David Knickerbocker

Building Intelligent AI Agents with GraphRAG

Packt Publishing Limited

Build AI Agents That Investigate, Reason, and Retrieve with Knowledge Graphs

<h3>Stop building AI assistants that simply retrieve documents. Learn how to build intelligent GraphRAG-powered agents that investigate complex questions, gather evidence across multiple retrievals, remember previous investigations, and deliver trustworthy answers for production AI applications.</h3><p><br></p><p>Traditional RAG works well when the answer exists inside a few documents.</p><p>Real-world business problems rarely do.</p><p>Enterprise investigations often require connecting information across people, organizations, events, products, research papers, regulations, and thousands of relationships that ordinary vector search simply cannot capture.</p><p>This is where GraphRAG changes everything.</p><h4>In this hands-on live bootcamp, you'll learn how to build Agentic GraphRAG systems, AI agents that don't just retrieve information, but actively investigate problems, decompose complex questions, gather evidence from Knowledge Graphs, remember previous investigations, and synthesize reliable answers.</h4><p>Unlike courses that spend hours teaching how to build Knowledge Graphs from scratch, this workshop focuses on what most AI developers actually need: using existing GraphRAG systems to build powerful AI applications.</p><p>Working with a production-ready Natural Language GraphRAG endpoint, you'll focus on designing intelligent agent workflows instead of wrestling with graph query languages or Prompt-to-Cypher generation.</p><h4>Throughout the workshop, you'll progressively build an investigative AI agent that evolves from a simple GraphRAG client into a conversational research assistant capable of:</h4><ul><li>Connecting an LLM agent to a GraphRAG system</li><li>Retrieving structured knowledge using natural language</li><li>Breaking complex investigations into multiple evidence-gathering tasks</li><li>Synthesizing findings from numerous GraphRAG retrievals</li><li>Maintaining investigation memory for follow-up conversations</li><li>Extending agents with advanced graph analytics, traversal, sentiment analysis, timeline analysis, geospatial reasoning, and other specialized investigative tools</li></ul><p>You'll also learn one of the most important architectural principles for production AI systems: treating the Knowledge Graph, the GraphRAG platform, and the AI Agent as independent, modular components that can evolve separately.</p><p>By the end of this workshop, you'll understand how modern investigative AI systems are designed and have built an intelligent GraphRAG-powered agent that can tackle problems far beyond the capabilities of traditional Retrieval-Augmented Generation.</p><h4>What You'll Build</h4><ul><li>By the end of the bootcamp, you'll have built an investigative AI agent capable of:</li><li>Connecting to a GraphRAG system through natural language APIs</li><li>Investigating complex questions using multi-step retrieval workflows</li><li>Gathering and synthesizing evidence from multiple graph queries</li><li>Maintaining conversational investigation memory</li><li>Extending its capabilities with specialized graph analysis and investigative tools</li></ul><p>Whether you're building enterprise copilots, research assistants, fraud detection systems, intelligence platforms, compliance solutions, or knowledge-intensive AI applications, this workshop provides practical patterns for building AI agents that reason over connected knowledge—not just retrieve documents.</p>

Agenda

Kick off the workshop by exploring the core concepts behind Agentic GraphRAG and why it's transforming how AI applications are built. You'll understand the differences between traditional RAG and GraphRAG, learn how Knowledge Graphs organize connected information, and discover how AI agents use GraphRAG to investigate complex questions. We'll break down the architecture into three independent components the Knowledge Graph, the GraphRAG layer, and the AI Agent and explain why keeping them modular leads to more scalable, maintainable, and production-ready systems. Through live demonstrations, you'll see a real Knowledge Graph and GraphRAG API in action and gain the architectural foundation needed to build intelligent agents throughout the rest of the bootcamp.

Understand: Foundations of Agentic GraphRAG

In this hands-on session, you'll build your first Agentic GraphRAG application by connecting an LLM agent to a GraphRAG API using natural language. You'll learn how an agent receives a user query, retrieves relevant knowledge from a GraphRAG system, interprets the returned dataset, and generates accurate, context-aware responses. Rather than writing complex graph queries, you'll focus on orchestrating the interaction between the agent and the GraphRAG service, giving you a practical "Hello World" implementation that forms the foundation for more advanced investigative and production-ready AI agents later in the bootcamp.

Connect: Connecting Your First GraphRAG Agent

Move beyond simple retrieval by teaching your AI agent to investigate complex problems using multi-step GraphRAG workflows. You'll learn how agents break a challenging question into smaller investigative tasks, perform multiple GraphRAG retrievals, gather supporting evidence from diverse datasets, and combine those findings into a single, evidence-based response. Rather than relying on one search, your agent will reason across multiple sources, synthesize insights, generate concise summaries, and produce comprehensive investigation briefs. This session introduces the core workflow that enables GraphRAG-powered agents to tackle real-world analytical and enterprise research problems with greater accuracy and reliability.

Investigate: Building an Investigative AI Agent

Transform your AI agent from a one-time investigator into a conversational research partner with memory. In this session, you'll learn how to persist investigations by storing the original prompt, supporting questions, retrieved datasets, evidence, and final conclusions, allowing the agent to maintain context across conversations. You'll build an agent that can intelligently answer follow-up questions using its existing investigation instead of starting from scratch, performing new GraphRAG retrievals only when additional evidence is needed. By the end of this session, your agent will deliver faster, more contextual, and more efficient conversations—mirroring how production-grade AI assistants support ongoing research and decision-making.

Converse: Building a Conversational Investigative Agent

Take your GraphRAG agent beyond retrieval and reasoning by equipping it with specialized tools that unlock powerful new investigative capabilities. You'll explore how to enhance agents with graph analytics, graph traversal, community detection, keyword and source analysis, sentiment and bias detection, timeline analysis, geospatial mapping, and other advanced techniques. Through live demonstrations, you'll see how combining GraphRAG with domain-specific tools enables agents to solve richer, more complex problems than traditional AI assistants. This session is designed to inspire creative thinking and show how modular tooling transforms capable AI agents into intelligent investigative systems ready for real-world applications.

Extend: Extending Your Agent with Advanced Investigative Tools

Wrap up the bootcamp by exploring the future of Agentic GraphRAG and where intelligent, graph-powered AI systems are headed. We'll discuss emerging agent capabilities, advanced investigative workflows, production architecture patterns, and strategies for building scalable GraphRAG applications in enterprise environments. You'll gain insights into how these technologies are evolving, the challenges and opportunities that lie ahead, and practical guidance for continuing your learning journey. The session concludes with an interactive Q&A, giving you the opportunity to discuss real-world use cases, implementation challenges, and next steps with the instructor and fellow participants.

Building Intelligent AI Agents with GraphRAG

More events from Packt Publishing Limited

Discover more events from Packt Publishing Limited, from Science & Tech to other experiences you might love.

Still looking for the right event?

Explore all online events to browse and filter by date, category, and more.

Building Intelligent AI Agents with GraphRAG

More events from Packt Publishing Limited

Discover more events from Packt Publishing Limited, from Science & Tech to other experiences you might love.

You might also like...

Browse more events with different dates, prices, and formats to find your next great experience.

Still looking for the right event?

Explore all online events to browse and filter by date, category, and more.