Learn and master Clay with Clay university

In the last lesson, we talked about the two ways you can use AI to transform data: deterministic and generative. Now we're going to focus on the first one: deterministic transformations using AI Formulas in Clay.

These are perfect for data cleanups, text extraction, and other repeatable operations where you want consistent, rule-based output.

🧮 When to Use AI Formulas

Anytime you're manipulating or calculating existing data in a structured way like:

• Totals
• Patterns
• Extractions
• Cleanups

You should use AI Formulas.

🔍 Ideal Use Cases for Deterministic Tasks

These are ideal for deterministic tasks when:

• You know exactly what kind of result you want
• You're applying the same logic to every row
• You don't need the AI to make a judgment call

Here are a few classic examples. You want to:

• Extract all the companies someone has worked at into a comma-separated list
• Calculate the total years of work experience someone has
• Get the tenure in their last role, expressed in years and months

This kind of transformation would normally take a bunch of custom formulas or scripting. AI Formulas make it simple.

📝 Use Cases for AI Formulas

You can use AI formulas for:

• Formatting dates or numbers
• Extracting text patterns (like pulling a domain from an email)
• Cleaning up inconsistent inputs like job titles or locations
• Doing math, like adding up years of experience or calculating tenure

If it's structured and consistent, AI Formulas are your best bet.

⚡ Why AI Formulas Optimize Your Credits

AI Formulas are particularly valuable because they:

1. Don't consume Clay credits; they're free to use
2. Process faster than API-based enrichments
3. Maintain consistent output formats
4. Can be reused across multiple tables and workflows

By using AI Formulas for deterministic tasks instead of credit-consuming enrichments, you can reserve your credits for more complex operations that truly require generative AI.

🔮 What's Coming Next

In the next lesson, we'll cover the other side of the Transform equation. We’ll use generative AI to take messy or unstructured inputs and turn them into clean categories or classifications.