AI Agents in Production: Stories from the Field and Systems That Last

Date: Wednesday, February 11th, 2026
Time: 17:00-18:30
Location: Online Webinar
*The webinar will be held in Hebrew

On the Agenda
17:00-17:05 Opening Words A short introduction and overview of what to expect.

17:05-17:45 AI Agents from Dream to Reality Amit Goldstein, Senior Backend Engineer, Tikal

Building an AI agent is only 20 percent of the work. The other 80 percent starts when it meets production.
In this session, Amit will share real world horror stories from agent based systems, including an agent that deleted a production database and claimed it did not. Using these cases, we’ll explore why most agent projects fail before reaching production.
The talk will cover the core challenges teams face when operationalizing agents, including governance, cost management, evaluation strategies, guardrails, and security. You’ll leave with a practical checklist designed to help you avoid becoming the next cautionary tale.

17:45-18:30 Guardrails for Production Grade AI Agents Gilad Ivry, CPO & Co-Founder, Qualifire

AI agents are evolving from simple chat interfaces into systems that plan, call tools, and take real actions. After understanding where agents fail in production, the challenge becomes designing systems that behave predictably under real traffic, real data, and real adversaries.
In this session, Gilad will walk through a practical blueprint for turning an agent from a promising proof of concept into a production grade service. We’ll discuss how to design evaluations that reflect real user outcomes rather than vanity metrics, how to run continuous evaluations using live traces, and how to turn error analysis into a repeatable improvement loop.
The session will also cover guardrails for multi step workflows, including structured output validation, grounding and citation checks, tool use constraints, and policies that prevent silent failures.

Please Note
*By registering for this event, you agree to receive emails from Tikal

See you online