Skip to content

adityanaranje/Agent-Builder-Research-Agent

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 

Repository files navigation

AI-Powered Multi-Agent Workflow

This repository contains an AI-powered multi-agent workflow built using Agent Builder by OpenAI (platform link: Agent Builder), designed to automatically analyze topics and generate actionable insights. Whether you need scientific research summaries, business intelligence, or social insights, this workflow can handle it all in an automated, structured pipeline.


🚀 Overview

The workflow leverages multiple specialized agents that work together to classify queries, gather information from credible sources, and synthesize insights. It acts like a personal AI research assistant capable of producing concise summaries and reports in real-time.

Key Features:

  • Smart query classification into Scientific/Technical, Business/Market, or Social/General.
  • Automated retrieval of top peer-reviewed papers, market reports, or social insights.
  • Summarization and insight synthesis in a concise and actionable format.
  • Integration with reliable sources such as ArXiv, IEEE, NeurIPS, Crunchbase, Statista, and CB Insights.

🛠️ Workflow Steps

1. Query Classification

The workflow begins by classifying the user input using the Starting Agent:

  • Categories:

    • Scientific/Technical – for research, technology, and science queries.
    • Business/Market – for industry, finance, and company-related queries.
    • Social/General – for societal, cultural, or general information.
    • Unrelated – for casual or non-analytical text.
  • The agent returns a short summary of the user query and identifies its category.

2. Specialized Agents

Depending on the classification, the workflow triggers one of the following agents:

a. Technical Agent

  • Searches top peer-reviewed papers from ArXiv, IEEE, NeurIPS.
  • Focuses on extracting methods, results, and key insights.
  • Prepares the information for summarization.

b. Business Agent

  • Collects latest industry news, market reports, and company data from Crunchbase, Statista, CB Insights.
  • Analyzes trends, metrics, and insights.
  • Generates actionable summaries for strategic decision-making.

c. General Agent

  • Retrieves social, cultural, or societal insights from credible sources.
  • Summarizes key trends and opinions concisely.

3. Summarization Agent

  • Receives output from the respective agents.
  • Combines and condenses information into a 500-word summary.
  • Ensures the result is easy to read and actionable.

4. Insight Synthesizer Agent (for Business/Market)

  • Compares summarized findings.
  • Identifies key trends, common insights, and open questions.
  • Produces a structured insight report for decision-making purposes.

🖼️ Workflow Preview

Workflow Preview


🌐 Sources & Tools

  • Technical Research: ArXiv, IEEE, NeurIPS
  • Business Intelligence: Crunchbase, Statista, CB Insights
  • Social Insights: Various credible online articles and studies
  • Agents & AI Models: GPT-5-mini, GPT-5, GPT-4.1

💡 Use Cases

  • Academic research summaries
  • Business market trend analysis
  • Social and cultural insights collection
  • Automated knowledge discovery

🔧 Customization

You can customize the workflow by:

  • Adding new tools or sources to the agents
  • Modifying agent instructions for different research objectives
  • Adjusting summarization length and output format

✨ Conclusion

This workflow demonstrates how multiple AI agents can work together to automate research, analysis, and insight generation, saving time and improving decision-making.

Check out the platform where it was built: Agent Builder

Feel free to contribute, suggest improvements, or integrate it with your own applications!

About

An AI-powered research assistant that uses multiple specialized agents to classify queries, gather credible information, and generate concise insights and reports in real time.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages