Krawp
← Back to Blog

Case Study

August 5, 202611 min read

How one co-op cut diagnosis time by 70%

A field-led look at faster answers, safer treatment and better follow-up.

JK

James Kariuki

Partnerships Lead, Krawp

Cooperative farmers working together in a rice paddy

The challenge at AgriUnion Co-op

AgriUnion Co-op manages over 2,400 hectares of mixed cropping — rice, maize, tomatoes, and beans — spread across four districts in Kenya's Rift Valley and Western provinces. The cooperative serves 623 member farmers, most of whom cultivate between 1 and 5 hectares. It's a typical East African cooperative: strong community bonds, limited infrastructure, and a chronic shortage of extension officers relative to the area under cultivation.

Before Krawp, every crop diagnosis followed the same slow chain. A farmer noticed a problem — discoloured leaves, stunted growth, unusual spots — and walked to the nearest extension office, sometimes a round trip of 15 kilometres. They described the symptom verbally, or carried a physical sample if the distance allowed. The extension officer, if available, scheduled a field visit. If the officer was managing multiple districts, the visit could take 5–14 days to materialise.

The average time from symptom recognition to confirmed diagnosis was 9 days. In that window, diseases like rice blast, bacterial blight, and tomato mosaic virus could spread across entire plots. The co-op estimated losing 12–18% of its annual tomato yield — approximately $34,000 in revenue — to problems that could have been managed effectively if caught in the first 48–72 hours.

Rolling out Krawp across 180 farms

In January 2026, AgriUnion began distributing Krawp to 180 of its member farmers — roughly 30% of the membership, selected across all four districts to ensure geographic and crop diversity. The rollout was deliberately simple: farmers downloaded the app during regular cooperative meetings, received a 20-minute hands-on walkthrough, and were asked to photograph any unusual crop symptoms they encountered during routine fieldwork.

The cooperative's leadership understood that technology adoption depends on trust and habit formation, not technical sophistication. Rather than conducting extensive training sessions, they identified 12 'farmer champions' — respected members with strong field skills — and gave them deeper training. These champions then supported their neighbours informally, answering questions and encouraging consistent app usage.

The offline-first design was critical to the rollout's success. Many of AgriUnion's members farm in areas with intermittent or no mobile data coverage, particularly during the rainy season when infrastructure is most vulnerable. Krawp's ability to queue diagnoses locally — storing photos and results on the device — and sync when connectivity returned meant the app worked regardless of signal conditions. Farmers didn't need to walk to a hilltop or find a specific spot with signal to use the tool.

The numbers after six months

The improvement in treatment adherence was particularly significant. When farmers received a clear, actionable diagnosis on their phone — in Swahili or their local language, with specific product names, application rates, and timing guidance — they were far more likely to act on it. The diagnosis wasn't just information; it was a prescription that removed ambiguity from the next step.

  • Diagnosis time dropped from 9 days to 2.6 days — a 71% improvement.
  • Diagnostic accuracy rose from 64% (extension officers without lab support) to 89% with Krawp's AI-assisted identification.
  • Treatment adherence within 48 hours of diagnosis went from 38% to 72% — nearly doubling.
  • Tomato yield loss attributed to delayed diagnosis fell from an estimated 15% to 6% — a 60% reduction in preventable losses.
  • Total estimated revenue recovery across the 180 pilot farms: $18,400 per season.
  • Extension officers reported spending 40% less time on routine diagnostic visits, freeing capacity for agronomic advisory and farmer training.

Voices from the field

Joseph Kamau, AgriUnion's lead extension officer with 14 years of field experience, described the shift: "Before Krawp, we were guessing from verbal descriptions. A farmer would come in and say 'the leaves are going yellow' — but yellow how? All over or just the edges? Old leaves or new leaves? By the time we got to the field, the symptom description had passed through three people and lost most of its diagnostic value. Now we see the same photo the farmer sees, taken in the actual field conditions, and the app tells us what it thinks it is. We still make the final call — the app doesn't replace our judgement — but we start from a much better place."

Grace Wanjiku, a 2.3-hectare tomato farmer in Kericho district, shared her experience: "Last season I lost almost half my tomato block to blight because I didn't know what it was. I thought it was just old leaves dying. This season I photographed the first spot I saw, and Krawp told me it was blight and what to spray. I treated it in two days and saved the rest of the crop. That one diagnosis paid for the whole season's app usage."

The cooperative has since expanded its Krawp deployment to cover all 623 member farmers and is exploring how the aggregated field data can inform seasonal planning, input procurement negotiations, and crop insurance assessments with partner financial institutions.

What other cooperatives can learn

AgriUnion's experience offers several transferable lessons for cooperatives considering similar deployments:

  • Start with a pilot. 180 farms across diverse districts gave AgriUnion enough data to evaluate impact before committing to a full rollout.
  • Invest in farmer champions rather than classroom training. Peer-to-peer support drives adoption faster than formal instruction.
  • Offline-first isn't optional. If the app doesn't work without signal, farmers in remote areas won't use it consistently.
  • Make the data visible to members. Farmers who see their own field data — and how it compares to the cooperative average — become more engaged with crop management.
  • Use aggregate data for collective decisions. Individual diagnoses help individual farmers; aggregated data helps the entire cooperative negotiate better input prices, plan spray schedules, and assess risk.

Spray compliance within 48 hours of diagnosis went from 38% to 72% — and tomato yield loss from delayed diagnosis fell from 15% to 6%.

When diagnosis is fast, accurate, and delivered in the farmer's language, treatment adherence doubles. AgriUnion's 71% reduction in diagnosis time translated directly into $18,400 in recovered revenue per season across 180 pilot farms.

Frequently asked questions

How long did the full rollout take?

The initial pilot of 180 farms was completed in 3 weeks. The full rollout to all 623 members took an additional 6 weeks, with farmer champions supporting onboarding in each district.

Did farmers need smartphones?

Krawp runs on Android devices running version 8.0 or higher. Most of AgriUnion's members already owned compatible smartphones — the cooperative provided devices to the 12 farmer champions and to members who didn't have access to one.

What happened to the extension officers' role?

Extension officers shifted from spending most of their time on diagnostic visits to providing agronomic advisory, farmer training, and seasonal planning support. Their role became more strategic rather than reactive.