A model must answer from company documents. Which option fits?
Why C is the answer
Pre-built RAG with Agent Search
RAG brings your own data into the response.
In detail
- Pre-built RAG with Agent Search answers from company documents.
- It brings your own data into responses.
Use case
An illustrative scenario, not a real incident.
An assistant answers from company manuals, using pre-built RAG with Agent Search.
See it in different ways
flowchart TD
A{"Where is the truth?"} -->|"Your company data"| B["Pre-built RAG with Agent Search or RAG APIs"]
A -->|"Public, current facts"| C["Grounding with Google Search"]stateDiagram-v2 [*] --> SourceOfFacts SourceOfFacts --> GoogleSearch: current public facts SourceOfFacts --> AgentSearch: company documents GoogleSearch --> GroundedAnswer AgentSearch --> GroundedAnswer GroundedAnswer --> HumanReview: high impact GroundedAnswer --> [*]
sequenceDiagram participant P0 as User participant P1 as Agent Search participant P2 as Company documents participant P3 as Model P0->>P1: Question P1->>P2: Retrieve P1->>P3: Question and passages P3-->>P0: Answer
Key point: RAG brings your data into answers.
Why the other options are wrong
- A. A higher top-p setting
- Top-p adds no facts.
- B. A role prompt with no data
- A role with no data adds no facts.
- D. Grounding with Google Search only
- Search reaches public facts only.