In xCP 2.0, Documentum Content Intelligence Services (CIS) is available out-of-the-box. To use CIS, simply decide what you want found in your unstructured content and xCP 2.0 does the rest. With out-of-the-box taxonomy-based categorization, NLP-based entity extraction and rules-based metadata extraction, analyzing unstructured content in xCP applications has never been easier. Let's walk through an example.


In xCP 2.0, I can specify the information I'm looking for in unstructured content much like choosing which attributes to make part of the content model. In the image below (left), I'm modeling a DailyReport - a document with attributes such as title, author, date, etc. These are typical Dublin Core like attributes. I also want to characterize my DailyReport model with attributes called Business Terms (USA), Organization, People, and URL. In xCP 2.0 these attributes are called Discovered Metadata. Discovered Metadata is valuable information that is "hiding" in the content waiting to be discovered and characterizes my content (adding context) and ultimately allowing me to make optimal decisions.


DailyReport attributes.bmp DailyReport context data.bmp

NOTE: These images are mainly for illustrative purposes only - they are taken from work in progress and may not be representative of the final product.

The window on the right is the Context Data Palette. When modeling a document, the Context Data Palette allows you to choose the attribute Data Types and Discovered Metadata. It also allows you to specify Relationships between this document (DailyReport) and other objects in the application (other documents or business objects, for example)


When the application is deployed, xCP 2.0 generates configuration files that inform CIS to analyze this type of content. For example, when an instance of the DailyReport comes into the system, it is tagged with URL's, Organizations, People and Business Terms.


Another benefit is that any place in the application you want to display, list or search DailyReport instances and its attributes, the Discovered Metadata is also available. Once configured, Discovered Metadata works like other attributes in xCP 2.0. By nature, Discovered Metadata types are repeating value and can be:


  • displayed with the content just like other attributes
  • displayed as output columns in a list of content items
  • used as user inputs or predefined inputs in a search
  • used as facets for faceted navigation


These last two points are of particular interest. The benefits of content analysis can be framed from several angles - and you'll get different answers depending on who you ask. From an end-user perspective, the primary benefit of content analytics is to characterize (e.g. tag) content with context so that it can be found later with efficiency and precision. Searching by context and leveraging it in faceted navigation serve this purpose directly.


With xCP 2.0, achieving the benefits of content analytics has become seamless and easier to leverage.


Please feel free to reply with thoughts/reactions/questions! I'll try to respond as soon as feasible.


Thank you.