Should Agents Say I Dont Know? | AirBnB at Interrupt 26 @LangChain
Should Agents Say I Dont Know? | AirBnB at Interrupt 26  @LangChain
Uploaded November 2026 | Updated November 2026, 2 days ago
Pedro Rodriguez, Director of Engineering, ML at Airbnb, explains why his team lets agents answer abstain when they aren't sure and sends those cases to a human, since in high-stakes cases a confident wrong answer is one of the worst possible outcomes. He asks viewers to imagine an agent that approves loans or pays insurance claims, where an answer that looks right isn't enough and you need to explain why the agent reached it.
Captured at Interrupt New York City, the agent conference by LangChain.
#AIAgents #HumanInTheLoop #Airbnb #LangChain #Interrupt
Should Agents Say I Dont Know? | AirBnB at Interrupt 26Why the Harness Matters More Than the ModelHow Bridgewater Built an AI Analyst That Does Hours of Expert Research in MinutesThe 3 Things Clay Measures on Every AI AgentMCP Explained: The USB-C Cable for AI AgentsBuilding Clinical AI Agents with LangGraph: Abridges Eval Stack for High-Stakes HealthcareWhy best model is the wrong question nowIntroducing OpenWiki, an open source agent for repo documentationBuilding ToyotaGPT: 50+ Production Agents, One Config File, Zero Architecture ReviewsWhat Is an AI Agent, Actually?Quickstart: LangSmith Fleet | LangChain AcademyBuilding Governed Agents: A Framework for Cost, Control and Compliance
LangChain |

Should Agents Say "I Don't Know"? | AirBnB at Interrupt 26

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER