Silent Failures and AI: Everpure, Greenlight, Reltio Panel
Platform leaders from Everpure, Greenlight and Reltio on catching silent failures, AI agents in production troubleshooting, and runbooks arriving with alerts.
Some of the costliest failures never trip an alert: every light is green and something is still wrong. At FLOCK '26, Francois Martel, Head of Field Engineering at NeuBird, talks with Dheeraj Khanna (VP, Platform Engineering, Everpure), Naveen Ede (Principal Platform Engineer, Greenlight) and Dmitry Blinov (Sr. Director of Engineering, Reltio) about finding those failures, and about what changes when AI agents join production troubleshooting.
Dheeraj argues for shared responsibility across software, data and infrastructure, so teams speak one language instead of pointing fingers. Naveen defines the invariants each layer has to hold, gives each one an owner, and brings every signal into a single view for AI to analyze. Dmitry describes the data quality checks Reltio built over its metadata, because a data platform can fail its customers while every service is up.
The panel covers where agents can safely help without breaking compliance (the same access rules for agents as for people, contracts for what an agent may do, and data masking), why traceability matters to auditors, and what eats triage time: most failures follow a change, and the hard part is finding which one. They close on habits to abandon, starting with static runbooks in a wiki, which are giving way to runbooks that arrive with the alert.
In this panel
Shared responsibility across software, data and infrastructure
Invariants, ownership and a single view of every signal
Silent failures hiding in the data itself
Where AI agents can safely help in production
Why traceability matters to auditors
From static runbooks to runbooks that arrive with the alert
Related Resources
See NeuBird in action
Root cause in minutes, not war rooms.
