[Design Early Warning System]
Helping Growers Stay Ahead by Identifying, Containing, and Mitigating Plant Threats
A Customizable Disease Early Warning System for Crop Health
Proactive Risk Management in Agriculture - Predict - Protect - Grow
[TL;DR]
DEWS (Disease Early Warning System) is a predictive tool powered by weather-based disease prediction technology to help agribusinesses and farmers anticipate crop diseases before they occur. I designed a flexible configuration system allowing users to customize disease warnings, select notification channels, and provide real-world feedback to improve prediction accuracy.
[NDA Disclaimer]
Due to the confidential nature of the project and adherence to a Non-Disclosure Agreement (NDA), certain details, designs, and proprietary information have been intentionally omitted or simplified. The content shared in this case study provides a high-level overview of the project, focusing on the problem-solving approach, design process, and outcomes without revealing sensitive or proprietary information.
While I have limited the use of detailed designs and specific data, the case study aims to highlight the key challenges, solutions, and impact of the project while respecting the confidentiality agreements in place.
[Summary]
Objective
Enable agribusinesses to proactively detect, contain, and mitigate crop diseases by providing a customizable early warning system powered by weather and satellite data.
The goal was to improve prediction accuracy, user engagement, and decision-making by allowing users to configure disease warnings, set risk thresholds, and select notification channels.
My Role
Led the design of the feature by collaborating closely with the product manager, engineers, data scientists, agronomists, and stakeholders throughout the project.
My contributions included user research, defining workflows, and designing UI elements, interactions, and micro-interactions to ensure an intuitive and effective experience.
What Was Achieved
Based on insights from user research, industry analysis, and internal data, the following improvements were made:
Customizable Warnings – Users could configure disease alerts based on crop type, location, risk thresholds, and prediction timelines.
Multi-Channel Notifications – Enabled In-app, SMS, Mail and WhatsApp warnings to ensure timely communication.
Feedback Loop – Introduced a mechanism for field users to confirm predictions or raise alerts, improving data validation.
Seamless Integration – Ensured smooth incorporation into Cropin Cloud (web) & Cropin Grow (mobile) , Cropin’s flagship product.
Impact
Proactive warnings reduced unexpected crop losses by allowing timely intervention.
Customizable configuration ensured relevant and non-intrusive notifications.
Feedback loop improved accuracy, making DEWS smarter (9.5%) with real-world validation.
Positive adoption from agribusinesses and Cropin cloud users (B2B SAAS platform), leading to higher engagement. Upscaled 70% of the strategic partners / customers.
[Behind The Solution]
Understanding User Needs & Research Insights
To design an effective early warning system, it was essential to understand the current disease prediction and management practices used by different stakeholders:
Growers & Farmers
often rely on visual inspection or take plant samples to local agronomists for disease identification. In some cases, they use precautionary chemical treatments without certainty, leading to unnecessary pesticide use.
Agronomists & Farm Managers
conduct regular field visits to monitor crop health. They sometimes detect diseases only after early symptoms appear, reducing the window for preventive action.
Business Stakeholders at Cropin
emphasized the importance of aligning the system with existing workflows, reducing manual effort, and ensuring seamless integration with Cropin cloud & Cropin Grow.
These insights highlighted the need for a proactive and data-driven disease prediction system that could alert users before visible symptoms appeared, helping them take timely action.
DEWS Diagram
The DEWS system flow integrates AI-driven disease prediction with user validation to ensure accurate alerts. It processes weather data to forecast risks, allowing users to configure alerts based on crop, location, and probability thresholds. Field users can review, verify, or refine predictions, creating a continuous feedback loop that enhances model accuracy over time.
Designing the DEWS Configuration Feature
Based on insights from user research and stakeholder suggestions, the DEWS configuration feature was designed to align with Cropin’s SmartFarm platform while addressing the specific needs of agribusinesses and field users. Here’s how it was done:
User Insights
Agribusinesses
needed flexibility to track specific crops and locations, set risk thresholds, and define how often they wanted updates.
Field Users
preferred simple, actionable alerts that could be accessed on-the-go via mobile devices.
Stakeholders
emphasized seamless integration with Cropin Cloud, ensuring the feature complemented existing workflows without adding complexity.
Key Features Designed
Project, Crop & Location Selection: Users could choose specific crops and geographic areas for disease tracking, ensuring relevance to their operations.
Model Run Frequency: Options for daily, weekly, or custom intervals allowed users to balance data freshness with operational needs.
Prediction Window: Users could define how far into the future they wanted forecasts (e.g., 7 days, 14 days), aligning with their planning cycles.
Risk Thresholds: Customizable probability thresholds (e.g., 50%, 70%, 90%) ensured users received alerts only for significant risks, reducing unnecessary notifications.
Notification Channels: Multi-channel delivery (In-app, SMS, WhatsApp) ensured alerts reached the right people, even in low-connectivity areas.
