were smartphones and 1 in 2 Kenyan adults used mobile data in the last 30 days. Smartphone enabled and savvy farmers are the rapidly growing, early adopting segment that we need to design for.
  • We must observe and build on existing digital behaviours: 100,000s of farmers and buyers are posting on digital forums (Facebook, WhatsApp groups, classified ads websites etc). Facebook is currently the most successful digital agricultural trading platform in Kenya by volume and growth. In just one farmer Facebook group alone there are almost 40,000 members, growing by over 1,000 members per month. We believe that a successful solution in this space must build on what’s already working and not try to force behaviour change.
  • We must respond to farmer stated problems and design suggestions: digitally savvy farmers can articulate the issue better than we can, and design interfaces for themselves better than we can. We don’t even have to ask farmers, online farmer forums are full of problem statements and design suggestions that we can learn from. This isn’t where a solution should be tested, but rather, incepted.

So what are farmers actually saying?

Although Facebook is currently the most popular online platform for buying and selling produce, many farmers are vocal about the fact that it is not an ideal solution. As demonstrated by this Facebook post, farmers posting produce for sale often don’t find the right buyer – even when they have what “should be an automatic disposable product” such as milk.

Another farmer agrees, yet this demonstrates one of the key challenges of these large Facebook message feeds – information is getting lost in the noise.

Image title

Another key limitation of these message feeds is that information is not recorded in a useful way. Many farmers have posted what they are growing and where they are located, but this information is getting buried in the noise.

Image title

But the most consistently raised challenge is around trust, this farmer speaks for many when she asks whether a list of trusted buyers could be generated.

Image title

The real farmer quotes above are not one-offs, these same issues are being posted again and again everyday!

These farmer voices pose technologists a question

How can we build a better digital solution, accessible to all farmers and buyers who use Facebook and other online tools like this (as well as those who could, but don’t yet)? A solution that means the hundreds of posts per day from farmers and buyers don’t get lost in the noise? A solution where there is relevant profile data on farmers and buyers? And a solution which gives farmers the ability to exclude bad brokers, listing ‘good brokers’ and buyers only?

Where should we start?

Too many solutions try and re-invent the wheel. Rather than starting from scratch, how can we use the millions of public digital data points online to build solutions? How can we connect to existing platforms and aggregate live streams of information already being generated online?


Join the conversation:

Photo of David Meek Jah

Hi, My name is David Meek Jah, i would want to know if this is going to be an app and if yes? How do we collaborate to ensure we integrate your app on our open source payment platform that takes all source of payment in the world which is like a smart phone that has an inbuilt catalogue and receipt system via mail, SMS, or paper receipt. We believe our deep pocket data italics on products, sales, and settlement which give industrial data will also help farmers get loan faster as it can be synchronized with quick books or once payments are done via mobile money financial institutions could use that information via our inbuilt report system that can be mailed to them daily and automated to give them feedback on each farmers financial status and business profile. Kindly let me what you think and we see how we could do business. You can contact me on +23225224656 or mail me (

View all comments