This is a submission for the AI Agents Challenge powered by n8n and Bright Data
Pre-requisite
- New users of Bright Data, please make sure to sign-up here - Bright Data
- n8n
- Google Gemini. Please Sign up on Google AI Studio to get the API Key.
Download the Workflow
What I Built
I built an AI-driven company deep research workflow that combines Bright Data scraping, Google Search enrichment, and Google Gemini intelligence to automatically generate a human-readable company research report in Markdown format.
The workflow brings together three key sources of truth:
- CrunchBase → Company funding, size, acquisitions, and market positioning.
- Glassdoor Company Info → General overview, employee ratings, and employer branding signals.
- Glassdoor Company Reviews → Employee sentiment, culture, leadership perception, and workplace insights.
These datasets are aggregated, cleaned, and then passed into Google Gemini, which transforms raw information into a strategic deep research report. The final output is a structured, insight-rich document ready for use in competitive intelligence, due diligence, recruitment, and sales prospecting.
The Problem
Conducting thorough company research traditionally involves:
- Visiting multiple platforms (Google, CrunchBase, Glassdoor).
- Copy-pasting data into spreadsheets or documents.
- Manually reading and summarizing employee reviews, which are often noisy and inconsistent.
- Struggling to merge hard business data (e.g., funding history) with soft cultural insights (e.g., employee sentiment).
This manual process is:
- Time-consuming (hours per company).
- Error-prone, since unstructured data is difficult to consolidate.
- Not scalable, especially when researching dozens or hundreds of companies.
The Solution
The workflow solves these challenges by automating the end-to-end company research pipeline:
-
AI Agent (Gemini) as Orchestrator
- Constructs Bright Data and Google Search queries.
- Identifies relevant Glassdoor and CrunchBase URLs.
- Normalizes the links into JSON using structured parsing.
-
Bright Data Scraping & Extraction
- CrunchBase → Funding rounds, acquisitions, employee counts.
- Glassdoor Company Info → Ratings, headquarters, overview.
- Glassdoor Reviews → Extracts employee feedback for culture and sentiment.
-
AI-Powered Interpretation with Google Gemini
- Cleans and summarizes noisy review text into insightful narratives.
- Synthesizes multiple datasets into a cohesive deep research report.
- Formats the report in Markdown for easy export, sharing, or conversion to PDF.
-
Final Deliverable: Deep Research Report
- Provides a balanced view by merging:
- Quantitative business data (CrunchBase).
- Qualitative cultural insights (Glassdoor).
- Reduces hours of manual research into minutes of automated reporting.
- Provides a balanced view by merging:
The result is a scalable, reliable, and AI-powered research assistant that empowers teams across recruitment, investment, sales, and strategy.
Introduction
This workflow is designed to perform deep company research by combining Bright Data scraping, Google Search enrichment, and AI-driven interpretation with Google Gemini.
At its core, the workflow focuses on building a comprehensive Deep Research Report by integrating three critical data streams:
- CrunchBase → Funding, acquisitions, size, and market positioning.
- Glassdoor Company Info → Company overview, general facts, and employer branding signals.
- Glassdoor Company Reviews → Employee sentiment, leadership feedback, and culture insights.
The extracted data is normalized, enriched, and finally synthesized into a human-readable Markdown report by Google Gemini. This ensures that raw data (e.g., JSON dumps from Bright Data) is transformed into strategic insights with clear narratives.
The workflow is particularly useful for:
- Competitive Intelligence → Compare multiple companies on funding, growth, and employee sentiment.
- Investor/VC Due Diligence → Validate funding data alongside employee perspectives.
- Market Research → Understand brand perception and workforce satisfaction in target industries.
- Recruitment Insights → Position employer branding by combining company facts with real employee experiences.
By merging CrunchBase’s hard business metrics with Glassdoor’s cultural insights, the workflow produces a well-rounded research report that supports both quantitative and qualitative analysis.
