There are many answers, and many of them are complex, but across my work in food systems over the last 10 years, I keep seeing the same pattern: agribusiness transformation is never really about the crop. It's about the system around the farmer — technology, design, data, partnerships, and financing. Get those five things right, and build them around farmers, and sectors can be positioned to really take off, moving from local markets to global significance. Get them wrong, and you can work for decades and remain exactly where you started. 

What Rwanda's coffee tells us about Nepal's tea 

Start with a country that got the system right. In 2002, only about 1% of Rwanda's coffee exports were fully-washed specialty coffee — the grade that commands premium prices. The rest was ordinary, semi-washed coffee sold into commodity markets, where someone else set the price and captured the value. Structurally, it's the same position much of Nepal's tea sits in today when it flows across the border. 

By 2007, fully washed specialty coffee was 21% of Rwanda's exports. Today it's roughly half. An entire sector moved up the value chain in a generation, and none of what changed turned out to be a miracle. 

First, policy. Following market liberalization and the 2002 National Coffee Strategy, the government removed barriers to private investment, and investors put money into coffee washing stations — the processing infrastructure that turns ordinary coffee into specialty coffee. 

Second, shared infrastructure. More than 250 washing stations went up. A smallholder with a hectare of coffee could never afford her own processing facility now but she didn't have to. The station handled sorting, processing, and cleaning to a consistent standard; her job was to grow excellent coffee cherries and deliver them to the station. Cooperatives, actively encouraged by the state, let smallholders pool resources and reach economies of scale. 

Third, those washing stations became hubs — places where farmers gathered, compared techniques, and supported one another, and where government and buyers could deploy training and inputs. The infrastructure created the community, and the community sustained the quality. 

Now look at eastern Nepal through that lens. Ilam's orthodox tea already commands premiums abroad. Nepal is also the world's leading producer of large cardamom, most of which leaves the country with almost no processing and almost no brand. It’s a similar story with ginger. The raw material for a Rwanda-style transformation is sitting in the hills of Nepal. What Rwanda teaches is that the gap between 1% and 50% isn't closed by farmers working harder. It's closed by building the system around them in ways that support the local context.  

The reliability problem 

Premium markets don't just pay for quality. They pay for consistent quality, at reliable volume, season after season. A buyer in Tokyo or Hamburg will forgive many things, but not surprise. And consistency is exactly where structural constraints bite hardest. Three stand out in eastern Nepal. 

Labor. Tea is labor-intensive — labor runs 45–60% of production costs — and the labor market in Nepal is under real strain as working-aged men often go abroad for stable incomes. When labor is short, harvest gets delayed; when tea is plucked late, leaf quality drops and so does the price.  

There's a remarkable dimension in Nepal, though: around 90% of the tea workforce is now locally trained women, and research shows their involvement in household decision-making has measurably increased. That matters for production, not just for the women themselves. When women gain more say over household resources, they reinvest in ways that show up directly in the fields: better inputs, steadier labor, more consistent care of the crop. It's a fact that has been proven out around the world: farms are more productive when women are more involved in both the work and the decisions that follow from it.  It was the core premise of a partnership I built with PepsiCo several years ago that spanned six countries, and research from that project proved the point. So in Nepal the labor constraint has become an opportunity: a story about who should be at the center of how the industry designs the systems of support that will define its future. This is a particularly important aspect during this UN Year of the Woman Farmer. 

Climate and disease. Nepal's agriculture sector is highly vulnerable to climate change, and tea is already feeling it: reduced water content in the leaf, sun scorch, extreme weather that damages quality and destroys crops outright. Warmer temperatures mean pests reproduce faster and spread further. Fungal disease can take up to 85% of a yield; viral disease can take all of it. But buried in the research on Jhapa and Ilam is a finding worth repeating: the farmers who adapted best to climate change were not the ones with the best land or the most experience. They were the ones with credit access, frequent training, and cooperative membership. The best climate adaptation tools we know are loans, peer-to-peer teaching, and community dialogue. Those aren't agronomy interventions. They're system interventions. 

Post-harvest loss. Nepal grows an enormous volume of fruits and vegetables, then loses 20-50% of it after harvest to poor handling, missing processing technology, inadequate storage, weak cold chains, and thin market linkages. This isn't unique to Nepal; my firm saw the same dynamic studying stone fruit loss on the U.S. West Coast. 

Notice what these three constraints have in common: none of them originate with the farmer. Farmers didn't design the labor market, the climate, or the cold chain. These are system failures, and they need system answers — not things we hand to farmers and walk away from. 

The solution glut 

Farmers are not short of people trying to help them. There are hundreds of NGOs, agtech startups, apps, and platforms hoping to serve farmers. Yet, to a farmer with limited capacity to add steps to an already full workday, much of it registers as noise. 

