Aniruddh's Design Portfolio

ANIRUDDH SHARMA

← Home

Crop Portfolio Monitoring Platform

Designing an enterprise platform to digitize India’s crop yield estimation process - transforming manual methods into a scalable, data-driven, and transparent system.

Context

Munich Re, a global reinsurance leader, has been advancing agriculture insurance in India through better portfolio monitoring and claim validation. With the Government’s YES-Tech Guidelines under PMFBY introducing satellite-based crop yield estimation, a major gap emerged - the lack of a digital platform to implement these guidelines.

My Role

Product Design, Analysis, Research

Team

2 Product Managers

8 Engineers

2 Product Managers

8 Engineers

Timeline

~ 50 Weeks

OBJECTIVE

Enhance Munich Re’s agriculture insurance portfolio management by leveraging geospatial data and YES-Tech yield models.

Problem Statement

Crop yield estimation was slow, manual, and error-prone and no digital tool existed to run the newly introduced YES-Tech models.

Solution

YES-Tech Workbench – Configure, compute, and qualify yield models.
YES-Tech Explorer – Map-based outputs with layers, charts, and role-based access.
Data Explorer – Integrated historical & real-time satellite/weather datasets.
User Management – Role-based onboarding for multiple insurers and clients.
Scalable Design – Modular system built for nationwide adoption.

Results

40%

faster claim validation compared to traditional processes

60+

districts covered across multiple states in the pilot phase

1st

digitised YES-Tech platform in India.

understanding basics

🌾 What is CCE ?

Crop Cutting Experiment (CCE) = the traditional way to measure crop yield.

🧩 How Does CCE Work?

Select a random small plot of land (say 5m × 5m) in a farmer’s field.
Harvest all the crops from that plot.
Weigh the harvest → get yield for that sample plot.
Use a formula to scale up to per-hectare yield for the entire field/region.

Select a random small plot of land (say 5m × 5m) in a farmer’s field.
Harvest all the crops from that plot.
Weigh the harvest → get yield for that sample plot.
Use a formula to scale up to per-hectare yield for the entire field/region.

Old Method

Why It’s a Problem

Manual, time-consuming
Limited samples → prone to errors
Delays insurance claim processing

Why It’s a Problem

Manual, time-consuming
Limited samples → prone to errors
Delays insurance claim processing

Old Method

🌾 What is YES-Tech?

YES-Tech = Yield Estimation System based on Technology

🧩 How Does YES-Tech Models Work?

Combine satellite-based estimates with manual CCEs
Create a blended yield value that is more reliable
Aim: gradually reduce dependence on manual, error-prone methods

Combine satellite-based estimates with manual CCEs
Create a blended yield value that is more reliable
Aim: gradually reduce dependence on manual, error-prone methods

New System

Why YES-Tech Matters

✅ Accurate yield estimates

✅ Faster insurance claim processing
✅ Transparency and fairness
✅ Builds trust between farmers, insurers, and govt.

Why YES-Tech Matters

✅ Accurate yield estimates

✅ Faster insurance claim processing
✅ Transparency and fairness
✅ Builds trust between farmers, insurers, and govt.

YES-Tech is like a fitness tracker for agriculture - replacing guesswork with satellite-driven, standardised yield measurements.

New System

research

Uncovering the Gap

Our research showed: despite YES-Tech guidelines, no digital tool existed in India. This ambiguity gave us the opportunity to build something entirely new.

To gain clarity, we:

Spoke with agriculture and internal domain experts to understand crop yield modelling and insurance workflows
Connected with a Claims Validator (primary persona) through the client to map real-world pain points and current processes
Studied the 60-page YES-Tech guideline document released by the governing body to understand policy logic and model expectations
Conducted two structured design thinking workshops with stakeholders to align scope and identify key problems

Who is the end user?

Through our research and design thinking workshops, we identified the key personas who interact with crop yield data and insurance workflows.

