Enterprise Information Access and Navigation: "Connect To KnowledgeTM!"
What To Look For in an ENSR Solution
Relevancy and navigation. Enhancing individual and collective productivity means making the best possible decision faster, whether you are writing a proposal, making a go/no-go decision, answering a client request, assessing risk, etc. This requires finding the most relevant available content or advice quickly in the growing maze of IT systems and the network of "those who know" without missing valuable sources and without endless expensive pre/re-processing of unstructured information and refinement of search queries.
It is key, therefore, to genuinely "understand" the true meaning of content (semantics) regardless of language, idioms and vocabulary. Without semantic capabilities (i.e. just basic full text and statistical analysis), the solution will bring little value to the enterprise.
An employee must be able to describe in plain English (or any other language), using business or technical terms, what they want to know. The ENSR solution then returns, in real time, all relevant information from all available sources. Simultaneously, it presents a structured business-relevant vision of results, based on generic entities such as date, location, people and additional business-specific categories that are extracted dynamically based on business relevancy, such as organization, product line, projects, subsidiaries, etc.
This point is critical: if the response was just a long list of documents and sources, the information worker would still be faced with the challenge of processing each document and selecting the most relevant one, leading to saturation, lack of visibility and poor or slow decision-making, based on an incomplete data set.
Dynamically organized responses bring two key benefits to solve the "long list of documents" problem. One, they organize the response by categories based on business-relevant concepts, presenting unforeseen relationships and so extending the coverage of the response. This is critical in many business situations where a user has a problem to solve but does not know where and what to look for, such as a police investigation starting with field observation but no direction to lead the investigation, or a field engineer facing a new situation. This helps the user to make not only faster but also better decisions. Secondly, they organize information in categories specific to the enterprise’s organization and processes, allowing the user to easily navigate among results and rapidly narrow the research down to the most useful information (whether it is a person, a document or an application).
Every enterprise has, as part of its culture and IP, its own vocabulary. It is essential to provide the capacity to easily manage the evolution of specific terms and concepts, for they belong to the enterprise DNA.
User adoption. Beyond the intrinsic quality of the solution, user adoption depends on how fast the system responds, how intuitive it is to use and how easily a user or group of users can directly customize their environment while respecting IT policies.
Combining performance, user friendliness and ease of customization leads to adopting AJAX technology. This allows moving the presentation layer to the client, so minimizing bandwidth consumption, enabling drag-and-drop screen configuration and a rich, intuitive user interface.
Connectivity and access-control management. Extensive data-source connectivity must be supported by a broad set of native "off-the-shelf" connectors in a true "configure-to-deploy" mode, thus minimizing coding and implementation time and cost.
Native: Connectivity is an intrinsic part of the ENSR solution provided out of the box. This "one-stop shopping" approach resolves the traditional support, compatibility, performance and viability issues encountered with a multiple vendors solution.
Configure-to-deploy mode: No coding is required to set up the system, leading to low-cost connectivity and maintenance and rapid deployment of new sources. Beyond connectivity to vendor-packaged applications, one must be able to extend connectivity to home-grown and unsupported legacy applications. A generic connector covers that requirement.
Deep connectivity: The connectivity to a given application must leverage the most complete and best-supported API for that application. This ensures robustness, long-term maintenance and stability. The benefits are better performance and full access to the security layers as well as to the complete set of data and metadata. An alternative approach is also possible using standard remote connectivity, such as Web services. Such an approach leads to poor performance, limited data set access and incomplete or no security layer access.
A special focus must be made on security and access control. The ENSR solution must indeed be able to interface with a large variety of heterogeneous access control mechanisms without having to create an additional, specific access control directory.
It must leverage existing enterprise access control standards using them for single sign-on and map them to all other application specific access control layers and taxonomy in a non-intrusive fashion. Each user, once authenticated, is automatically associated with their respective access rights in each and every connected source. This approach allows quickly deploying and expanding to new data sources. Other approaches are error-prone and could lead to the violation of information access control policies.
Robustness and architecture. High availability is critical. Search and navigation that is on and off, or very slow, or has incompletely indexed target sources would be unreliable and soon be rejected. The ideal architecture is one where each critical component is interchangeable, fault-tolerant and can be dynamically load-balanced for performance and volume optimization.
The great challenge for an ENSR solution is to accommodate simultaneously large amounts of data, large numbers of users, provide fast responses and update information in a timely manner. In order to achieve this extreme scalability, components—whether they are connectors, indexers, search and retrieval engines or Web servers—must be deployable in a grid architecture or on a single server, depending on the requirements.
Leveraging these state-of-the-art components, this architecture guarantees the best performance available. A GUI-based administration and monitoring interface must be used to optimize the architecture with no reinstallation as the IT topology evolves.
The vision and requirements detailed above led Sinequa to create its Enterprise Information Access and Navigation solution. Sinequa’s patented semantic search technology, based on 200 human-years of research and development and embedded in a state-of-the-art architecture, offers an advanced multipurpose information access and navigation solution. This solution already serves global leaders and is a key element of their competitiveness. In 2006, Sinequa doubled its headcount and its revenues grew by 82%. Sinequa’s solution requires minimum effort to be deployed and maintained while delivering productivity by leveraging the totality of an enterprise’s information asset. Details of Sinequa’s multipurpose solution are provided in a technical Sinequa white paper available at
www.sinequa.com. For more information please contact sinequa@sinequa.com