We saw this firsthand while running grower workshops for an AI-based agronomy tool we helped build with PepsiCo. Our biggest finding was digital fatigue. Growers were being asked to manage hundreds of data points scattered across disconnected platforms. Agronomists were spending more time training farmers on software than working with them on crop health. And when a farmer can't tell which app to open for which problem, the damage compounds: they don't just lose confidence in that app. They lose confidence in digital tools as a whole. Every poorly designed tool poisons the well for the next one. 

Sometimes the barrier is even more basic: language. A pilot by Opportunity International in Malawi called Ulangizi AI — an advisory chatbot for smallholders built on Ministry of Agriculture data and designed with a WhatsApp front end— found that among farmers using it in English, 41% would recommend it. Among farmers using it in Chichewa, their own language, the figure was 75%. Nothing about the technology changed. The tool was transformed by respect for local conditions.  

Apply that to eastern Nepal, where Nepali, Maithili, Tamang, Bantawa, and many other languages are spoken, and the lesson is clear. The question isn't "is the technology good?" The question is "was this built for the person who has to use it?" 

I saw a great example of a similar, localized solution here during the Code for Impact pitch competition finals in Kathmandu. One team — made up of farmers from Madhesh— built an app that was designed to use AI to help local farmers with agricultural questions.  

Build for the farmer and they will adopt 

Human-centered design is a simple commitment with radical consequences: you design with the people who will use the solution, not from a conference room or a set of far-off assumptions. You start with their lives — how they make money, what language they speak, where their trust networks live — and let the solution take whatever shape those realities demand. 

In our PepsiCo work, we facilitated user workshops, sat for a dozen one-on-one grower interviews, and walked fields in the U.S. and Europe. Two insights changed the entire product. 

First, ROI must be visible before behavior changes. Growers adopt when they can see greater yield, fewer diseases, higher profit, or time saved. As one grower told us, he makes dozens of decisions every day, and any new technology has to show evidence that it makes those decisions easier. Second, meet users where they are, which translated into local language support, integration with familiar tools like WhatsApp, and functionality in low-connectivity environments. 

At the other end of the scale, consider DeHaat, one of South Asia's most successful agri-platforms. For its first several years, the founding team didn't build much of anything; they immersed themselves in farmers' daily lives and discovered the real enemy wasn't a missing app; it was fragmentation. Their answer was "phygital": more than 11,000 physical centers run by local micro-entrepreneurs, because trust travels through familiar faces; advisory in more than thirty regional languages; and — maybe the most important design choice of all — a business model in which advisory is free and DeHaat earns only when the farmer sells his goods. The company cannot succeed unless the farmer does. 

For eastern Nepal, the raw material for this already exists: cooperatives, collection centers, established pathways to market. The trusted local layer is here. The design task is to build onto it, not around it. 

Data, partnerships, and the last piece: catalytic capital 

Once farmers begin using a digital system that is designed for them, it opens up access to the most underrated asset in this sector: evidence. Data protects quality, shows you where to improve, and becomes the credential that persuades partners and financiers to show up. But it will only work if it delivers value to the farmer first, and only if collecting it is nearly effortless. A ten-dollar moisture sensor tells a cardamom farmer exactly when curing is done, the difference between premium grade and a charred discount. Data is the key that opens doors: One Acre Fund, a global finance NGO, gets meetings with commercial banks and development finance institutions that non-profits normally never get, because it built a documented, multi-year record of farmer repayment and impact. Data is how a handshake scales. 

No single organization transforms a sector alone, though. Nepal's cooperative architecture — roughly 30,000 primary cooperatives serving over 10 million members — is a structured entry point for exactly the kind of multi-stakeholder partnership that moved Rwanda.  

Honduras offers a proven template: the De Mi Tierra program, where supermarket chain La Colonia issued binding purchase contracts to farmer associations, Banco Ficohsa used those contracts to underwrite low-interest loans for greenhouses and irrigation, and the nonprofit FUNDER embedded technical training in the fields. It began with 16 potato farmers; today it reaches more than 1,200 producers across 100 communities and helped cut vegetable import dependency by over 86%. No massive aid influx. Just local actors aligning their commercial incentives. 

Which lands the argument on its most counterintuitive point. Conventional wisdom treats farmer welfare and profitability as a trade-off. The evidence says these markets are profitable precisely when — and only when — solutions are built around improving farmer livelihoods. Rwanda's washing stations work because farmers earn more from specialty coffee. DeHaat earns only when the farmer sells. Farmer prosperity isn't the cost of the business model. It is the business model. 

The Nepal Tea Collective has proven that a Nepali commodity can scale into a global brand. With our partners, we’re working on an ag system that can make this the rule, rather than the exception. The work ahead isn’t easy but ensuring that Nepali products have a place on the world stage will mean long-term livelihoods for millions of farmers and their communities.