This is a submission for the AI Agents Challenge powered by n8n and Bright Data
What I Built
I built Company Research Analyst, an AI-powered research agent that takes a company’s domain name as input and generates a structured, comprehensive research report. The agent automates the tedious process of gathering company insights from multiple trusted sources like LinkedIn, Crunchbase, PitchBook, and Google News.
This tool is designed for analysts, investors, and business developers who need quick, credible, and actionable company intelligence without spending hours across multiple platforms.
Demo
n8n Workflow
Company Research Analyst Agent
Crunchbase Scraper
LinkedIn Scraper
Pitchbook Scraper
Social URLs
Google News Search Query
Technical Implementation
The workflow is powered by n8n + LangChain and orchestrates multiple research tools into a single AI agent.
- System Instructions: Configured the agent as a Company Research Specialist with a clear task breakdown: overview, products, market presence, key personnel, financials, tech stack, news, and competitors.
- Model Choice: Google Gemini (flash-lite) for reliable, low-latency reasoning with a temperature of 0.2 for concise, fact-driven outputs.
- Memory: Implemented a Buffer Window Memory node for short-term context (25 message window).
-
Tools Integrated:
- Social Profiles Extractor (auto-discovers LinkedIn, Crunchbase, PitchBook URLs).
- LinkedIn Company Scraper (company overview, size, founders, updates).
- PitchBook Scraper (funding, valuations, M&A).
- Crunchbase Scraper (funding rounds, key personnel, competitors).
- Google News Search (latest articles and coverage).
Bright Data Verified Node
The Bright Data node is integrated to ensure verified, up-to-date, and high-quality data is retrieved across different company data sources, boosting the reliability of the final research report.
Journey
The biggest challenge was synthesizing fragmented data across multiple platforms into one cohesive report. Using n8n’s modular workflow design, I chained scrapers and search utilities while keeping the agent’s reasoning focused with structured system prompts.
I learned how powerful n8n can be when combined with LLMs and external data sources—it’s possible to automate what traditionally required hours of manual research. This agent can be extended for VC due diligence, competitor benchmarking, or sales prospecting use cases.
Top comments (1)
Super impressive build 👏. Love how you combined n8n + LangChain + Bright Data into a single research agent that actually feels practical for analysts and investors. The structured breakdown (overview, products, market, financials, competitors, etc.) makes it way more actionable than just dumping raw data.