About This Event
Text-to-SQL benchmarks top out around 78% accuracy — not good enough when business decisions depend on the answer. Ryan Boyd walks through what it takes to close that gap and make LLM-powered analytics actually reliable. Covers the failure modes that trip up AI agents on real business data, the metadata and schema patterns that improve accuracy, and how MotherDuck approaches the problem differently.
What You'll Learn
- Why text-to-SQL accuracy plateaus at 78% and what causes the remaining failures
- Schema and metadata patterns that improve LLM query accuracy on business data
- MotherDuck's approach to making AI-powered analytics reliable enough for production use
Who Should Attend
This event is ideal for:
Data EngineersData AnalystsAnalytics EngineersTechnical Leaders
