Agentic Retrieval: Building Self-Optimizing Search Systems
summary
Claims that agents compress relevance feedback from months to seconds and can tune scoring, schema and indexing in real time, demonstrated by Hornet’s agent-native retrieval stack with Agentic Retrieval workflows that expose a schema-first API design and verifiable state changes. Benchmarks cite BrowserCompPlus (105k docs → 740M tokens) and an 8.4× cost gap vs human-style queries. Describes using agents as judges, auto-research loops and RL-style feedback to hill-climb hyperparameters (BM25, ranking fusion) while warning about cheating and the need for guardrails.
AI-written summary · about
chapters (6)
- 0:00Introduction and overview
- 1:23What is agentic retrieval
- 9:37Agents as new search users
- 13:10Relevancy for agents
- 20:41Self-optimization loops
- 30:04Q&A session
related talks
01h 30mOptimizing Retrieval for Agentic Systems
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch
13mThe State of AI Powered Search and Retrieval
12mAgentic GraphRAG: Autonomous Knowledge Graph Construction and Adaptive Retrieval
22mAgentic Hybrid Retrieval: Orchestrating LLM Agents, Hybrid Search & Explainability in OpenSearch
21mAgentic GraphRAG: Autonomous Knowledge Graph Construction and Adaptive Retrieval
16mUser Signal Dies at the Retrieval Boundary
36mBuilding Agentic AI Assistants: Lessons from productionising agents in Preclinical R&D