Use Cases & Real-World Applications
1. Competitive Intelligence
- Gather data on competitors from CrunchBase and Glassdoor.
- Analyze company size, funding, employee sentiment, and growth trajectory.
- Build benchmark reports for strategic decision-making.
2. Recruitment & Employer Branding
- Extract Glassdoor reviews to understand employee sentiment.
- Present AI-enhanced summaries of culture, leadership, and employee satisfaction.
- Aid recruiters in positioning companies more effectively when pitching roles.
3. Investor/VC Due Diligence
- Pull CrunchBase funding data and combine it with Glassdoor reviews.
- Generate AI-curated summaries of risks, strengths, and employee perspectives.
- Accelerate investment decision-making with reliable research reports.
4. Sales Intelligence / Account Research
- Enable B2B sales teams to perform deep prospect analysis.
- Extract data from public search, Glassdoor, and CrunchBase before outreach.
- Provide sellers with AI-driven one-pagers on target accounts.
Workflow Overview
The workflow follows these main steps:
-
Chat Trigger
- Starts when a user sends a company name via chat.
-
Set Input Fields
- Captures the company name and prepares it for downstream nodes.
-
AI Agent (Google Gemini)
- Constructs Bright Data search queries.
- Identifies and retrieves relevant URLs (Glassdoor, CrunchBase, reviews).
- Uses Structured Output Parser to normalize URLs into JSON format.
-
Bright Data Extraction
- Glassdoor Company Info → General company overview.
- Glassdoor Reviews → Employee sentiment and reviews.
- CrunchBase Data → Funding, size, acquisitions, market insights.
-
Human-Readable Content Extraction (Glassdoor Reviews)
- Uses Google Gemini to convert raw reviews into natural Markdown summaries.
-
Merge Responses
- Combines CrunchBase, Glassdoor overview, and review summaries into a single dataset.
-
Deep Research Builder
- AI Agent generates a comprehensive research report in Markdown format.
- Includes Glassdoor insights, CrunchBase data, and AI-curated analysis.
-
Respond to Chat
- Sends the research report back to the requester in real time.
Key Components
AI Agent (Gemini-powered):
Acts as the orchestrator for data retrieval, URL discovery, and query building.Bright Data Scraping Tools:
Handle web scraping and structured data extraction from CrunchBase and Glassdoor.Structured Output Parser:
Ensures the AI output is formatted as JSON for reliability.-
Gemini Chat Models:
- Used at three levels:
- Query building & orchestration (AI Agent).
- Human-readable review extraction.
- Deep research synthesis into Markdown.
n8n Merge Node:
Combines multi-source insights into a single structured object.
Output
The final output is a Markdown Deep Research Report containing:
- Company Overview (Glassdoor + CrunchBase data).
- Employee Sentiment Summary (Glassdoor reviews).
- Funding, Size, and Market Data (CrunchBase).
- AI-generated strategic insights (growth trends, risks, opportunities).
Deep Research Report Sample: Bright Data
1. Company Overview
Website: www.brightdata.com
Headquarters: Netanya, Israel
Founded: 2014
Type: Private Company
Industry: Internet & Web Services
Employees: 201 to 500
Revenue: $100 to $500 million (USD)
Description
Bright Data positions itself as the world’s #1 web data platform, providing crucial public web data solutions to a diverse client base including Fortune 500 companies, academic institutions, and small businesses. Their platform enables users to research, monitor, analyze data, and make informed decisions through efficient, reliable, and flexible data retrieval.
Serving over 15,000 customers globally across nearly every industry, Bright Data offers a range of no-code data solutions for business owners, alongside a robust infrastructure tailored for engineers and IT professionals. Key benefits highlighted by users include cost-effective, fast, and stable public web data collection at scale, effortless conversion of unstructured data into structured data, and superior customer experience. The company emphasizes its commitment to transparency and compliance.
