Agentic troubleshooting: One pattern for customer-facing AI

In this short talk, Nik Kale breaks down a single, bounded pattern for agentic troubleshooting in customer-facing platforms.

Instead of broad discussions about AI autonomy, the session focuses on one use case: AI systems that observe product state, detect anomalies, and guide users through resolution steps while keeping humans in control. The talk explains how this pattern differs from chatbots or automation scripts and why it can be safely introduced incrementally.

Attendees will learn:
• What “agentic troubleshooting” means in practical terms
• One safe pattern for AI-guided issue resolution
• How platforms introduce intelligence without losing trust