Research Agent

LLMs confabulate facts. A dedicated research agent with retrieval tools and citation requirements grounds answers in verifiable sources.

A research agent is an investigative journalist — it doesn't opine until it has sources, and it shows its work.

Scenario

Research agent tasked with 'EV battery suppliers in EU' returns 8 findings with URLs, key quotes, and a synthesis paragraph flagged low-confidence where sources conflict.

Solution

In Multi-Agent Systems, apply Research Agent to this scenario: Research agent tasked with 'EV battery suppliers in EU' returns 8 findings with URLs, key quotes, and a synthesis paragraph flagged low-confidence where sources conflict. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to Research Agent (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where Research Agent happens in the code.

1research_agent = Agent(  # key line for Research Agent2    tools=[web_search, fetch_url, rag_search],3    instructions="Cite every claim. Return ResearchReport with sources.",  # key line for Research Agent4    output_type=ResearchReport,  # key line for Research Agent5)

• Treating Research Agent as a black box without evaluation
• Ignoring cost and latency in production
• Skipping error handling for research agent

Quick recap — the most important points from this module.