In 2021, Bright Data launched The Bright Initiative, a separate organization dedicated to providing pro-bono access to Bright Data’s technology and expertise. This initiative partners with NGOs, NPOs, academic institutions, and public bodies globally to drive positive change, having collaborated with over 600 organizations to date.
Mission Statement
Bright Data's mission is to create technologies that preserve a transparent internet, ensuring easy access to and collection of public web data. They believe that making public web data readily accessible is vital for maintaining openly competitive markets, which ultimately benefits everyone. The company actively partners with and serves entities that align with its core values of transparency, innovation, and trust.
2. Ratings & Reviews
Based on 182 anonymous reviews, Bright Data maintains an overall rating of 3.7 out of 5 stars.
Overall Rating: 3.7 ★
Would Recommend to a Friend: 70%
CEO Approval (Or Lenchner): 75%
Total Reviews: 182
Ratings by Category (out of 5.0)
Culture & values: 3.5
Diversity, Equity & Inclusion: 3.8
Work/Life balance: 3.8
Senior management: 3.5
Compensation and benefits: 3.8
Career opportunities: 3.5
3. Awards & Accolades
Bright Data has received notable recognition, including:
Top 50 EMEA companies, G2, 2023
Best Estimated ROI, G2, 2023
4. Pledges & Certifications
- Pledge to thrive: Bright Data has committed to taking steps to prioritize employee well-being.
5. Employee Insights & FAQs
Employee Sentiment
Overall Employee Rating: 3.7 out of 5 stars (based on 182 anonymous Glassdoor reviews).
Recommendation Rate: 70% of employees would recommend working at Bright Data to a friend.
Business Outlook: 72% of employees believe Bright Data has a positive business outlook.
Compensation and Benefits
- A specific employee review (Dec 20, 2024) indicates a concern regarding compensation and benefits for remote workers, stating that the company "purposefully staff[s] the company with a ton of remote workers from foreign countries so they don't have to pay them living wages or provide benefits...."
Interview Experience
Positive Interview Experience: 45% of job seekers rate their interview experience at Bright Data as positive.
Interview Difficulty: Candidates report an average difficulty score of 3 out of 5 for job interviews.
How to Get a Job
- Prospective candidates are advised to browse open positions, apply, and prepare for interviews, noting a moderate difficulty level in the process.
This report can be:
- Delivered in-chat.
- Stored in Google Sheets, Notion, or a database.
- Exported as a PDF for reporting purposes.
Major Challenges and Solutions
Challenge 1: Fragmented Company Data Across Multiple Platforms
Problem: Company data was spread across Glassdoor, CrunchBase, and general Google Search results, each with unique structures and reliability levels.
Solution: Implemented Bright Data scrapers and standardized them with a Structured Output Parser in n8n, ensuring normalized JSON formats that could be merged seamlessly.
Challenge 2: Noise in Glassdoor Reviews
Problem: Raw employee reviews on Glassdoor often contained excessive noise, slang, and irrelevant commentary.
Solution: Applied Google Gemini summarization to extract human-readable insights from employee reviews, focusing on sentiment, recurring themes, and leadership perception.
Challenge 3: Identifying the Right URLs Dynamically
Problem: For each company, Glassdoor and CrunchBase URLs differ and may include duplicate or outdated results from search engines.
Solution: Used an AI Agent (Gemini) to construct smart search queries and parse Google/Bright Data results, filtering for the most relevant URLs with higher accuracy.
Challenge 4: Generating a Cohesive Research Report
Problem: The extracted data was too fragmented, making it difficult for end-users to gain actionable insights.
Solution: Designed a Deep Research Builder Agent (Gemini) that merged data streams (Glassdoor insights, CrunchBase funding info, and employee sentiment) into a single Markdown-formatted research report.
Challenge 5: Maintaining Reliability in Multi-Step AI Orchestration
Problem: Errors in one step (e.g., failed Bright Data scrape) could break the entire workflow.
Solution: Implemented error handling and fallback prompts in Gemini Agents, ensuring that partial data could still generate a useful report instead of failing completely.
Top comments (0)