AI News Automation Bot

I built an end-to-end AI-powered automation pipeline that scrapes, summarizes and rates the latest AI news twice daily with zero manual effort. Everything lands in a clean Google Sheet ready to scan in under 2 minutes each morning.

This project demonstrates my ability to design automated workflows that combine web scraping, AI-powered content processing and structured data storage into a single reliable system.

The bot runs on a scheduled cron job twice a day. It pulls the latest AI articles from sources like TechCrunch, The Verge, MIT Technology Review and VentureBeat. Each article gets passed through OpenAI's API which generates a concise 2-3 sentence summary. The system then assigns a relevance score from 1 to 10 based on how useful the news is for small business owners. Everything gets logged into a Google Sheet with the date, title, summary, source URL and relevance rating.

The scraping layer uses BeautifulSoup and Requests to pull article data from multiple sources with error handling for rate limits and site changes. The AI summarization layer calls OpenAI's API with custom prompts engineered to extract business-relevant insights not just generic summaries. CrewAI orchestrates the workflow as a multi-agent system where one agent handles scraping, another handles summarization and a third handles relevance scoring. Each agent has a defined role and goal so the pipeline stays modular and easy to extend. Flask serves as the lightweight backend that exposes an endpoint for the cron trigger. The cron job fires via cron-job.org hitting the Flask endpoint on schedule.

Processes 15-20 articles per run in under 3 minutes

Runs reliably twice daily with 99%+ uptime on Render

Summaries average 40-60 words keeping them scannable

Relevance scoring tuned for small business decision-makers

Python 3.12 with CrewAI for agent orchestration. Flask as the trigger endpoint. OpenAI API for summarization and scoring. BeautifulSoup and Requests for web scraping. Google Sheets API for structured storage. Deployed on Render with scheduling via cron-job.org. All credentials handled securely through environment variables and base64-encoded service account keys.

Multi-agent AI workflow design using CrewAI

API integration with rate limiting and error handling

Automated scheduling with external cron services

Clean data pipeline from raw web content to structured output

Production deployment with secure credential management

Need an automation system that saves you hours every week? I build AI-powered bots and pipelines that handle repetitive research, data collection and content processing so you can focus on decisions not data gathering. Let's talk.