Help Center
Get started
Key concepts
Troubleshooting
Your workspace
Your account
Get started
Key concepts
Troubleshooting
Your workspace
Your account
What is a transformation and how to create your own
Transformations are the engine room of a workflow. They’re the step where raw input turns into something structured, enriched, or customized for your needs. Without transformations, workflows would just move data around as-is: useful, but limited.
With transformations, you can clean, enrich, and reshape data on the fly. Here’s how they work.
1. Choose the transformation mode
The first step in creating a transformation is selecting the mode.
The mode defines what operation should be applied to your data. There are many pre-built transformations available, each grouped by category and designed for a specific purpose. Most come with a small example (just click the question mark icon) so you can quickly see how they behave.
Some transformations work immediately with no setup, while others require additional configuration, such as specifying a rule, setting parameters, or writing an AI prompt.
Transformations are the engine room of a workflow. They’re the step where raw input turns into something structured, enriched, or customized for your needs. Without transformations, workflows would just move data around as-is: useful, but limited.
With transformations, you can clean, enrich, and reshape data on the fly. Here’s how they work.
1. Choose the transformation mode
The first step in creating a transformation is selecting the mode.
The mode defines what operation should be applied to your data. There are many pre-built transformations available, each grouped by category and designed for a specific purpose. Most come with a small example (just click the question mark icon) so you can quickly see how they behave.
Some transformations work immediately with no setup, while others require additional configuration, such as specifying a rule, setting parameters, or writing an AI prompt.
2. Define the inputs
Once you’ve picked a mode, the next step is telling the transformation what data to work with.
Think of a transformation like a factory: one or more ingredients go in, and one final product comes out.
Inputs are the data points you select to feed into the transformation. For example, if your trigger data contains an email address, you can point the transformation input to that value.
Some transformations require just one input, while others can accept multiple.
If you provide multiple inputs to a single-input transformation, it will automatically use the first non-empty value in your list.
This makes it easy to build fallback logic: “use value X, but if empty use value Y, and if that’s empty use value Z.”
2. Define the inputs
Once you’ve picked a mode, the next step is telling the transformation what data to work with.
Think of a transformation like a factory: one or more ingredients go in, and one final product comes out.
Inputs are the data points you select to feed into the transformation. For example, if your trigger data contains an email address, you can point the transformation input to that value.
Some transformations require just one input, while others can accept multiple.
If you provide multiple inputs to a single-input transformation, it will automatically use the first non-empty value in your list.
This makes it easy to build fallback logic: “use value X, but if empty use value Y, and if that’s empty use value Z.”
3. Add a clear name
Every transformation should have a readable name. This helps you, your teammates, and even your future self quickly understand what it does without having to open it up.
It may sound like a small detail, but naming transformations well saves a lot of confusion when building complex workflows.
3. Add a clear name
Every transformation should have a readable name. This helps you, your teammates, and even your future self quickly understand what it does without having to open it up.
It may sound like a small detail, but naming transformations well saves a lot of confusion when building complex workflows.
4. Test the Transformation
The final step is testing. Here you can provide mock data for the inputs you’ve defined, run the transformation, and immediately see the output.
This is especially powerful when working with AI-based transformations. You can iterate quickly by adjusting prompts, tweaking variables, and running tests until you’re confident the output is exactly what you need.
4. Test the Transformation
The final step is testing. Here you can provide mock data for the inputs you’ve defined, run the transformation, and immediately see the output.
This is especially powerful when working with AI-based transformations. You can iterate quickly by adjusting prompts, tweaking variables, and running tests until you’re confident the output is exactly what you need.
Building Complex Chains
Once you’ve built your first transformation, you can start chaining them together. The output of one can serve as the input of the next, creating layered transformations that handle even the most complex data processing needs.
From cleaning and normalizing values to generating personalized content with AI, transformations are what make workflows truly flexible and powerful.
Remember
A transformation follows four simple steps:
Choose a mode → Define inputs → Add a name → Test the output.
Master these steps, and you’ll unlock the ability to take any piece of data and shape it exactly how your workflow needs it.
Building Complex Chains
Once you’ve built your first transformation, you can start chaining them together. The output of one can serve as the input of the next, creating layered transformations that handle even the most complex data processing needs.
From cleaning and normalizing values to generating personalized content with AI, transformations are what make workflows truly flexible and powerful.
Remember
A transformation follows four simple steps:
Choose a mode → Define inputs → Add a name → Test the output.
Master these steps, and you’ll unlock the ability to take any piece of data and shape it exactly how your workflow needs it.