Prompt Library ⚙️ Technical Design a Multi-Agent Research Pipeline
GPT-4o ⚙️ Technical Advanced

Design a Multi-Agent Research Pipeline

Architect a coordinated multi-agent system where specialized AI agents divide research tasks, share findings, and synthesize a final report.

👁 0 views ⎘ 0 copies ♥ 0 likes

The Prompt

# Design a Multi-Agent Research Pipeline

You are an AI pipeline architect. Design a multi-agent research pipeline for [RESEARCH DOMAIN OR TOPIC].

## Agent Roles

Define each agent's specialization:
- **Search Agent:** queries sources, returns raw results
- **Analysis Agent:** extracts claims, scores relevance
- **Fact-Check Agent:** cross-references claims against [TRUSTED SOURCE LIST]
- **Synthesis Agent:** assembles the final report

## Inter-Agent Communication

- Message format between agents: [FORMAT, e.g. JSON with confidence score]
- State management: how agents share context without duplicating work
- Error handling: what happens when one agent fails or returns low-confidence results

## Orchestration Logic

- Parallel vs sequential execution: which agents run simultaneously
- Timeout and retry rules for [EXPECTED LATENCY]
- Consensus threshold before synthesis begins

## Sample Pseudocode

Write orchestration pseudocode in [LANGUAGE OR FRAMEWORK, e.g. Python with LangGraph].

## Quality Controls

How does the pipeline detect hallucinations and flag gaps in coverage?

📝 Fill in the blanks

Replace these placeholders with your own content:

[RESEARCH DOMAIN OR TOPIC]
[TRUSTED SOURCE LIST]
[FORMAT, e.g. JSON with confidence score]
[EXPECTED LATENCY]
[LANGUAGE OR FRAMEWORK, e.g. Python with LangGraph]

How to use this prompt

1
Copy the prompt

Click "Copy Prompt" above to copy the full prompt text to your clipboard.

2
Replace the placeholders

Swap out anything in [BRACKETS] with your specific details.

3
Paste into GPT-4o

Open your preferred AI assistant and paste the prompt to get started.