Enrich CSV with AI-Generated Columns

How to Use This Prompt

This prompt is called with two parameters:

1. Enrichment directive (required): Description of the new column to add (e.g., "estimated GDP 2025", "population in millions", "capital city")
2. Filter (optional): Description of which rows to process (e.g., "European countries only", "countries with population > 50M")

When this prompt is invoked, Zo will:

1. Read a handful of rows from the CSV to understand the current schema
2. Generate a Python script tailored to your specific CSV and enrichment request
3. The script will prepare the CSV by adding any new columns before processing begins
4. Execute the script to process the CSV in batches, making sub-API requests to get structured output for each row
5. The Python script will write the enriched data back to the CSV immediately as each API call completes (not in one big transaction at the end)
6. The user's enrichment directive may be either very simple or very detailed, specifying specific tools to use etc. When translating their ask from a chat into the script, you are free to reword their initial query for clarity and completeness into the API prompt, but never omit any special requests that they included in their original ask. Your job is to write the best possible prompt into the script for the API request to fulfill the user's original intent with quality and accuracy. Each API request will only process one row at a time and will as a Zo sub-agent to return the requested data, after using any of the tools and capabilities that Zo has access to on this host. You can include specific instructions in the prompt to guide the API on how to get the data, if needed.

Example usage: