DigiNews

Tech Watch by Johan Denoyer

← Back to articles

Building Reliable Agentic AI Systems

Quality: 8/10 Relevance: 9/10

Summary

Thoughtworks' Bayer PRINCE case study details a production-ready agentic AI system for preclinical data using Retrieval-Augmented Generation (RAG) and Text-to-SQL. It explains the multi-agent architecture, context and harness engineering, data pipelines, governance, explainability, and resilience, highlighting production considerations in regulated environments.

🚀 Service construit par Johan Denoyer