Beyond the Answer: How to Check AI’s Work

​Beyond the Answer: How to Check AI’s Work

​AI systems are getting better at producing convincing answers - but how do we know when we can actually trust them?

​As we move from simple LLM applications toward more complex AI and agentic systems, generating an answer is only part of the challenge. We also need ways to understand what happened, what evidence supports the result, and where our visibility ends.

​In this one-hour webinar, we’ll explore AI Observability and how it helps us understand and evaluate AI systems in practice.

​We’ll look at:

​🔍 Observation — what can we actually see about what happened inside an AI workflow?

​✓ Verification — how can we examine the evidence and sources behind an AI-generated result?

​📊 Evaluation — how do we assess whether an answer is correct, relevant, and useful?

​🧩 The limits of observability — what can we know about an AI system’s behavior, and what remains hidden?

​We’ll connect these ideas to real-world AI development and see how observation, verification, and evaluation work together to help us build systems we can understand and trust — rather than simply accepting the final answer.

​Speaker: Tamar Peretz
GenAI / LLM Researcher @ Andy agent lab

​📅 8.10.26
🕘 09:00–10:00
💻 Online

​AI Updates for Moms – קבוצת הוואטסאפ:
Join WhatsApp Group