AI Interviewer

The AI Interviewer Agent is the entry point for non-technical domain experts. Its job is to turn conversation into structured, retrievable knowledge without requiring the expert to learn vector search or RAG architecture.

Responsibilities

Interview experts using guided, domain-aware prompts.
Extract concepts, definitions, examples, caveats, and source references.
Organize raw answers into structured notes and metadata.
Identify gaps or contradictions that need follow-up.
Generate embeddings and prepare local knowledge collections.
Publish or update a controlled RAG endpoint when the provider approves.

User Experience

The expert should be able to review what the interviewer captured before anything is exposed to other agents. Early versions can use a simple local workflow: interview, review notes, build embeddings, test retrieval, then publish metadata to discovery.

Design Principle

The interviewer should assist the expert, not replace them. Human approval is central because quality and provenance matter more than bulk ingestion.

Related pages: Knowledge Nodes and Privacy and Data Sovereignty.