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
- Connect to the data source (Supabase, Airtable, or CSV)
- Identify the fields to enrich and the target output fields
- Build the prompt template for each enrichment task
- Process records in batches of 20 to stay within rate limits
- Call the LLM API and parse structured output (JSON mode)
- Write enriched fields back to the source record
- 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"