Honeycomb Observability Day SF - Kesha Mykhailov, Fin.ai: Human-Centric Observability in AI Systems
Empathy is one of the superpowers of modern engineering teams, especially when building tools that interact with humans. This talk by Kesha Mykhailov tells the story of Fin, Intercom's Customer Support AI agent, and how they transformed their approach to Fin's observability.
In this video:
- Hear how Intercom used Honeycomb to surface customer-centric signals to improve end-to-end latency of its Fin.ai customer support agent
- Learn how distributed tracing with Honeycomb helped them surface business-centric signals to bring costs down
- See why Intercom established SLOs to prevent degradation of its systems
- Understand why "data available" is different from "data observable"
For more: Read the case study on how Honeycomb helped Intercom observe and operate Fin.ai.