Creating and Understanding Custom Targets

1. Accessing custom targets

Go to https://app.mappingclarity.com/custom-targets.

2. Choosing a target type

Choose a target type based on the data you are mapping against.

List + Definition

This is the most powerful option because it gives the AI the most context.

• Use this for taxonomies, benchmarking categories, or account codes where values have written rules or explanations.
• Definitions help the AI make smarter and more confident mapping allocations.
• The upload file must have headings in row 1.
• For multiple classification levels, repeat the higher-level value in earlier columns and place the more specific item in later columns.

List Only

Use this when you have a static set of values but no formal definitions.

• This works well for lists such as vendor names, product IDs, or internal codes.
• The algorithm maps uploaded data only to values from the provided list.
• The file structure is the same as List + Definition, but definition columns are omitted.

Pattern

Use this when the AI should learn how to extract or normalize information from contextual patterns.

• This can support tasks such as extracting a specific value from text or normalizing inputs based on examples.
• Uploading known inputs and correct outputs is highly recommended so the AI has a contextual reference.

3. Creating the custom target

Provide a custom target name, select the target type, and upload the structured file when applicable.

After creation, the custom target is available when you create or edit a data pipeline.