Over 85% of rich contextual enterprise information is in unstructured form and has doubled over the past three years according to recent industry research reports. While structured data provides answers to “when”, “who” and “what”, unstructured data answers “how”, “where” and “why” by uncovering situational contexts around events that helps provide essential cause, complaint, and correction knowledge. An enterprise’s core business intelligence is hidden in such unstructured forms (various textual or image formats), and decision support will remain incomplete without unless the gap between unstructured text and structured enterprise data is bridged.

Regular search engines’ basis of keyword search does not meet the unstructured information analytics challenge. A recent 2004-2005 IDC Content technologies Study reported that on an average an employee spends 3.5 hours every week on searches that fail to locate the desired document and another 3 hours every week to recreate content that wasn’t found. The annual salary costs alone for extracting required decision making intelligence to a 1000 employee enterprise could be more than $9.7 million.

But this astronomical expense is only one part of the story; other significant areas of impact include extended wait-times due to the “switch-time across types of data” for business decision support, and out-dated or incorrect information.

Significant challenges of processing unstructured information stem from:

	Enormous volumes
	Multiple formats
	Multiple languages
	Differing grammar and vocabulary
	Multiple situational contexts around events
	Skills Gap (not everybody is a mathematician/statistician, and a statistician does not know pharmacology or medicine in a healthcare enterprise, for instance.)

Enterprises Information Retrieval as a specialized technology domain is characterized by the following:

	Business requirements are specialized and information is role-dependent
	Each business has unique requirements over and above the expected impact on culture change and using, securing, and presenting information uniformly & accurately Each employee has a different semantic perception of the information desired, and uses it differently, depending on professional requirements.
	Delivery of accurate information is critical
	Multiple search methods (pre-categorized, advanced full-text search, spelling suggest, relevance tuning, etc.) are important to enable the user to get to the right answer quickly Delivery of information should be compliant with Corporate Security Policies and Procedures
	360° view of Enterprise Knowledge
	Multiple repositories with structured and unstructured information in multiple formats and languages need to be “locatable”, “findable” and “correlated” for effective decision support

PluralSoft recognizes that enterprises need a comprehensive knowledge recognition-extraction-visualization framework that is driven by domain-specific concept-and-relationship precision over time. The significance of the ability of such a framework to turn unstructured information into effective business intelligence is something that cannot be underestimated in the global business scenario. PluralSoft addresses this need of bridging the gap between structured and unstructured analytics with its unique, multi-tier, multi-platform operable (operating system, hardware and database agnostic), flexible ontology-driven knowledge recognition-extraction-visualization framework called Ontos.

To know more about PluralSoft’s Ontos solution and the value we have delivered to our clients, please contact us.