AI Data Enrichment Agent

When to use

Use this skill to batch-process records and enrich them with LLM-generated fields such as summaries, sentiment, categories, or extracted entities.

Instructions

1. •Connect to the data source (Supabase, Airtable, or CSV)
2. •Identify the fields to enrich and the target output fields
3. •Build the prompt template for each enrichment task
4. •Process records in batches of 20 to stay within rate limits
5. •Call the LLM API and parse structured output (JSON mode)
6. •Write enriched fields back to the source record
7. •Log token usage and cost for the enrichment run

Environment

• •Runtime: python-3.12
• •Trigger: API
• •Category: Data and AI Agents

Examples

• •"Classify all supplier records by industry and add a risk score"
• •"Generate product descriptions for all catalog items missing summaries"