AI + IoT for precision agriculture in Africa ︎︎︎
Go-to-market, Startup, Agriculture, Climate change, AI, IoT, Africa, Sustainable development, Financial inclusion
AgOptimized’s first sensor pilot test.
Co-led and executed requirements gathering, user onboarding and testing, and go-to-market strategy for AgOptimized, an integration of microfinance, IoT, hardware, and neural networks to help smallholder farmers in Africa practice precision agriculture, securing yield amidst rapidly changing climate conditions.
Charlie Nichols (CTO & co-founder)
Andrew Denu (Product manager)
Hanif Wicaksono (Design lead)
Tom Zhou (Business lead)
Alex Kung’u (Engineer)
Sophie Mutheu (Agronomist)
Jimmy Gitau (Fixer)
Sunculture AgOptimized (launched as MITI)
Pilot result: 1000% revenue growth with 97% repayment rate. Funded by grants Microsoft and Shell Foundation, raising seed
AgOptimized system orchestration (illustrated by Microsoft AI for Earth)
Problem: climate change on farmers
96 percent of Africa’s smallholder farmers (who comprise 60 percent of the population) rely on rain instead of irrigation for farming. But Africa’s rainfall has declined more than 100mm annually since the 1970s, requiring farmers to be extremely precise with when and how they farm. To guide planting, fertilizing, and harvesting, they rely on traditional weather forecasts, which are typically too broad for the small land areas they are farming. As a result, their yields often fall well below the world average. Additionally, farmers typically overpay for household and farming supplies by 20-30%, affecting their ability to maximize profit and maintain their livelihoods.
AgOptimized is an integrated system with three main components:
AgOptimized uses cloud computing to compare sensor data about current conditions against historical climate models, then provides detailed forecasts and recommendations for each farm plot. These recommendations help farmers optimize planting, irrigation, fertilization, and pest control.
SunCulture’s AgOptimized collects soil, weather, and meteorogical satellite to build a hyperlocal database (knowledge specific to a site)
New technologies such as mobile banking and digital payments have brought financial services to millions of rural consumers across Africa, creating an opportunity to build on this momentum. Farmers can order products that are appropriate to predicted risks input.
My role: Lead designer
I co-led a research team with a business specialist using human-centred design (HCD) methodologies to help us understand the behavior, challenges, and needs of smallholder farmers in Kenya, specifically on knowledge gain + delivery, agronomy practices, and credit financing which are the main interests and touchpoints of AgOptimized (AO).
We also analyzed and mapped out potential opportunities and partners for the business development, go-to-market strategy, and design guidelines for AO product features & prioritization.
Project timeline and milestones.
Relationship map is one of our preliminary research milestones since it helps us paint a picture on upcoming findings and adjustments.
Assembling a team
Jan Chipchase famously stated in his design research practices: “you are only as good as your local team.” In field research, a mix of different expertise are needed to unlock hidden perspective and valuable insights since getting a full picture is really important.
This project was also my first time in Africa and I definitely need to plan carefully beforehand. To answer this need, at early stage of the project I talked with co-founder and CTO of Sunculture and recruit best people in the team that could do this project full-time, starting from our office in Nairobi.
Aside from my business partner for this project, we have an agronomist, Engineer, and our driver / translator. Camaraderie is our essential fuel; we bond together and travel across Kenya together for around 2 months using our four-wheel-drive truck.
Understand the farming ecosystem
We started by talking to the existing customer base of Sunculture’s solar pump and branched out the network to other stakeholders. By the end of this project, we gathered data from these places:
counties: Laikipia, Meru
- 3 locations:Nanyuki, Matanya,Timau, Meru
42 farmers interviewed +
10 group discussions
- 4 contract companies (outgrowers)
This project covered a long spectrum in product development and various methods and prototypings were used, both paper & digital.
One of the essential finance features of farmers is how they earn revenue.
Most of transactions are made in lump sum after harvest and before planting.
One of the generated farmer archetypes. All of these information would keep the product development and design aligned with the actual user in mind.
