Agentic AI for unstructured data: how to transform enterprise documents into operational knowledge
May 19, 2026 — 12:45 pm - 01:00 pm
Tech Stage
Already registered? Log in now to personalize your experience!
Description
In modern organizations, over 80% of information resides in unstructured content: documents, reports, procedures, proposals, knowledge bases, and historical archives. This represents a valuable asset that is often difficult to access, fragmented, and hard to leverage.
This talk explores how to design an Enterprise Agentic AI platform focused on corporate knowledge management to enable autonomous process orchestration, advanced document retrieval, and automated document generation.
The presented solution implements an adversarial Actor/Critic agentic model supported by a Reinforcement Learning system integrating RAG and GraphRAG engines, specialized agents for automated document generation, and advanced search capabilities across enterprise document repositories, with deployment options also in on-premise or isolated environments.
Special focus will be given to typical enterprise concerns often overlooked in AI projects, such as evolutionary maintenance, non-regression testing, performance benchmarking, observability, scalability, and operational continuity.
A real-world case demonstrates how agentic AI can evolve beyond a technological demo into an industrial-grade platform supporting business, operations, and decision-making.
TIM Enterprise is TIM’s business unit delivering integrated digital solutions to enterprises and public sector organizations, focusing on innovation, sustainability, and security to drive digital...
