Empowering Business Users with AI Agent-Driven Insights based on an Enterprise Knowledge Graph providing Trusted and Explainable Answers: A Hands-on Tutorial
Daniel H, Metaphacts, COO

Decision-makers and business users must trust the data that informs their decisions. Collecting and integrating data distributed across an enterprise is often labor-intensive and manual. LLMs offer to streamline this process, providing business users with easy data access using their own language. However, LLMs lack transparency in their answer derivation, have limited knowledge of enterprise-specific data and lack the scoping to concrete business domains.

This tutorial demonstrates how to connect an LLM-powered AI agent for a conversational interface to an internal Enterprise Knowledge Graph. The natural language capabilities of the LLM provide easy data access for business users, while the Knowledge Graph supports explainability, fosters trust in the answers provided and allows users to trace the origin of results. We will employ a domain model represented as an ontology to configure the interface for an enterprise-specific use case. This semantic model ensures responses are domain-specific and tailored to the enterprise context, covering information from several company departments, each with its own data model.

Participants will receive their own metaphactory instance to follow along and build a conversational interface grounded by a Knowledge Graph based on a modular ontology. No technical knowledge is required to set up and configure the interface in metaphactory.