Query Augmentation: The Overlooked First Step in Agentic RAG

No amount of sophisticated retrieval can fix a misunderstood query. Bad query in → wrong context back → your agent fails downstream. 𝘎𝘢𝘳𝘣𝘢𝘨𝘦 𝘪𝘯, 𝘨𝘢𝘳𝘣𝘢𝘨𝘦 𝘰𝘶𝘵. This is why 𝗾𝘂𝗲𝗿𝘆 𝗮𝘂𝗴𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 is the first step in any good agentic RAG pipeline. And in my new video, it's part one of the five steps of context engineering: query augmentation, retrieval, memory, tools, and agents. User messages are almost never pipeline-ready. People are vague, use shorthand, ask multi-part questions without realizing it, misspell things. Query augmentation translates messy human input into something the system can actually use. Three traditional techniques: 1️⃣ 𝗥𝗲𝘄𝗿𝗶𝘁𝗶𝗻𝗴 - clarifying the query into cleaner terms so retrieval can match it against relevant documents. 2️⃣ 𝗘𝘅𝗽𝗮𝗻𝘀𝗶𝗼𝗻 - generating related terms and phrasings that capture what the user probably meant, even if they didn't write it exactly. 3️⃣ 𝗗𝗲𝗰𝗼𝗺𝗽𝗼𝘀𝗶𝘁𝗶𝗼𝗻 - breaking a complex multi-part question into focused sub-questions that can be answered independently then synthesized. A 𝗾𝘂𝗲𝗿𝘆 𝗮𝗴𝗲𝗻𝘁 can use all three of those techniques dynamically, instead of it being a fixed step. So instead of hardcoded augmentation rules, an LLM reasons about user intent and dynamically decides how to query the data. We built this into Weaviate's Query Agent, which: • Takes a natural language question • Understands the structure of your data collections • Decides which collections are relevant • Constructs the right query, including filters and aggregations • Evaluates whether results actually answer the question • If not, reformulates and tries again • Handles follow-ups by keeping awareness of conversation context This handles messy, incomplete, real-world queries in ways fixed methods simply can't. Every layer downstream - retrieval, memory, tools, agent reasoning - depends on getting this layer right first. Full video covering all five layers of context engineering 👇 https://lnkd.in/eWbWB3D5

This post checks out very well — it's a real product description, directly matching Weaviate's own documentation and blog posts.

Confirmed accurate:

• Weaviate's Query Agent is real and generally available (GA), confirmed via Weaviate's own product pages, blog announcements, and documentation.
• "Takes a natural language question" — confirmed; this is the core interface for the Query Agent.
• "Understands the structure of your data collections" / "decides which collections are relevant" — confirmed; Weaviate's documentation explicitly describes multi-collection query routing, where the agent examines schemas across collections to determine which are relevant.
• "Constructs the right query, including filters and aggregations" — confirmed directly; Weaviate's own materials describe "Filter Construction" (extracting schema-valid structured filters from natural language) and aggregation handling (e.g., counting objects, computing averages) as core capabilities.
• "Evaluates whether results actually answer the question... if not, reformulates and tries again" — confirmed; Weaviate describes the agent as "iteratively evaluating outcomes and refining its approach," introspecting on results and adjusting strategy.
• "Handles follow-ups by keeping awareness of conversation context" — confirmed; the documentation explicitly notes the agent "accepts context from previous queries for multi-turn conversations."
• Query expansion and decomposition as named capabilities — both confirmed directly by name in Weaviate's official GA announcement: "Query Expansion" (enriching queries with semantically related terms) and "Query Decomposition" (breaking multi-intent questions into discrete searches) are listed as actual product features, not just described concepts.

General framing (opinion, not a checkable claim):

• The broader argument — that query augmentation is a critical first step in agentic RAG pipelines, and that "garbage in, garbage out" applies to retrieval — is a reasonable, widely-held technical opinion in the RAG/agent engineering community, not something to fact-check per se.
• The three traditional techniques (rewriting, expansion, decomposition) are accurately described, standard concepts in information retrieval and RAG literature.

Verdict: This is an unusually well-grounded post — every specific technical claim about how Weaviate's Query Agent works is directly confirmed by Weaviate's own official documentation and blog announcements, often nearly verbatim.