How Fair Trade USA uses AWS to improve working conditions for farmers

Fair Trade USA™ is a nonprofit organization that is committed to eliminating poverty by promoting sustainable development through ethical trade. They work to ensure fair compensation, safe working conditions for farmers and workers, and sustainable farming practices. When farmers receive a fair price for their harvest, they can farm sustainably, keep their kids in school, and reduce the risk of displacement. You may have seen Fair Trade Certified™ products, which means the product or ingredients meet rigorous social, environmental, and economic standards. Additionally, sourcing Fair Trade Certified products generates additional revenue known as the premium. These funds go back to the producers, who assess needs and democratically determine where to invest in projects for families and communities. In this post, you’ll learn how Fair Trade USA leverages Amazon Web Services (AWS) to improve working conditions for farmers and producers around the world.

Fair Trade USA collects data from a variety of sources, including their customer relationship management (CRM) tool, audit platform, survey tools, and data from external data providers. This data includes impact data for more than 1,500 of their business partners; compliance data against their 120-plus social and environmental protection criteria; socio-demographic data for more than 1.6 million farmers, fishers, and workers in 51 countries; and transaction data for 15,000-plus Fair Trade Certified products. This data, collected in multiple languages, comes in various formats and degrees of completeness and quality.

Architecture

To begin extracting insights from this data, Fair Trade USA performs data extraction and cleanup and stores the data in Amazon Redshift and Amazon Simple Storage Service (Amazon S3). Fair Trade USA utilizes AWS Glue for extract, transform, and load (ETL) processes, which clean and transform the data that is stored in Amazon Redshift and Amazon S3. User-generated content (UGC) from community development projects is processed using Amazon Comprehend and Amazon Translate to translate multilingual content into English. Amazon Rekognition adds AI-driven tagging of UGC media and images, streamlining the discovery of relevant materials for impact reporting. This architecture enables the transformation of raw data from farms, factories, and fisheries into timely, high-quality quantitative and qualitative insights, supporting partner impact evaluation and enhancing transparency across the supply chain. The following Figure 1 shows the architecture.

Figure 1. Fair Trade USA data processing architecture. The major components are Amazon Rekognition, AWS Glue, an Amazon S3 bucket, Amazon Redshift, Amazon Translate, Amazon Comprehend, Amazon EventBridge, and AWS Lambda.

In Q4 2024, Fair Trade USA piloted Insights Hub, a self-service analytics platform built on ThoughtSpot for Fair Trade USA partners to see their data insights. This year, Insights Hub will gradually roll out to most Fair Trade USA partners, with increased understanding of the impact of the partner sourcing practices on farmer and worker livelihoods around the world. As partners gain access to data, they’ll have more questions. Insights Hub is designed as a flexible solution that evolves to address these complex inquiries.

By delivering focused insights, Fair Trade USA leads the way in helping corporations understand how their purchase decisions impact the lives of farmers and workers throughout the world. With the transition from manually created annual reports to on-demand impact data delivery through an automated business intelligence (BI) solution, they’re cutting time to insight from three months to minutes. The gained efficiencies empowers commercial partners to make timely, data-driven decisions that drive supply chain transparency and sustainability.

Insights Hub

In the following example, the organization sees the exact amount of fair trade certified ingredients they’ve purchased during the year, the volume over time, and the premium that was generated. The premium goes directly back to the producer (the farmer, for example) to improve their operations and community.

Figure 2. Premium generation and supply chain tracking via the Insights Hub. User see a bar graph for their production volume and premium, a breakdown of how the premium is being used by producers, and a data visualization of their global supply chain.

Organizations often wonder how the ingredients they purchase are sourced. It’s particularly easy to lose track of where an ingredient comes from, especially as the ingredient passes through multiple steps of being exported, imported and processed. The supply chain map, showing sourcing by producer and origin makes it easy to see where each fair trade certified ingredient was sourced.

Fair Trade USA’s Insights Hub will help companies better understand how their decision to source fair trade certified ingredients and products is helping improve the lives and livelihoods of farmers and workers across the world. As demand for robust Environmental, Social, and Governance (ESG) reporting grows, the Insights Hub empowers companies to track and report on key sustainability metrics, enabling them to demonstrate their impact and enhance their sustainability efforts.

Going further with Stories of Impact

Most consumers have low awareness of ethical trade issues in the global supply chain. Fair Trade USA has identified that by providing meaningful stories sourced from fair trade producers, they can help educate consumers on how their purchase decisions have positive, global impacts on underserved communities. To achieve this goal, Fair Trade USA partnered with the AWS Prototyping team to explore the art of the possible.

The goal of an AWS Prototyping engagement is to remove blockers and ultimately help customers build the foundation to bring systems into production. Working together, the AWS Prototyping team was able to build a prototype that leverages generative AI to allow Fair Trade USA to filter and highlight high-impact stories and images from producers, and then create a story summary that can be edited by Fair Trade USA marketing and shared with partners. The following Figure 3 shows this solution’s architecture.

Figure 3. A prototype architecture to generate stories of impact for Fair Trade USA. The major components are an Amazon S3 bucket, AWS Amplify, Amazon Cognito, AWS AppSync, AWS Step Functions, AWS Lambda, Amazon Bedrock, AWS Secrets Manager, Amazon Aurora, and AWS Key Management Service.

This prototype allows Fair Trade USA to process text and images from producers, evaluate their impact effectiveness based on Fair Trade USA’s assessment criteria of relatability, engagement potential, relevance, and image quality.

The system is accessed through a React TypeScript web application hosted on AWS Amplify, which integrates with Amazon Cognito for secure authentication. The application makes use of AWS Step Functions to process images and analyze text. These step function workflows leverage Amazon Bedrock for generative AI and large language models (LLMs) to generate impact scores and story summaries.

Next steps

Learn more about Fair Trade USA’s impact and begin shopping for Fair Trade Certified products.

If you’re trying to build a data warehouse, consider getting started with Amazon Redshift. Also consider getting started with Amazon Rekognition and Amazon Comprehend as part of your intelligent document processing (IDP) pipeline. Explore what you can do with generative AI by getting started in the Amazon Bedrock console. Finally, learn about AWS for Nonprofits  and connect with the AWS account team.

This initiative is made possible in part by an AWS IMAGINE Grant. Support from AWS and partners like them empowers Fair Trade USA to innovate, elevating their impact and driving meaningful progress toward a future with fair trade for all.