Data Enricher

What This Skill Does

The Data Enricher skill, developed by visualdeptcreative, is an automated lead intelligence engine designed to transform raw domain data into structured, actionable profiles for CRM systems like Notion. It functions by systematically searching for contact information through a prioritized hierarchy of sources, starting with website scraping, moving through Instagram social analysis, and utilizing the Hunter.io API for high-confidence email retrieval. It features an intelligent deduplication mechanism to ensure your database remains clean and efficient. The skill leverages local LLM processing (ollama/llama3.2:8b) for data normalization and formatting, ensuring privacy-conscious, consistent data output. The workflow is designed for high-volume batch processing, automatically generating localized JSON files before pushing data to your destination platforms.

Installation

To integrate the Data Enricher into your OpenClaw environment, ensure you have your Hunter.io API key ready and execute the following command in your terminal:

clawhub install openclaw/skills/skills/visualdeptcreative/data-enricher

Use Cases

• B2B Outreach: Identify key decision-makers and contact points for potential leads found in Meta ads or market research.
• CRM Cleanup: Populate Notion databases with verified, deduplicated contact information derived from social media handles.
• Sales Automation: Efficiently scale outbound efforts by automating the tedious task of searching websites and Instagram bios for valid email addresses.

Example Prompts

1. "Data Enricher, please process the list of beauty brands in the imports folder and update my Notion lead tracker with contact emails."
2. "Search for contact information for these 10 fashion brand websites and save the output as a JSON file in my workspace."
3. "Run a fresh enrichment cycle for the brands collected in my 'meta_ads' category, prioritizing founder-level contact emails."

Tips & Limitations

• Rate Limiting: Respect the 5-second interval between API calls; the skill is configured to batch 10 leads at a time to stay within Hunter.io limits.
• Confidence Thresholds: The system is hard-coded to ignore any email address with a confidence score below 70%, reducing the likelihood of bounced emails in your outreach campaigns.
• Deduplication: Always normalize your input lists. The skill automatically ignores existing domain keys to prevent duplication.
• Model Selection: For cost-efficiency, Hunter.io calls use Haiku, while complex data structuring is handled locally by Llama3.2:8b, ensuring zero inference costs for formatting tasks.