From POC to Profit: Deploying Customer-Facing AI Agents That Actually Work
summary
Claims that 90% of customer-facing agents stay stuck in POC and that production requires ~50× more work, plus a platform that creates synthetic users and simulated databases to run multi-turn scenarios at scale. The system connects catalogs, reviews and tools, tests return and voice flows (example: retailer Sparky and an Emerald Preferred-style credit card voice path), surfaces missing domain info, integrates into CI/CD and runs nightly regression tests, and replaces much manual QA to catch drift from changing APIs.
AI-written summary · about
chapters (6)
- 0:00AI agent applications intro
- 3:52Deploying customer-facing agents
- 8:04Why AI agents fail
- 10:33Need for agent evaluation
- 13:02Regression testing for AI
- 17:42Q&A
related talks
56mAI Agents That Don’t Suck: How to Build Ones That Actually Drive Business Value
43mTesting Agentic AI Applications: Beyond Traditional QA
19mRise of the AI Architect
18mDispatch from the Future: building an AI-native Company
49mHow to be an AI First company
16mThe AI Agent Iceberg
07h 37mAI Agents Conference 2026 - Day 3
37mThe Production AI Playbook: Deploying Agents at Enterprise Scale