Configuring Smart Mapping settings

Set up Smart Mapping settings to receive mapping recommendations that are tailored to your requirements.

1. In webMethods Integration, click Settings > Preferences.
2. Select Use community data for predictions to get mapping suggestions based on community insights. If selected, the field mappings in your flow services and workflows and other user's flows are anonymized and aggregated to provide mapping suggestions to all users. To benefit from community suggestions, enable this option.
3. Set the prediction score by using the Default prediction score slider.

   Default prediction confidence is the initial confidence level that the Smart Mapping recommendation engine assigns when it suggests mappings between source and target fields. This confidence is based on historical patterns, structural similarities, and aggregated community insights.

   Confidence levels are typically categorized in the range of 0 to 1, helping users prioritize which recommendations to review and accept. This value becomes the baseline for your recommendations. For example, setting a score of 0.7 indicates that the recommendations higher than 70% confidence are shown. The Default prediction score value takes precedence over the value set in the canvas page.

4. Select App and field names to enable context for the Smart Mapping engine to deliver more accurate mapping recommendations.
   • An app represents a client application that is configured in webMethods Integration that interacts with integration endpoints. The app provides contextual information about the integration scenario, which helps the AI refine mapping recommendations based on the type of services and data structures that are associated with that application.
   • A field name identifies a data element within a source or target schema. Smart Mapping indexes field names and their hierarchy to provide structural context for mapping suggestions. By analyzing field names and their position in the hierarchy, Smart Mapping improves mapping accuracy for complex data structures.
5. Select Field hierarchy in the tree structure, then suggestions are provided considering only the leaf part of the field hierarchy. By default, this option is selected.
   A field hierarchy represents fields that are organized in a tree-like structure, where each node corresponds to a field, and child nodes represent subfields or nested properties. The leaf is the last segment in the path. If the hierarchy is Connector.cc.name, only Name is considered to provide suggestions. Parent nodes (Connector, cc) are ignored during suggestion lookup.
6. Enable One-to-many mapping recommendation if you want the Smart Mapping to intelligently associate a single input field (like a user profile, product, or query) with multiple relevant outputs. This setting is referred to as one-to-many mapping, where one source can produce several recommendations. The process relies on machine learning models, contextual analysis, and historical data to generate diverse and personalized suggestions.
   If disabled, the system supports only one-to-one mappings, meaning that each input corresponds to a single output. This limitation affects the diversity and effectiveness of recommendations.
7. Enable LLM for more context aware recommendations, then Smart Mapping uses AI large language models to generate intelligent, personalized, and context-sensitive suggestions based on a deep understanding of user input, history, preferences, and surrounding context. By default, this option is enabled.
8. Click Apply.
   The Smart Mapping settings are enabled, and the mappings are recommended when you are building integrations. Updates to recommendations occur at scheduled intervals and might not reflect immediate changes.
9. Optional: Click the Restore defaults button to restore the initial configurations.