Enterprise Knowledge Search
Bayer Crop Science needed agronomic advisors to answer highly technical crop and product questions faster, but the expertise was spread across decades of agronomy content, product research, proprietary insights, spreadsheets and tables that were difficult to search, synthesize and apply in the moment.
Reported impact
Bayer Crop Science proved that a grounded GenAI assistant could make agronomic knowledge easier to retrieve, help advisors answer questions faster and more completely, and extend into training, marketing and future multilingual farmer support as confidence in the system grew.
Bayer Crop Science worked with EY professionals and Microsoft to build a proof-of-concept generative AI assistant for agronomic advisors. The team grounded a large language model in Microsoft Azure with decades of aggregated agronomy content, product research data and proprietary insights, then translated legacy spreadsheets and tabular sources into structures the model could understand. Additional algorithms reformatted information from different sources into digestible data so the assistant could answer detailed natural-language agronomy and product questions. Retrieval Augmented Generation and prompt engineering were used to retrieve relevant context in real time and improve response quality, while an automated validation process compared outputs against subject-matter experts and open-source LLMs. The system gave Bayer a reusable knowledge layer that could support employees, partners and customer-facing interactions.
Agronomy content, product research data, proprietary insights, spreadsheets and tables were normalized and loaded into an Azure-based LLM stack. RAG retrieves relevant passages at query time, prompt engineering shapes the answer, and automated validation compares outputs with open-source models and subject-matter experts before advisors use the results.
"In just a few months, a few hundred BCS employees have successfully started using the GenAI system."
"The LLM, built in the Microsoft Azure environment, was grounded with decades of aggregated agronomy content, product research data and proprietary insights."
"The team employed Retrieval Augmented Generation (RAG) methodology that could dynamically retrieve relevant information to respond, in real time, to prompt inquiries."
"Together, the team built the POC in just 90 days."
"We recognize that AI holds tremendous potential to assist our agronomists to positively impact farmers worldwide."
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Microsoft Sample Sep 3, 2026 Bayer Bayer needed to help agronomists and other farmer-facing employees answer agronomy, farm management, and Bayer product questions more quickly and accurately, while simplifying crop protection decisions amid regulatory,… 90 percent of E.L.Y.'s users are willing to continue using it and recommend it to a colleague Azure AI model catalog · E.L.Y. · E.L.Y. Crop Protection 
Google Cloud Sample Modified · Jul 21, 2026 Cropin Cropin needed to centralize very large agricultural data streams, reduce infrastructure cost, and speed both analytics delivery and model-training cycles as it expanded OrbitAI across hundreds of crops and global… 3 to 4x increase in output Apigee · BigQuery · Cloud SQL 
EY Sample Archived · Nov 14, 2025 Microsoft Microsoft's finance organization wanted to improve data literacy and speed product-launch work, but critical knowledge was spread across tens of thousands of historical financial documents and disparate databases,… 50% reduction in the time required to gather finance requirements for new product introductions Finance Launch AI · Natural Language Processing (NLP) · Replay AI 