The Transform Step allows you to use AI to intelligently modify, analyze, or reformat data as it moves through your workflow. Instead of writing logic or code, you can describe the transformation you want, and the AI generates new or updated data based on your workflow inputs.

This makes it easy to format content, generate new values, summarize text, extract information, adjust tone, validate content, and more.

What the Transform Step Can Do

The Transform Step supports a wide range of AI-powered operations, including:

Data Transformation

• Converting formats such as capitalization, currency, or date formats
• Combining or splitting fields
• Changing data types (string to number, date to number)

Text Processing

• Summarizing content
• Generating descriptions, titles, or categories
• Translating content
• Adjusting tone or rewriting text
• Extracting entities such as names, dates, and locations

Analysis

• Sentiment analysis
• Categorization or classification
• Identifying patterns or inconsistencies
• Comparing data points
• Assessing content quality

Logic and Reasoning

• Conditional transformations
• Multi-step reasoning
• Rule-based restructuring
• Validating or cleaning input text

How to Use the Transform Step

1. Add the Transform Step

Click Add Step in your workflow and search for Transform Step.

Give your step a name if you’d like.

2. Open the Step to Edit

Click the Transform Step to open its configuration panel.

3. Write Your Prompt

The Prompt is where you describe the transformation you want the AI to perform.

You can:

• Write natural-language instructions
• Insert data fields from previous workflow steps (Forms, Documents, Search Data, etc.)
• Refer to multiple fields in the same prompt

Example instruction:

“Create a new ID number using the combination of fields.”

4. Configure Output Fields

Below the prompt, define the fields that should be generated by the transformation.

For each output field, you will select:

• The field name
• The data type (String, Number, or Boolean)

These outputs become part of your workflow’s data and can be used in future steps.

About Output Field Types:

• String - Use this for text-based values such as names, descriptions, categories, or IDs.
• Number - Use this for numeric values such as scores, counts, totals, or calculated values.
• Boolean - Use this for true or false values. This is helpful when the transformation determines whether a condition is met.

Choosing the correct data type helps downstream steps interpret and use the transformed value properly.

5. Test Your Prompt (Recommended)

Testing allows you to preview what the transformation will look like using sample inputs.

To test:

• Provide sample values for any mapped fields
• Click Test
• Review and adjust the output as needed

While testing is not required, it is strongly recommended to ensure reliable behavior during live workflow sessions.

6. Save the Step

Click X to save your configuration.

Your new output fields from the Transform Step will now be available for use in all downstream steps such as Documents, branching logic, Deliver Data, and additional Transform Steps.

Considerations When Using the Transform Step

Workflow Structure Considerations

• The Transform Step works with existing workflow data types and cannot create new object types
• Each Transform Step performs one transformation; multiple transformations require multiple steps
• Separate steps help maintain clear data mapping and support referencing intermediate outputs

AI Behavior Considerations

• The AI can perform basic math but is not designed for complex calculations
• It cannot access external systems or APIs
• Outputs may vary slightly between identical inputs
• Accuracy may differ for highly technical or specialized subjects

Output Field Considerations

• Output must fit within the limits of the destination field
• Mapping long text into fields with short character limits may result in truncation
• If needed, guide the AI with length constraints such as “Limit output to 250 characters”

Technical Considerations

• Transformations typically take between 2–30 seconds depending on complexity
• AI cannot process binary files such as images or PDFs without preprocessing

Summary

The Transform Step brings AI-driven data transformation directly into your workflow, enabling dynamic updates, intelligent formatting, content generation, and more. By using natural-language instructions and workflow data, you can enrich and automate your processes with ease.