Recipient Management: Users could assign alerts to specific roles (farmers, agronomists, managers), ensuring the right stakeholders were informed.
Integration with Cropin Cloud
The configuration feature was seamlessly integrated into Cropin’s SmartFarm platform, leveraging its existing infrastructure and data models.
Designed to work harmoniously with other Cropin Cloud features, such as farm monitoring and analytics, ensuring a cohesive user experience.
Ensured scalability to handle large datasets and diverse user needs across different regions and crop types.
Designing the DEWS Feature in Cropin Cloud & Cropin Grow
The DEWS feature was designed to seamlessly integrate into both Cropin Cloud (web dashboard) and Cropin Grow (mobile app), ensuring a consistent and intuitive experience across platforms. Here’s how the feature was implemented:
Web Dashboard (Cropin Cloud)
DEWS Card on Home Screen
A DEWS Card was added to the web dashboard, displaying:
Total Number of Warnings: The overall count of active disease warnings.
Plots Affected: The number of plots impacted by disease risks.
Users could click on the card to view a detailed list of DEWS warnings.
Detailed Warnings View
Clicking the DEWS card opened a list view of all active warnings, including:
Crop type, location, disease risk level, and prediction timeline.
A summary of affected plots and recommended actions.
Users could click on individual warnings to view plot-specific details.
Plot Cards & Navigation
Each warning was linked to plot cards, providing a snapshot of the affected plot’s details.
Clicking a plot card navigated users to the plot page, where they could:
View all DEWS warnings specific to that plot.
Voiew action taken at the field through Cropin Grow app.
Actions taken on the web were synced in real-time with the mobile app, ensuring consistency across platforms.
Action Tracking
Users could view actions taken from the mobile app directly on the web dashboard.
A history of all actions (e.g., alerts raised, predictions validated) was maintained for future reference and analysis.
Mobile App (Cropin Grow)
DEWS Card on Home Screen
The mobile home screen featured a DEWS Card, showing:
Total Number of Warnings: Active disease warnings for the user’s assigned plots.
Plots Affected: The number of plots with disease risks.
Tapping the card opened a list of DEWS warnings.
Detailed Warnings View
Users could view a detailed list of warnings, including:
Crop type, location, disease risk level, and prediction timeline.
A summary of affected plots and recommended actions.
Tapping a warning opened plot-specific details.
Plot Cards & Navigation
Each warning was linked to plot cards, providing a quick overview of the affected plot.
Tapping a plot card navigated users to the plot page, where they could:
View all DEWS warnings for that plot.
Take action by either raising an alert or marking the prediction as “Looks Good.”
Actions taken on the mobile app were synced in real-time with the web dashboard.
Action Tracking
Users could view actions taken from the web dashboard directly on the mobile app.
A history of all actions (e.g., alerts raised, predictions validated) was maintained for future reference and analysis.
Key Design Considerations
Consistency Across Platforms
Ensured a seamless experience between the web dashboard and mobile app, with real-time syncing of data and actions.
Actionable Insights
Designed plot cards and warnings to provide clear, actionable insights, enabling users to make informed decisions quickly.
User-Friendly Navigation
Simplified navigation between warnings, plot cards, and plot pages to reduce friction and improve usability.
Real-Time Feedback
Enabled real-time feedback from field users, improving the accuracy and reliability of DEWS predictions over time.
[Testing & Iteration]
Conducted prototype testing with internal teams and pilot agribusiness users to validate usability and effectiveness.
Refined UI elements and workflows to make configuration faster and more intuitive.
Simplified feedback collection to encourage higher engagement and response rates from field users.
[Impact]
Proactive Alerts
Proactive warnings reduced unexpected crop losses by allowing timely intervention.
Customizable Configuration
Empowered users to define relevant alerts. Customisable configuration ensured relevant and non-intrusive notifications.
Feedback Loop
Improved model accuracy (by 9.5%) with real-world validation.
Increased Engagement
Positive adoption and usability feedback from agribusinesses and field teams. Upscaled 70% of the strategic partners / customers.
[Key Learnings]
User-Centric Design is Non-Negotiable
Always prioritise user needs and context, especially when designing for users with varying levels of technical expertise.
Flexibility Drives Adoption
Providing flexibility in features empowers users to tailor the system to their unique workflows, increasing engagement and satisfaction.
Feedback Loops Enhance System Intelligence
Incorporating real-world feedback into predictive models ensures continuous improvement and user confidence.
Seamless Integration is Key
Designing for integration with existing systems demands close collaboration with cross-functional teams and a deep understanding of technical limitations.
Simplicity is Powerful
Strive for simplicity in design, even when dealing with complex data or workflows.
Iterative Testing Drives Refinement
Continuous testing and iteration are essential to creating a polished, user-friendly solution.
Cross-Functional Collaboration is Essential
Collaboration across disciplines is critical to delivering a solution that is both functional and impactful.
Designing for Impact
Great design goes beyond aesthetics—it solves real problems and creates meaningful change.