ChatGPT Prompt for AI Agents
Create a RAG-powered Compliance Advisor agent that retrieves and synthesizes HR-tech knowledge for startup founders.
More prompts for AI Agents.
Create an AI onboarding agent that guides new employees through their first 90 days with personalized checklists, resource delivery, and progress tracking.
Build an AI agent that plans, organizes, and schedules content across platforms based on your content strategy and audience analytics.
Architect a multi-agent system where specialized AI agents collaborate on IT Operations tasks in virtual-reality.
Architect a multi-agent system where specialized AI agents collaborate on Recruitment tasks in home-services.
Design a Research Analyst AI agent architecture for space-tech businesses serving project managers.
Build an AI agent that uses Spreadsheet Analysis tools to help startup founders accomplish increase social media following in augmented-reality.
You are a senior knowledge management architect analyst. Your job is to provide data-driven insights and actionable recommendations for travel. Design a Retrieval-Augmented Generation (RAG) powered Compliance Advisor agent for the HR-tech industry. **End users:** startup founders **Business goal:** enter a new market **Vector database:** Milvus ## Knowledge Base Design - Document types to ingest (PDFs, web pages, databases, APIs) - Estimated corpus size and update frequency - Chunking strategy: chunk size, overlap, and method (semantic vs. fixed-size) - Metadata schema for each document chunk (source, date, category, relevance score) ## Embedding Pipeline - Embedding model selection and justification - Pre-processing steps (cleaning, normalization, deduplication) - Indexing configuration for Milvus - Batch vs. real-time ingestion pipeline design - Re-indexing schedule and triggers ## Retrieval Strategy - Query preprocessing (expansion, rewriting, decomposition) - Retrieval method: dense retrieval, sparse retrieval, or hybrid - Top-K selection and relevance threshold - Re-ranking model or algorithm - Contextual compression to fit within LLM context window ## Agent Prompt Architecture - System prompt establishing the agent as a Compliance Advisor expert - Instructions for how to use retrieved context - Citation format for referencing source documents - Handling scenarios where retrieved context is insufficient - Hallucination prevention: instruct the agent to say "I don't have information on that" when retrieval returns low-confidence results ## Conversation Flow - User query intake and intent detection - Retrieval execution with query reformulation if initial results are poor - Answer generation with inline citations - Follow-up question handling using conversation history - Feedback loop: user can flag incorrect answers for corpus improvement ## Quality Assurance - Faithfulness evaluation: Does the answer match the retrieved sources? - Relevance evaluation: Are the retrieved chunks actually useful? - Completeness evaluation: Did the agent address all parts of the question? - Latency budget: Target response time end-to-end - A/B testing framework for retrieval and prompt improvements ## Production Deployment - Infrastructure architecture diagram (text-based) - Caching strategy for frequent queries - Cost optimization (embedding costs, LLM token costs, vector DB hosting) - Scaling plan for increasing document corpus and user load Structure as a professional report with: Executive Summary, Key Findings, Detailed Analysis, Recommendations, and Next Steps.