Risk Assessor

Looks at data to spot risks in crop portfolios
Shares risk categories with managers and underwriters

Looks at data to spot risks in crop portfolios
Shares risk categories with managers and underwriters

Underwriter

Uses risk data to decide insurance pricing
Creates detailed reports supported by satellite insights

Uses risk data to decide insurance pricing
Creates detailed reports supported by satellite insights

Claims Validator

Reviews farmer claims and checks them against crop data
Ensures payouts are accurate and fair

Reviews farmer claims and checks them against crop data
Ensures payouts are accurate and fair

research

Insights from users & research

Manual & Slow Estimation

Yield calculation relied on Crop Cutting Experiments, which were time-consuming and error-prone.

No Digital Tool for YES-Tech

Government guidelines existed, but there was no platform to run SPM/CHF models.

Low Transparency & Delayed Claims

Fragmented workflows caused delays, disputes, and reduced trust between farmers, insurers, and government.

user Journey

Current User Journey

problem breakdown

Problems to solve

Yield estimation was manual, slow, and prone to errors.

Heavy reliance on Crop Cutting Experiments (CCE) without technology support.

No digital tool existed to run YES-Tech models (SPM, CHF, AI/ML).

solutioning

Early attempts to understand workflow

My first step was to sketch and prototype early visual design concepts.

These weren’t final solutions but conversation starters - they helped me and the team understand how models might be configured and run.

The visuals were key in discussions with data scientists, product managers, and clients - revealing what wouldn’t work and exposing conceptual gaps in the workflow.

My first step was to sketch and prototype early visual design concepts.

These weren’t final solutions but conversation starters - they helped me and the team understand how models might be configured and run.

The visuals were key in discussions with data scientists, product managers, and clients - revealing what wouldn’t work and exposing conceptual gaps in the workflow.

process

Improved User Flow

process

Core Modules

After several back-and-forth iterations and joint sessions, we narrowed down the platform structure to two core modules:

process

Early explorations

A few early explorations helped me identify the overall navigation structure and uncover key gaps in the concept.

solution

The Final Outcomes : Highlight

YES-Tech Workbench

yes-tecH workbench

Raw Data

In this step, clean the data and initiate raw data preparation by adding raw data for a specific combination of region, season, and source configuration.

yes-tecH workbench

Model Configuration & Computations

Model Configuration → Users can create and experiment with multiple setups by tweaking parameters.
Model Computation → Users can run different methods (like blending models through ensemble or applying correction factors), review outputs, and qualify the best model by locking it.

yes-tecH explorer

Map Visualisation & Analysis

Analyse outputs with map views and charts to make informed decisions.

other modules

Sneak-Peak

launch

Finally, When it all went live !

Launch of the crop monitoring platform, attended by leading crop insurers across the country.

Impact

👥 Successfully onboarded 10+ clients and 100+ users

🗺️ Over 60+ districts covered across multiple states in the pilot phase

🕒 40% faster (estimated) claim validation compared to traditional processes

💬 India’s first digitised YES-Tech platform. Praised by both clients and MNCFC (guideline creators) for configurability and transparency

Learnings

Designing in ambiguity

Working on a first-of-its-kind platform taught me how to structure workflows when no prior reference exists.

Systems thinking

Small design choices (like parameters or navigation) had ripple effects across the entire model workflow, reinforcing the need for holistic design.

Cross-functional collaboration

Alignment with data scientists, PMs, & engineers was essential to bridge gaps b/w design, guidelines, and backend feasibility.

Alignment with data scientists, PMs, & engineers was essential to bridge gaps.

Alignment with data scientists, PMs, & engineers was essential to bridge gaps b/w design, guidelines, and backend feasibility.

Simplifying complexity

Translating government guidelines and scientific models into a clear, intuitive experience.

Scalability mindset

Designing not just for one client, but for a platform that can scale across insurers and states, strengthened my ability to think long-term in enterprise product design.

← Home

Portfolio built through sleepless nights, bursts of joy and countless iterations.

Twitter

Instagram