Google Research Adds Agentic RAG to Gemini Enterprise Agent Platform with a Sufficient Context Agent for multi-hop queries
Google Research has developed a new agentic RAG framework integrated into the Gemini Enterprise Agent Platform, enhancing its Cross-Corpus Retrieval capabilities. This framework is designed to address the limitations of standard RAG in handling complex, multi-hop queries across various data sources. By employing a multi-agent architecture that plans, reasons, and iteratively searches, the system achieves up to a 34% improvement in accuracy on factuality datasets and better grounding on domain-specific tasks. AI
IMPACT Enhances enterprise search capabilities by improving accuracy and handling of complex, multi-hop queries across diverse data sources.