Syngenta Accelerates Agricultural Innovation with Progress-Powered Enterprise AI Search Platform

Founded in 2000, Syngenta provides millions of farmers with sustainable solutions to grow and distribute crops and seeds worldwide. A science-led agricultural technology leader, the company leverages more than 70 years of its industry research to discover new chemical compounds and bring cutting-edge products to market.

Syngenta’s scientists were spending thousands of hours looking for the relevant information to support the research and development phases and unintentionally increasing operational costs trying to pull what they needed. In some cases, scientists relied on the “collective memory” of other Syngenta employees, and when those employees left, critical information could be lost forever.

The organization’s vast amounts of research data were not in a centralized location but scattered across numerous repositories and diverse silos. The data was also in different formats, from Word documents to JPEG images to PDFs. Syngenta was using an enterprise search tool that only understood basic keywords and not industry-specific terms. When employees looked for information regarding chemical structures, the tool could not recognize the structure in a document file or provide the extra context. Existing study data often became buried, resulting in duplicated research effort and increased costs.

These obstacles highlighted a major bottleneck in the R&D process. Product Safety Lead Geraint Duck put it simply: “If we can speed up the time it takes our scientists to find relevant information, they can decide quicker, and we can understand risks quicker. They can spend more time on value-added activities as a result.”

Duck and his team wanted to empower R&D scientists to discover and explore relevant studies in an intuitive and efficient way. A search application had to be equipped with advanced filters, interactive visualizations and elaborate document previews to show the complete set of information related to a question and enable scientists to navigate complex chemical data down to the smallest detail. 

Together with Datavid, a Progress implementation partner, Syngenta built Synapse, an enterprise AI search application powered by the Progress Data Platform with Progress® MarkLogic® software at its core. Both organizations wanted to utilize the platform's powerful multi-model data management, search and knowledge graph capabilities.

Datavid and Syngenta first centralized years of R&D data into a single, searchable knowledge base by ingesting and harmonizing millions of documents into the MarkLogic platform. Most of these documents came from internal file systems, such as the company's own SharePoint, and external data sources, like the EPA website.

Thanks to the native semantic capabilities of the Progress Data Platform, the documents were semantically tagged, connected and organized into a knowledge graph, which, just like synapses in neural networks, allows the most relevant information to a researcher’s question to be instantly retrieved, also laying the groundwork for AI systems to understand context, answer complex questions and reason over vast amounts of data.

In this way, Syngenta was able to counter the impact of deepening silos and terminology ambiguity resulting from multiple company mergers. By integrating optical character recognition (OCR) capabilities with the platform, Synapse extracts text from different file formats and provides accessibility to hard-to-search content to make AI-driven search and analysis easier.