Understanding stakeholders in the system
We were interested in doing the research on these aspects for the development of AgOptimized plaftorm:
- Knowledge gain and learning
- Agronomy + market practices
- Credit + financing
We synthesized our findings into farmer archetypes, a classification based on their operational mindsets and characteristics. On each archetype, we put together their key findings, insights, channels, and trust circles. Since we were doing qualitative research, the data is granular and nuanced. For instance, here’s a snippet on data for the type of farmer whom we called the “Climber”:
Climber: Key Characteristics
Fully invested in gaining higher quality yield. Climbers are committed to crop production as a business and put in a considerable amount of effort. They are willing to take risks on farming and make their own input purchases. They also do their own market research to find buyers. At the end of the day, they want the best price in the market.
Dream to scale. Putting in very high effort to achieve better yields, climbers carefully think and plan their farming business. They know what they want in the next 2-3 years and the steps to achieve that ambition. However, other priorities, limited capital, resources, & market access act as barriers that prevent rapid scaling of their operations.
Hungry for knowledge. Climbers have access to agronomy support either through hired agronomists, trusted agrovets and/or online agronomy chat and information sites. They also conduct their own research on what crop does best in their respective area in terms of general management and availability / prices in the market.
Testing knowledge delivery: SMS, monitor, or encyclopedia?
UX framework and early UX prototypes
Aligning needs with design for AgOptimized
From our conversations, we propose to pilot the AgOptimized solution with contract farmers and outgrowers as the main users.
Through focus groups, interviews, and prototype testing, we further talked to outgrowers and contract farmers to understand how to design AgOptimized’s advisory information to each respective end users. Areas covered:
Initial appetite for advisory
- Knowledge delivery
We distill our findings into design opportunity areas with recommendations on how do we proceed and execute. For example:
Opportunity area/Design space for outgrowers
Actionable, instructional advisory: For mobile-based advisory, farmers find shorter, practical recommendations to be more useful vs. comprehensive guidelines that cover all topics in a crop harvest cycle. Farmers view AO advisory as a complementary solution, and will double check before making any decisions.
How to proceed:
Delivering brief, actionable, instructional advisory will help farmers on current/ upcoming problems. AO advisory pieces could serve to deliver “high immediate priority” info that enables them to act accordingly.
Findings on UX
We built a series of initial advisory prototypes and put together a proposed MVP model of mobile application UX for AO’ next iteration. We tested the card stack model and gathered feedback on usability, content, and information access.
Typical field day.
Testing different concepts and prototypes.
AgOptimized is now soft-launched as MITI: Financing the Next Billion.
Go-to-market: findings on financing
Even when farmers can access traditional bank loans, they have an extreme aversion to these products. The downside potential of failing to make payments can include land seizure.
Access to capital, however, is consistently reported by farmer as being one of their top challenges. There are simply few or no viable options for farmers to access credit and in many cases the costs of farm management and investment exceeds a farmer’s cash holdings.
From the research, we addressed these questions on financing:
“ Pay As You Go” Asset Financing
- Buy Now Pay Later (BNPL)
BNPL Big Ticket vs. Bundling
When asked to picked items from a catalog featuring both agriculture goods and consumer goods, farmers leaned heavily towards productive agriculture assets. Consumer goods also were chosen, and can be fit into 2 general (sometime overlapping) categories.
As part of AgOptimized, every farmer will receive a standard smartphone and data package on AO credit. This eliminates the technology access barrier to both marketplace and advisory. We tested two phone products with different quality and price points. The overwhelming preference was for phone 2. We tested payment schedule weights (equal installments or backloaded. The preference was for equal installments.
Product release: MITI
These research findings informed the first release of AgOptimized as MITI: MITI solution works by financing a high-quality smartphone preloaded with the MITI eCommerce application. Customers use the MITI app to shop for household appliances, electronics, foodstuffs and farm inputs from one of our marketplace partners, make payments in monthly installments and collect their purchases from nearby collection points.
MITI leverages smart partnerships to optimize operating expenses and customer acquisition costs, achieves high margins by financing a mix of products and generates revenue through both commission and finance charges.