How to Catch Tool-Call Errors in AI Agents

Lightning Lessons

Practical skills & tools to accelerate your career

How to Catch Tool-Call Errors in AI Agents

Hosted by Jason Liu and Julia Neagu

Tue, Nov 25, 2025

7:00 PM UTC (1 hour)

Virtual (Zoom)

Free to join

By continuing, you agree to Maven's Terms and Privacy Policy.

Invite your network

Go deeper with a course

Featured in Lenny’s List

Systematically Improving RAG Applications

Jason Liu

View syllabus

Tue, Nov 25, 2025

7:00 PM UTC (1 hour)

Virtual (Zoom)

Free to join

Invite your network

Go deeper with a course

Featured in Lenny’s List

Systematically Improving RAG Applications

Jason Liu

View syllabus

What you'll learn

Detect silent tool-call failures

Learn to identify when AI agents call wrong tools or pass incorrect parameters in production

Evaluate tool-use accuracy at scale

Apply practical techniques using evaluators like limbic-tool-use to measure agent reliability

Trace and improve agent tool interactions

Implement real-time analysis methods to debug and enhance how agents use external tools

Why this topic matters

AI agents fail most often through silent tool-call errors—wrong tools or bad parameters—not bad responses. Learning to detect and fix these failures is essential for building production-ready agents. You'll gain practical skills to evaluate, debug, and improve agent reliability, making you more effective at deploying AI systems that actually work at scale.

You'll learn from

Jason Liu

Consultant at the intersection of Information Retrieval and AI

Jason has built search and recommendation systems for the past 6 years. He has consulted and advised a dozens startups in the last year to improve their RAG systems. He is the creator of the Instructor Python library.

Julia Neagu

Co-founder & CEO, Quotient AI

Julia is the co-founder and CEO of Quotient AI, building the critical monitoring layer for AI agents in production. Before founding Quotient, Julia was the Director of Data for GitHub Copilot, where her team built the systems behind one of the world’s most widely used AI coding assistants. She previously served as Director of Analytics at Tamr and built quantitative models for Aon’s Intellectual Property Solutions group. Julia holds a PhD and MA in Physics from Harvard and an AB in Physics from Princeton.

worked with