Orchestria provides a rich set of analysis methods to intelligently analyze every message, web and file activity. In addition to subject and body text, Orchestria performs a full textual analysis of more than 250 file and attachment types. Content analysis methods include:

Content
Keywords and Phrases	While of limited use in isolation, the presence or absence of certain keywords and phrases can be used to identify known entities such as company, product and service names, ticker symbols, boilerplate, disclaimers, etc. Orchestria allows expressions to be combined with wildcards and constructs (such as money or dates).
Semi-Structured	The identification of semi-structured data, such as account, credit card and Social Security numbers, product, service and procedure codes, cost centers, etc can be specified in a manner similar to Regular Expressions, but operating many times faster. This technique can extract identified items (with optional obfuscation of credit card numbers) as metadata.
Statistical Analysis	Uses a scoring mechanism based on positive and negative indicators, including 'must [not] have' and 'should [not] have' fields. Fully supports date and money constructs, and provides automatic scoring adjustment based on document size. This technique can also extract selected indicators (including scores) as metadata.
Context
Context Based Analysis	This is an advanced analysis technique that looks at the “intent” of the message, web or file action in question.
Conceptual Understanding	This technique uses linguistic pattern matching to identify a electronic action violation. It is particularly useful in situations where domain terminology is frequently used, such as financial trading, insurance, and medical environments. Additionally, it may include file size, attachment size, as well as the presence or absence of information.
Learn by Example	Known by a variety of names, including Bayesian Inference, this pattern matching technique allows a representative document set to be analyzed for common characteristics, which are then used to build a policy which detects similar patterns in documents presented at a later date. Unlike other vendors who employ a 'black box' approach, Orchestria fully exposes the policy and allows it to be tuned, for example to ignore irrelevant patterns.
Transactional	This technique identifies monetary transactions, such as quotations, orders, invoices, payments and the like. It analyzes the proximity of terms and figures, as well as the mathematical relationship between figures in a document. Combined with other analysis techniques described above, this not only allows documents to be identified, but also for values to be automatically extracted as meta data.
Person
Who	This third facet of intelligent analysis determines who is performing the electronic action and where or to whom it is destined to go. Using Active Directory/LADP, policy can be applied appropriately based on the parties involved in the electronic action. This technique significantly reduces the false-positive rates traditionally experienced by lexicon based systems.