by: Greg Leighton, PhD, Palomino Inc.

Business intelligence (BI) is all the rage, as organizations seek to improve their operational efficiency, client relationships, and decision making through computer-assisted analysis of internal data. Currently, people chiefly associate BI with structured data repositories — such as data warehouses – but what about documents? It turns out that they constitute a rich source of information as well.
Let’s start by considering a document’s metadata: its author, date of creation, and other basic attributes. Using such information, it becomes possible to establish basic relationships between documents and users. Taking things a step further, one can incorporate historical information from activity logs to paint a picture of how that document has changed over time, and who is responsible. It now becomes possible to establish patterns in how an individual user interacts with documents (person-centric workflows) and, additionally, to discover the most common sequences of actions that are performed on a given document (document-centric workflows).

Once collected, both types of workflows can assist an organization in answering important questions about their business processes. Person-centric workflows can be used to identify those tasks an employee spends the bulk of their time doing, together with common sequences of actions they perform. The greatest gains in employee productivity are then most likely to come from focusing efforts on improving that employee’s ability to carry out these tasks. Document-centric workflows, on the other hand, provide valuable insights into the nature of collaborative activities conducted across the organization. They facilitate the discovery of redundancies within business processes, determination of which employees most frequently work together, and verification that industry regulations and other formal requirements are being correctly observed by employees.

Up to this point, we’ve neglected perhaps the most valuable information associated with a document –its content. By utilizing natural language processing techniques, it becomes possible to extract the key concepts, actors, and relationships from document text, to classify documents by topic, and to identify which documents in the system are closely related. Finding relevant needles in the document haystack is greatly simplified.

Of course, realizing such benefits requires an “intelligent” document management system, capable of mining the most valuable information from documents and presenting it in an intuitive way to the end user.

Over the coming weeks, we’ll be following up with a series of posts discussing various aspects of making document management more intelligent. Stay tuned!