Automation
Lead Handling Agent
Built an automated pipeline that scrapes job listings, deduplicates and enriches them, then uses AI to score legitimacy and route verified opportunities into a CRM. Runs unattended, with error handling and logging built in so nothing breaks silently.
The problem: A job platform sourcing listings for users seeking remote and international roles had no way to separate legitimate postings from scam or low-quality ones at scale. Manual review couldn't keep pace with volume pulled from Indeed, and there was no structured way to track which listings had already been processed. The result was inconsistent data quality and hours spent on manual checks that should have been automatic.
How it was solved: Built an n8n pipeline that pulls listings via an Apify scraper, deduplicates against existing Airtable records using position-based data matching, and normalizes fields from raw scraped data into a clean, structured schema. An AI layer then scores each listing for legitimacy, checking for scam patterns, missing company information, and salary inconsistencies alongside existing company rating and review data. Verified listings get written straight into Airtable with a trust score and reasoning attached, ready for the team to act on.
Solution template: The core pattern, source, deduplicate, enrich, AI-verify, log, generalizes to any workflow that needs to pull external data and filter it for quality before it reaches a human or a CRM. Swapping the source (a different job board, a lead list, a directory) or the AI verification criteria is a config change, not a rebuild. Error handling and run logging are built into the template itself, so any version of this pipeline reports failures instead of failing silently.