This is a submission for the AI Agents Challenge powered by n8n and Bright Data

🚀 LinkedIn Growth Radar

What I Built

I built LinkedIn Growth Radar, a real-time AI agent that turns any public LinkedIn company URL into actionable business intelligence.

Instead of scrolling through pages or manually collecting data, this workflow automatically:

• Fetches live LinkedIn company data using the Bright Data verified node
  


• Analyzes growth signals such as follower increase, employee headcount, and hiring activity


• Scores each company with a priority index for business development and competitive monitoring


• Delivers digestible insights directly in Telegram, with messages chunked to avoid API limits

Use case: Sales teams, recruiters, and analysts can instantly qualify companies and spot high-growth opportunities — without leaving their chat app.

🎥 Demo

🔗 Live Demo Bot: @n8nsight_bot

Test it in real time: just paste a public LinkedIn company URL and get back a structured growth radar report.

📹 Video Demo Bot: Click to view the demo

🛠️ n8n Workflow

The full workflow JSON is available here:

👉 LinkedIn Growth Radar Workflow (GitHub Gist)

⚙️ Technical Implementation

System & Agent Setup

• 
  n8n AI Agent Node orchestrates the logic and reasoning


• 
  Function Node ensures Telegram output is split into safe chunks (<4000 chars)


• 
  Telegram Node delivers results to end users

Bright Data Verified Node

• We use the LinkedIn Company Profile extractor
  


• Input: company URL (e.g., https://www.linkedin.com/company/openai/)


• Output: structured JSON with fields such as employees, locations, followers, and about section

Redis Optimization

To reduce costs and improve performance, we implemented a Redis caching layer:

1. 
   Function Node → Hash the LinkedIn company URL into a cache key (e.g., linkedin:company:openai).


2. 
   Redis Node (GET) → Check if a cached result exists.


3. 
   IF Node → If data exists, return it directly. Otherwise, call Bright Data.


4. 
   Redis Node (SETEX) → Store new Bright Data results with a TTL (e.g., 24h).

This ensures:

• 🚀 Faster responses on repeated queries


• 💸 Reduced Bright Data credit usage


• 🔒 Consistency during a session or short monitoring window

Processing & Scoring

• Custom scoring logic weighs factors like size fit, headcount growth, hiring volume, ICP relevance, and engagement activity


• Generates a 0–100 score and classifies companies into Priority A, B, or C
  


• Outputs in a compact report, including engagement digest and growth signals

💬 Commands & User Experience

The Telegram bot is kept intentionally simple with two main commands:

• 
  

  /prospect → Analyze saved LinkedIn company URLs and return:

  
  
  
  
  
  • Company profile summary (industry, size, followers, growth signals)
  
  
  • Opportunity score (0–100) + Priority level (A, B, C)
  
  
  • Next best sales actions tailored to the company context
  
  
  
  
  
  

• 
  

  /content → Analyze a company’s recent public LinkedIn posts and return:

  
  
  
  
  
  • Top-performing posts in the last 30 days with engagement rates
  
  
  • Key content insights (themes & formats that resonate)
  
  
  • 3–5 fresh LinkedIn post ideas, each with title, hook, angle, bullets, and CTA
  
  
  
  
  
  

]

Flow:

1. User pastes one or more public LinkedIn company URLs.


2. Bot stores them in session.


3. User triggers either /prospect or /content.


4. Workflow calls Bright Data → Extracts structured company or post data → Normalizes it.


5. 
   AI Agent generates insights → Results are delivered in chunked Telegram messages (<4000 chars).

✨ Journey

The main challenge was to handle LinkedIn’s complexity: Bright Data only works with known URLs, so we designed the workflow around enrichment, not discovery. This made the tool perfect for analysts who already have target lists but need richer, real-time context.

Other hurdles included:

• 
  Telegram message limits → solved with a custom chunking function


• 
  Balancing signal weights → iterated on scoring logic until results felt useful


• 
  Clarity of insights → formatted reports with key metrics, scores, and digests


• 
  Avoiding redundant API calls → solved with Redis caching

What I learned:

• Bright Data shines as a stage-2 extractor once you have URLs


• n8n makes it easy to orchestrate AI agents and external data pipelines


• Redis is a powerful addition to manage costs and latency in real-time workflows


• Simplicity matters: the most powerful workflows are often the ones that users can trigger in one click (or one paste in Telegram)

Turns public LinkedIn company URLs into real-time business intelligence.