Scaling RAG

As AI systems mature, one thing becomes clear: Retrieval-Augmented Generation (RAG) is a design space that requires thorough planning. Vector search, keyword search, hybrid retrieval, reranking, graph-based retrieval, agentic workflows… the options keep expanding. But more choice doesn’t automatically mean better systems.
In this session, we’ll step back from individual techniques and focus on decision-making: which retrieval approach to use, when, why? We’ll explore how traditional search, advanced retrieval methods, RAG pipelines, and agentic systems complement each other and where they don’t.

💡 You’ll Learn:

• The core retrieval techniques powering modern AI systems (keyword, vector, hybrid, structured, graph-based)
• When traditional search beats RAG and when it absolutely doesn’t
• How advanced retrieval methods (reranking, multi-stage retrieval, query rewriting) improve results
• Where Agentic + RAG architectures add real value and where they introduce unnecessary risk or cost
• Common architectural mistakes teams make when “stacking everything together”

🧠 Who Should Attend?

• C-level executives and VPs shaping AI strategy
• Product and Innovation Leaders designing AI-powered experiences
• Data, Analytics, and AI Leaders responsible for system reliability
• Teams already using RAG or Agentic AI and questioning their current approach
• Anyone asking: “Are we overengineering this… or underengineering it?”

The conversation will remain high-level, practical, and jargon-free, with examples and plenty of time for discussion and Q&A.