The customer is a leading Argentine pharmaceutical company founded in 1940. With more than 80 years of commitment to healthcare excellence, the customer has built a legacy supported by a community of medical professionals and specialists. The company offers a wide range of primary care and specialty pharmaceutical products across Argentina and exports to over 40 countries.
The customer and Netrix Global have built a strong relationship over the past two years, beginning with the successful deployment of an Industrial Data Lake to harness and evolve the production data from the customer’s manufacturing facilities.
As the customer continues its digital transformation, the company sought to streamline data access for its APMs (medical sales representatives). These field-based employees needed faster and simpler access to commercial data—such as doctor prescription trends and pharmacy sales—without relying on complex dashboards or traditional BI tools that were often impractical to access from mobile devices or remote locations.
The objective was to develop a minimum viable product (MVP) that leveraged Generative AI to answer natural language questions based on commercial data housed in the company’s corporate Data Lake. For instance, a user could ask, “Which doctor prescribed the most [product name] in the last three months?”, and receive a precise, contextual answer without technical knowledge of SQL or dashboards.
Netrix Global designed and implemented an AI-powered natural language interface that connects directly to the customer’s AWS-hosted Data Lake, enabling intuitive and secure data querying for business users.
The solution leverages AWS Bedrock to interpret and classify natural language questions. The core solution utilizes four collaborative agents that prepare, create, validate, and customize data queries, employing Anthropic models such as Haiku and Sonnet.
Microsoft Teams serves as the query interface, allowing users to ask questions in conversational language. The workflow is orchestrated by Frontend and Answer Question Lambda functions. Bedrock LLMs classify the question to the correct data domain and generate the appropriate SQL query.
Amazon Athena validates and executes the query against AWS Glue catalog tables. Finally, Bedrock finalizes and structures the response to deliver a human-readable answer. All invocation metadata is stored securely in Amazon DynamoDB.
This architecture enables seamless interaction with vast datasets, while ensuring data integrity, security, and performance.
By reducing friction and time in how employees access data, The customer is equipping its commercial teams with real-time insights and enabling faster, better-informed decisions in the field. With this MVP, The customer is setting the stage for future AI-driven innovations that further its mission of healthcare access and excellence.
Netrix Global brings a wealth of expertise and capabilities to support The customer’s digital transformation:
Netrix’s multidisciplinary approach and commitment to excellence ensure that customers like The customer can achieve their strategic goals and drive innovation in their respective industries.
You can learn more about this Netrix solution on AWS Marketplace.
No matter what challenge you’re facing today, our team of technical experts can get you started on a path to a better solution, whether you’re modernizing finance operations or rethinking resource allocation through incremental improvements. We’ll partner with you to:
