Message Image  

App Connect

 View Only

IBM App Connect: Developing Integrations with Mapping Assist 

Fri July 31, 2020 05:35 AM

In the world of integration, data mapping is a key step since the domain itself deals with a wide variety of aligned tasks like data transformation and data mediation between various heterogeneous data sources and similar destinations. But integration developers spend a great deal of time and energy in just getting mapping of fields between different endpoints right. The reason is that this continues to be complex and time-consuming as the type and number of endpoints grow. This has the potential to slow the speed at which business can respond to changes and remain competitive.

Mapping Assist introduces Artificial Intelligence (AI) capabilities in IBM App Connect to accelerate the speed of development, shorten time to value, and improve overall user experience. Mapping Assist uses a pre-trained AI algorithm to provide users with intelligent, customized data map suggestions that are private to your business.

Now let’s discuss about the key features of Mapping Assist, with some real time use cases and see how it helps to reduce your development and deployment time.

Key features of Mapping Assist:

  • Auto map all fields wherever the matching confidence is higher than 80%.
  • Relevant and fine-grained field level suggestions are provided when confidence is between 30%-100%.
  • Suggestions are provided based on multiple sources, including all the previous nodes in the flow, for greater accuracy.
  • Mapping suggestions for flat structures as well as complex, nested mapping fields are provided.
  • Previously selected mappings are remembered and shown as top suggestions when a similar source and target mapping is attempted.

The Sync the Salesforce Contacts with Microsoft 365 scenario is used to demonstrate this.

Figure 1: Sync the Salesforce Contacts with Microsoft 365

Once the Microsoft 365 Create Contact node is added, mapping suggestions will be generated and best possible matches (top suggestions) are presented which can be automatically inserted into fields with a single click.

Figure 2: Top suggestions with Insert suggestions

These top suggestions have an 80% (or higher) accuracy rate, and the count (N suggestions) identifies the total number of fields that will be populated with mappings. Mapping Assist Insert suggestions won’t count or apply mapping for the fields which have:

  • Existing mappings that are manually populated.
  • Multiple suggestions for a field with same confidence rating. In that case, all suggestions have an equal chance to be mapped. Those suggestions will be provided in field level suggestions and can be mapped manually by selecting an appropriate mapping.

By clicking on Insert Suggestions, all suggestions with a confidence score of above 80% will be automatically inserted into the Microsoft 365 Create Contact node.

Figure 3: Populated mappings for simple and nested fields

Figure 3, demonstrates all the fields that are automatically mapped by Insert Suggestions.
Note the suggestions for simple fields as well as the nested fields.

For remaining unpopulated fields, a click on textbox will present the suggestions if Mapping Assist found suggestions for that field in the form of a Suggested mappings list. This displays up to five suggestions with percentages that are above 30% from all previous nodes.

Figure 4: Field Level Suggestions for a simple field

Figure 5: Field Level Suggestions for a nested field

In figure 5, suggestions are from two nodes - For each node and Salesforce node. For each mapping, the field hierarchy identifies the origin of the content, such as the application or node name, action and object.

If you wish to map the fields that are not suggested by Mapping Assist, you can do so by clicking on All mappings, that displays all inputs available for the target field (figure 6).

Figure 6: All available mappings are shown for a target field

Clicking on Suggested mappings will switch back to the Suggested mappings list.

To improve accuracy in future mappings, Mapping Assist collects and stores the mapping data in an internal database by tracking the mapping history of flows that are started. For example, if you mapped Birthdate to Birthday, this mapping will be remembered and Mapping Assist will suggest it in future flows with a 100% confidence rating and be mapped any time a user chooses to Insert Suggestions.

Figure 7: Suggestion from user history

Mapping suggestions for nested array fields are supported for exact schema matches only. This means root element names should exactly match in the source and destination array and fields that have the same field name under the root element will be mapped.


To illustrate this feature, let’s consider a scenario to Sync data from one account to another account in Microsoft 365.

Figure 8: Sync data between two different accounts

Clicking the Insert Suggestions will populate the fields with mappings, including nested array fields.


Figure 9: Populated mappings for nested array fields

You can also use the field level suggestions available for nested array fields as shown in Figure 10.


Figure 10: Suggestion for child element field in nested array fields

IBM App Connect Mapping Assist reduces integration flow building time and enables you to build flows confidently.