Loading the page…
Shaping the future of software engineering
This 45-minute on-demand session explores how Coinbase achieved 72% faster investigation time by integrating AI directly into production operations.
AI inside the incident channel, not beside it
AI as first responder on noisy production signals
Daily SLO check-ins that close the loop
A playbook you can copy
Angelo Marletta
Staff Engineer, Coinbase
AI plugs directly into the incident workflow, joins channels automatically, and collaborates with humans in real time rather than sitting beside the process as a separate tool.
Coinbase taught Resolve to query custom Datadog events for recent Terraform applies and deployments, and to check a dedicated load-testing dashboard, so false alarms get flagged fast.
Daily SLO check-ins, automated reports on what breached, and continuous back-testing — so the system gets sharper with every incident.
The concrete patterns and design choices that move a team from “AI as an experiment” to AI as the default way production gets run.
Watch the full session
Forty-five minutes with Coinbase on a 72% reduction in investigation time, and exactly how they got there.