Find Root Causes Faster with AI (Part 1)
Christoph Berlin explains how agentic AI can speed up root cause analysis by correlating signals across multiple systems, ranking likely causes by confidence, and presenting supporting evidence so engineers can start investigations faster after stoppages or slowdowns.
Overview
The video introduces an agentic AI approach to root cause analysis (RCA), focused on reducing the time it takes engineers to identify where to start when a manufacturing site experiences a stoppage or slowdown.
Key points covered:
Why root cause analysis can take days
- RCA often requires engineers to manually correlate information spread across multiple systems.
- The investigation work (finding and correlating signals) can be time-consuming even before remediation begins.
How agentic AI speeds up investigations
- An AI agent can pull relevant data from scattered systems and do correlation work on the engineer’s behalf.
- The goal is to accelerate investigation and triage, not replace the people who ultimately resolve incidents.
Ranking likely causes with evidence
- The agent produces a ranked list of likely causes.
- Each candidate cause is associated with:
- A confidence level
- Supporting evidence to help the engineer validate where to start
Resources
- Microsoft for Manufacturing: https://www.microsoft.com/AI/manufacturing
- Azure resources (Azure Essentials): https://azure.com/AzureEssentials
Chapters
- 0:00 Why root cause analysis takes days in manufacturing
- 1:08 How agentic AI speeds up root cause analysis
- 2:11 How an AI agent ranks likely causes for engineers