The Agentic Stack: Relevance, Reliability & AI Agents in OpenSearch

The Agentic Stack: Relevance, Reliability & AI Agents in OpenSearch

Join us for the OpenSearch User Group Munich as we dive deep into the next frontier of search: AI Agents. Whether you're building intelligent pipelines, tuning relevance at scale, or trying to make your agent workflows more reliable and observable, this evening has something for you. We'll explore how agentic architectures are reshaping the way we think about search — and how you can build, evaluate, and iterate faster with the right tools and practices. Bring your questions, your curiosity, and your appetite for the future of OpenSearch!

Agenda

• 🍕 18:00 — Doors Open — Networking & Pizza
• 🤖 18:30 — Talk 1: "Building Reliable AI Agents in Weeks, Not Months" — Dotan Horovits & Megha Goyal
• 🔎 19:00 — Talk 2: "Agentic Search and the Relevance Agent" — Aswath Srinivasan & Daniel Wrigley
• 🍺 19:30 — Networking & Refreshments

Talk 1: "Building Reliable AI Agents in Weeks, Not Months" — Dotan Horovits & Megha Goyal

How do you validate AI agents before production? Unlike traditional software with clear pass/fail tests, agents operate probabilistically across reasoning, tool selection, and multi-step decisions.

OpenSearch's open source Eval-Driven Development tool brings Automatic Relevance Tuning (ART) to agent development. Define test suites, run benchmarks, and measure performance across reasoning quality, tool accuracy, and task completion.

We'll share a compelling case study: one agent built this way and shipped in 6 weeks; another retrofitted later took 12 months. See how systematic evaluation transforms dev cycles.

Talk 2: "Agentic Search and the Relevance Agent" — Aswath Srinivasan & Daniel Wrigley

Agentic search lets you ask questions in natural language and have OpenSearch plan and execute the retrieval automatically. A preconfigured agent reads the question, plans the search, and returns relevant results.

The OpenSearch relevance agent helps you identify and resolve search relevance issues through natural language conversation. Using a chat interface embedded in OpenSearch Dashboards, you can interactively diagnose problems, receive step-by-step tuning guidance, and execute complex workflows. The agent leverages the tools provided by the Search Relevance Workbench.