New Approaches to KM in Government User-Centric Enterprise Information Retrieval

Today's $12 billion knowledge management market provides comprehensive solutions to large organizations—implementing such solutions requires significant resources and depends on behavioral changes by employees. As a result, KM efforts often fall short of their goals.

This paper describes a new user-centric approach to managing structured and unstructured organizational documents and data. Focusing on end-user features offers significant benefits to government agencies, their employees and their IT managers.

The Problem: Capturingand Interfacing Data

In the current approach to KM, capturing the structured and unstructured data in an enterprise or government agency is a complicated and difficult task since information resides in a variety of:

• Formats—documents, messages, relational data, Internet and intranet content and multimedia;

• Systems—ERP, financial, human resources, customer relationship management and content management; and

• Locations—enterprise data and file servers, departmental network/shared drives and employee desktops and laptops.

Each business function or information system has its own requirements for accessing and manipulating data. Additionally, each has widely different user interfaces and storage systems with varying degrees of access by the rest of the enterprise. As a result, an enterprise- or agency-wide KM solution contends with such a disparate collection of silos and stovepipes that the eventual system makes too many compromises to be of much use to anyone.

Currently, KM vendors typically work with an IT department on a hugely complex, expensive and labor-intensive effort to build a system capable of the full range of enterprise data management tasks.

Current KM solutions often fail in the following ways:

1. Inability to find and act on data, hampering an agency's ability to identify, analyze or respond to dangers, opportunities or constituent needs, especially in times of crisis;

2. Lack of mobile solutions, preventing data access by employees in the field;

3. Extensive and complex user behavior and business requirements, leaving KM systems only partially utilized;

4. Lack of extensibility and flexibility, failing to accommodate legacy, existing and future data sources;

5. Poor search of local files or e-mail, often obscuring the most relevant data;

6. Lack of scalability, making the addition of data sources and systems challenging for IT; and

7. Need for management, training, and deployment, resulting in high ongoing cost of KM system.

Although most enterprise and government organizations are now pursuing knowledge management, KM technologies have not yet realized significant impact on organizational productivity. Several factors cause this failure, including the methods of data capture and management, as well as interface and utilization.

A New Approach: Search-based User Interface

A new approach to the information retrieval component of knowledge management assumes that companies already have legacy data and file storage and management systems. Instead of replacing those systems, a user-centric approach seeks to make existing data more immediately searchable, actionable, useful and ultimately valuable by adding search-based user interfaces. Understanding the significant investments already made in KM systems, new search companies seek to give options to organizations by making existing systems work the way they want them to work, not the other way around.

In the presence of existing, largely taxonomy-centered systems, the new user-centric search interfaces are silo and stovepipe "agnostic." They provide connectors into existing sources of structured and unstructured data. Software development kits (SDKs) provide for customization to federate results from all available data sources.

Competing vendors who provide this user-centric approach will differentiate themselves in terms of how multiple data sources and types can be located and leveraged. Increased accessibility and usefulness of data, as well as enhanced interface, will increase the ROI of these vendors' solutions.

New search technology is information processing in a real-time environment. The traditional way of finding data is for the user to launch one application at a time, enter keywords and/or other search parameters and hit enter. Frequently, retrieving all the information needed by a user requires repeating the process within multiple systems, such as document management, customer service, logistics and e-mail.

The new user-centric approach starts with the entire data set and then excludes non-matching items from it. The user easily pares away the non-relevant data from the known collection via the new interface.

New search technology also provides "horizontal" information awareness. The new systems scale in three dimensions: the number and bandwidth of inputs; the total number of searchable information items; and the number of concurrent users and queries. The distributed yet integrated systems accommodate an unlimited number of structured or unstructured data inputs regardless of data format (see figure 1).

Just as user-centric search does not seek to recreate existing storage systems, it leverages third-party applications for locating the most relevant data in an organization.

A relevance algorithm provides a sorting order—from most relevant to least relevant—for a set of indexed content. The new approach does not provide any relevance algorithms out of the box. Instead, user-centric interfaces are relevance engine/algorithm agnostic, integrating multiple third-party and end-user defined relevance functions on-the-fly.

