Unverified50% confidenceFactExact time
Early RAG was a linear pipeline (user question, vector retrieval, context concatenation, generation); now more systems make retrieval one of an agent's tools, with the model itself deciding when to retrieve, what to retrieve, and whether multiple rounds of queries are needed.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/27/2026
First Seen
Valid until: 1/25/2027
Sources
Related Entities
Related Claims
UnverifiedTraditional RAG is a one-time, static process that relies on a pre-built knowledge index, whereas Agentic RAG introduces agent autonomy to dynamically decide whether to perform additional retrieval and rewrite queries.80% similarVerifiedIntent classification in RAG architecture routing mechanisms is sometimes referred to as Query Routing, and is one of the important optimization techniques for advancing from simple RAG to advanced RAG.77% similarVerifiedProduction-grade RAG systems typically build and fully fuse three retrieval paths: vector retrieval, knowledge graphs, and keyword inverted indexes.76% similarUnverifiedWeb search is a simplified implementation of RAG, in which the model directly calls a search engine to obtain real-time results.75% similarUnverified检索增强生成（RAG）技术让Agent每次推理前从向量库中检索最相关的历史信息片段注入上下文74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1102887API
curl https://kongchang.com/api/v1/knowledge/claims/1102887MCP
get_claim(id=1102887)