ONLINE REVIEW and ANALYSIS
Stroz Discovery’s online review platform, Stroz Review, is a hosted database repository that enables counsel, litigation support personnel, and their clients to review their documents securely from multiple locations. Users only need their web browser to navigate easily to our website and do not need to install specialized software to review Electronically Stored Information (ESI) that may be in non-standard formats. Stroz Review allows users to quickly connect to a secure, hosted Oracle database and review documents from anywhere in the world. With 24-hour client support and 99.5% system availability, Stroz Review is a robust and user-friendly platform that stands out dramatically in the eDiscovery field by virtue of our advanced text analytics and QC mechanisms that reduce the time, manage the costs, and improve the quality of eDiscovery outcomes. Many of Stroz Review’s features are described below. • Stroz Extract performs automated objective coding derived from document text. This artificial intelligence technology employs targeted data extraction to recognize, extract, and automatically code bibliographic metadata such as document type, document date, title, etc., from the content of the document, along with individual and organization names, dates, locations, and addresses. All of this information is searchable, enabling the reviewer to find key information in both electronic and paper (scanned) documents. Stroz Extract then displays the information in a searchable, highlighted format, which allows users to speed through first-level review by finding the right information quickly and efficiently. • The Intelligent Privilege Filter incorporates sophisticated search technologies that target key textual features of privileged documents to refine the search process, while identifying the universe of potentially privileged documents for manual review. By classifying all such documents into recognizable patterns, highlighting key sections of the document, and ranking the likelihood of privilege, reviewers can significantly accelerate the review process, while simultaneously increasing the confidence level and defensibility of their determinations. • Near Duplicate Detection identifies and displays documents which are syntactically similar to a specified document. Honing in on text, it ignores formatting and file type (.doc, .pdf, etc.). The degree of similarity is dynamically adjustable by the user, who can also display the differences between the near duplicate documents. • Foreign Language Support enables reviewers to conduct English or foreign language searches, with automatic language translation. The Stroz Review platform is fully Unicode compliant, including support for CJK languages. This support includes automatic language identification, OCR (including OCR of mixed alphabet documents), and machine translation. Machine translation is available on a bulk basis during processing or on demand during review. • Email Thread Analysis and Visual Representation enables reviewers to identify emails that are part of a thread and to group and tag threads for easier, faster review. The structured representation of email threads allows bulk coding of documents in any portion of a complete thread. • Customized Review Options facilitate project management and document review with broad capabilities for distributing, managing, and directing reviews, assuring the separation of attorney work product among different firms through a single database, drastically reducing data management costs. • Stroz AutoSuggest helps reviewers make quick work of large unstructured data sets by automatically grouping like documents. It relies on an initial sample of documents that have been previously classified in some manner, either manually or by application of Stroz Elevate. This “learning set” is run through a statistical engine (presently the Oracle Support Vector Machine) which performs a textual analysis of the documents to create a model embodying the criteria used to classify the initial learning set. This model is then applied to the larger set of documents subject to review for automated classification. Among its several applications, clients can use AutoSuggest to perform bulk classification of document sets for prioritization or to route various types of documents to specific reviewers; to suggest coding to reviewers as they review documents; or, as a quality control function, to perform behind-the-scenes classification and designate for second-level review those documents whose manual and AutoSuggest coding differ. • Stroz Elevate uses the logic described in attorney subjective coding manuals to run sophisticated processes that can partially or fully automate first-level coding. It dramatically reduces the time and cost of first-level review by applying Stroz Extract’s objective coding methodology to optimize subjective coding. Far more than a keyword search, Stroz Elevate performs sophisticated pattern matching and uses decision tree logic to quickly, consistently, and cost-effectively identify relevant documents. An iterative process of running Stroz Elevate, examining results, and adjusting Elevate logic is used to arrive at the most satisfactory result. • Event Extraction analyzes the document for events (i.e., who did what to whom when) which are identified by parsing the document text and identifying Actions, Actors, Acted On’s, and Dates. This facility enables high precision searches on any combination of Action (optionally including synonyms), Actor, and Acted On. Events identified for a document are displayed with the document’s metadata; clicking on an event highlights its location in the document. • Themes and Gists provide further text analytic sophistication. Themes, derived with the assistance of a knowledge base that can be edited for a specific domain, are searchable and can be used for conceptual grouping of documents which are then foldered and ranked for any number of designated themes. Gists provide either a generic document summary or a summary from the perspective of any displayed Theme. • Comprehensive Search/Redaction employs easy-to-use advanced query tools for many different kinds of logic searches—both Boolean and natural language—with numerous options for setting search parameters and a broad array of redaction tools for user-tailored flexible editing. • Exception, Workflow, and Administrative Reports can be generated on demand or at scheduled intervals across all metrics. Deep reporting on exceptions, the option for password cracking on protected files, and real-time monitoring of reviewer activity are among the available features. • Data Privacy and Security Assurances are provided with the pioneering Mobile Processing Unit (MPU) which brings the Stroz Review platform and its analytical tools to any global location that client needs may demand. Rapidly deployable in any venue, the MPU keeps data within physical country borders and provides a secure processing environment for resident reviewers, thus delivering ironclad assurance of compliance with country-specific ESI privacy regulations. Safe Harbor-certified, Stroz Friedberg is thus able to perform all phases of eDiscovery for an international client base that must adhere to disparate data privacy regulations arising from a multinational business environment.
