Recommendation Smart Asset

Recommendation smart assets use AML Recommender model outputs to personalize inbound and outbound realtime content. Recommender models provide a series of ranked recommendations based on a series of input data.

A recommendation smart asset's content panel contains the following:

Configure Outbound Support section: the following is displayed:

Note that recommendation assets are supported in Outbound Realtime and Inbound Realtime.

• Model Project: a mandatory Recommender model project, which supports Database table or Value list lookups. You can populate the property by browsing the file system for a Recommender model or using drag and drop. A Create New Item option is available. Note that only Recommender model projects are supported. For more information on model projects, please see the Model Project Designer documentation.

• Content Type: this dropdown allows you to define the type of lookup content that will be served by the smart asset. The following values are available:

  • Text (the default)

  • Image URL

  • Link URL

  • HTML

  • JSON

  • XML

Note that this property is output as ContentType in the payload returned from the RPI Realtime API, e.g.:

Image URL: '<img alt=\"RecSA02\" style=\"float:none; \" src=\"George\" />'

• Recommendation Field: allows you to select a field from the recommendation record output by the smart asset. The property is only displayed if the selected Model Project's Recommendation Lookup Type is set to 'Database table'. A mandatory dropdown field lists the fields sourced from the model project's Database Lookup table.

• Default Recommendation Rank: this integer property is set to 1 by default, and accepts a value range from 1 to 99. It allows you to specify the rank of the recommendation to be output by default by the asset. You can edit the embedded smart asset in the same context to change this default value.

• Default Value: this optional string property allows you to specify a default value to be returned by the smart asset when unable to return a recommendation lookup value. It accepts a maximum of 1000 characters.