How Successful Enterprises Measure AI Quality

Leaderboard scores won’t save you in production. In this session from COLLIDE Data + AI Conference 2025, Micaela Kaplan, ML Evangelist @ HumanSignal, unpacks how successful enterprises actually measure AI quality—beyond the hype. You’ll learn why generic benchmarks often miss the mark, what it takes to evaluate models before they hit production, and how to define “quality” in ways that map to real-world business impact. If you’re building or deploying AI systems in a high-stakes environment, this talk offers a pragmatic, no-fluff framework for evaluating what works—and what doesn’t. Brought to you by HumanSignal. Subscribe for more conversations and insights from leaders shaping the future of data and AI.