New search technologies also provide security. All large organizations require security features on their data storage and retrieval systems. Government agencies are, of course, mandated by law to use data systems with robust permissions management. The new approach to information retrieval will not require additional design, configuration or administration of a new security infrastructure. Rather, they will leverage the existing permission infrastructure already in place, including Active Directory permissions that are centrally maintained and familiar to system administrators. After the Search

In many cases, an information system's search functionality simply discovers if the requested information exists at all. Rarely can users act upon retrieved items immediately, without a "two-handed" cut-and-paste operation, or even worse, having to jot down the ID number and switch screens to retrieve the item.

In the new world of information retrieval, users will save time and work smarter by going to a single search interface to both initiate the search and complete the task.

Another unique, powerful feature of user-centric information retrieval technology is the possibility of virtual classification and the capability of easily sharing virtual classification between workstations and users.

A user can refine a query and then store that query, rather than the query results. The stored query then acts as a virtual, dynamic grouping of information: virtual, because the system does not tag the actual items in order to group them, and dynamic, because as additional information is entered into the system in real-time or data items age and are removed from consideration, the result set continuously updates itself.

Stored queries are saved as files that can be shared easily with other users via the network and through e-mail. Users share information by sharing how to find it, rather than by trying to move the information itself—collaboration by information referencing, rather than by information sharing.

Benefits for Government Organizations

What do all of these user-focused features add up to for government organizations, their IT managers, and their employees?

1. A new user-centric approach in KM will support faster response times, especially in times of crisis. Finding, making sense of, and acting on information will accelerate greatly across the organization. 2. Business processes no longer require being at the office. New platforms support access from anywhere, via secure and reliable mobile, Web and client interfaces centered around the employee.

3. Business requirements now drive technology solutions, rather than the other way around. User-focused KM gets business done faster.

4. User-centric solutions provide state-of-the-art, extensible and flexible information infrastructure. They can accommodate legacy, existing and future data sources and can adapt to a variety of sources without requiring costly data transformations.

5. Search-based interfaces allow for management and search of the user's most immediate work (files on local PC and in e-mails).

6. User-centric technologies mean less work for the IT group in managing an ever-increasing array of silos. 7. User-centric approaches to KM will be less expensive in every aspect, because they require less effort to manage, train and deploy. Focusing on end-user features with search-based interfaces can offer significant benefits to government agencies, their employees and their IT managers.

The X1™ Solution

X1's KM Model is not based on an ever-expanding taxonomy of existing data, but rather the intelligent indexing and searching of a company's data, no matter where it is stored and no matter whether it's organized or not.

The X1 KM Model recognizes the central truth that the traditional KM solution does not: people work in different ways, with various organization skills. They generally seek the shortest way between point A and point B irrespective of the "approved" path, and all knowledge workers have too much information and not enough time to find it and make sense out of it. X1's next-generation customizable search interface:

• Returns entire data set of matching items (not a page at a time);

• Allows random access to any element in the results data set (top, bottom, anywhere in-between, not just a list of 20 at a time);

• Provides high-fidelity previews of any data item without opening the underlying native application (not a 30-word extract, and you don't have to download the item and load the native application to simply view the item); and

• Allows real-time modification (add, delete, change, etc.) of multiple search parameters with instantaneous recalculation of results set.

X1 frees the IT department from expensive KM applications. Because X1's KM Model eliminates the need for locating, organizing and aggregating data across the enterprise, much of an organization's IT department resources are freed from the management tasks normally required of KM systems.

---

Dejan Nenov is a veteran entrepreneur and technologist who has held management positions with leading Silicon Valley and international technology firms. From 2002 to 2004, he was VP wireless messaging and VP engineering at Intellisync (Pumatech). As a founder and CEO of Jolly Object Software LLC, he funded and launched EZtearsheets, a newspaper and ad agency ad management system, and worked with Norwegian directory integration startup Metamerge, acquired by IBM/Tivoli in 2001. As a CTO at Instill Corporation, Nenov developed patent-pending technologies in the area of trading partner integration, data analysis and data warehousing, which today handle a unified system of over 2,000 physical locations and analyze over $10B in foodservice transactions.

X1 Government Edition has been accepted for use and deployed at secure installations in both military and agency accounts. For more information, please contact Gary Ward at 978-741-8094, or go to X1