Part 1 of this blog series introduced the need for knowledge management (KM) software applications as part of a more comprehensive and strategic service management (SSM) suite. One such broad SSM suite has been offered by Servigistics.
Servigistics’ Service Knowledge Management (SKM) solution, the newest module within Servigistics SSM, is designed to meet the requirements of the technical service organizations that manage complex problem resolution. The crucial issue is that technicians, dealers, agents, partners, and customers need in-depth knowledge to solve complex problems.
Diagnosing issues in these complex environments (e.g., motor vehicles, aerospace & defense [A&D], medical equipment, appliances/white goods, high tech) requires interaction and a comprehensive understanding of the essential diagnostic variables. As a good example, medical doctors have been provided with a framework that allows them to be masters of the diagnostic method because of their years of diagnostic training. As a trained diagnostician, the doctor is able to capture the essential diagnostic information from the patient and match it against prior treatment experiences.
However, in the world of technical support services, a formal diagnostic framework does not usually exist, since technicians and support personnel do not commonly receive standardized training in how to diagnose a technical problem. The SKM solution’s role is thus to help in the field as newer technicians will be less experienced while the experienced ones are about to retire in droves. In addition, loyalty has been down lately so future technicians will likely not obtain that same experiential knowledge. Moreover, fewer people are going into field service these days.
So, What Can We Do?
One solution is to leverage (collect, retain, and reuse) documents, experience, and expertise within a knowledge repository (base), with powerful retrieval methods (search algorithms) to provide information and resolutions via streamlined access to this stored experience. To that end (which is much easier said than done), Servigistics’ approach enables a single repository to serve as a knowledge store and as a source for interactions through many channels (Web, call center agent, and wireless) and in multiple languages simultaneously.
The SKM’s Knowledge Advisor (described later) can be set up to provide a way for the knowledge base to be queried in many languages. By translating the words in the domain model, this would allow a user to type in his or her searches in one of the languages and he or she would be walked through the same guided-search process. Although not all of the solutions might be translated into the user’s language of choice, Servigistics’ analysis has shown that the most critical factor is that the users be able to search for the “right” answer in their native language.
The SKM solution combines interactive diagnostics with so-called Text2Data processing to gather information along with Knowledge Studio and Knowledge Advisor in a single application. The retrieval engine has been provided by Kaidara in an original equipment manufacturer (OEM) fashion.
Assembling and maintaining a KM solution relies upon the service organization’s ability to manage relevant content and tag the content in a way that the target user base can find it when needed. In this regard, Servigistics has broken deployment of the system into the following four distinct steps, with appropriate products to address these specific areas:
1. Gathering History — Most KM deployments fail due to lack of content and lack of a defined process to author and populate the knowledge base. SKM is able to analyze and capture accurate and succinct problem-resolution pairs from disparate historical files to pre-populate and jump-start a problem resolution project.
Volumes of product data and information are generally plentiful in both structured data (database rows, fields, tables, etc.) and unstructured content (forms, documents, pictures, etc). Nonetheless, this data and content is not useful for reuse until it is normalized, validated, and tagged for retrieval.
The aforementioned Text2Data capability provides an automated means to gather, analyze, and categorize text documents or database records and transform them into a usable form. Text2Data processes a set of input records that generate output that describes each of the records. The input records are generally a set of text documents that are relevant to the target knowledge domain or a database table containing rows of information, such as call records, service requests, trouble tickets, and service bulletins.
This means that materials in various formats (i.e., structured and unstructured) can be incorporated into the system, thus adding value to content and future searches. The results from Text2Data processing identify important concepts within the data and converge upon a normalized set of solutions.
This convergence of product data into solutions provides a solid “starter” set of content and may be run iteratively throughout the SKM lifecycle to provide additional content. In other words, the system has a “self-learning” capability (i.e., to create new resolution cases based on history of symptoms, questions, solutions, etc.).
2. Organizing Knowledge — The abovementioned Knowledge Studio module creates a structured representation of content and maps the content to a standard vocabulary, which is commonly referred to as a domain model. The domain model is an abstract image of the application area in which the system operates, and it provides the vocabulary by which objects, states, actions, and relations can be described. It offers a structured representation of the attributes, their types, and the values that can be used to represent information pertinent to the domain.
Knowledge Studio provides the basis for representing, accessing, and handling information, in order to drive the interactive diagnostics. Since all representation schemes are based on the domain model, consistency is always ensured.
The Studio enables administrators and authors to efficiently tag solutions and add new metadata to the model (via translation, synonyms, acronyms, attribute tagging, etc.). Moreover, the domain model approach eliminates the need for any static (hard-coded) decision-tree maintenance, so all of the interactive diagnostics are dynamically automated from the domain model itself.
3. Distributing Knowledge — The cornerstone of the SKM offering and the means to realize the full potential of interactive diagnostics is called Servigistics Knowledge Advisor. The solution functions much like emulating the most experienced technical support agent’s thought process in dealing with customer problems. The Advisor interacts with the user in a natural language and flexible manner, similar to that of skilled technicians.
The goal of Knowledge Advisor is to help the user find the “right solution” to his or her specific problem, from anywhere (via intranet or Internet) in the shortest possible time with the greatest possible accuracy. The user does not want to receive a “no solution found” notice, nor does he or she conversely want to be told that there are “150 possible solutions to your problem.”
Different Strokes for Different Folks
Within the flexibility theme, SKM offers multiple search interfaces dependent upon the users’ experience (novice vs. advanced) and the context of what problem they are looking to solve. Critical to the success or failure of any system is the “look and feel” of the user interface (UI).
The approach adopted by Servigistics is to provide a flexible UI design framework that can accommodate the specific requirements of any user level. The tools to modify the system’s interface are delivered with the system, including a comprehensive set of instructions. The modes of user interaction with the system are as follows:
- In a Guided Search mode, users can describe their problems in their own words and Knowledge Advisor dynamically generates the most relevant questions to quickly address the problem at hand. Through this process, the user is walked through a natural question and answer (Q&A) dialogue to results. Even if a user is not able to answer a question from the Advisor, he or she can skip it and the Advisor will simply narrow the results by asking a different question. This approach is fully automated and requires no pre-programming to execute;
- In a Text Search mode, regardless of how a problem or question is stated in Knowledge Advisor, the system reads every word input and has the intelligence to understand the context of the words (i.e., it can account for the use of synonyms, acronyms, abbreviations, misspellings and lexical patterns). For example, words that are searched can include the “exact” word matches or perhaps if a user enters a code number, Knowledge Advisor would be able to detect if the code entry matches, e.g., the pattern of a product model number or an error code pattern; and
- Finally, if an experienced Knowledge Advisor agent uses the system, he or she can bypass the Guided Search and drill down through the system to quickly find the “right” solution with Expert Search. This search method directly exposes the metadata of the domain model, allowing the user to query the specific metadata areas directly and even change the weight of the areas.
4. Maintaining Knowledge — While building the knowledge base is the starting block for the SKM system’s instance, maintaining the content relevancy and verifying data tagging are the keys to the long-term viability of the system. To that end, Servigistics includes within the SKM system a set of tools and features to ensure that the knowledge stored in the system is not only easily imported but is also accurate and relevant to solving intricate technical issues.
The Knowledge Advisor module provides a means for authors, knowledge administrators, and subject matter experts to add or edit existing problem resolutions in the knowledge base. The authoring tools are designed to safeguard entry of duplicate solutions, and may be set for access and/or editing per role.
Furthermore, through Knowledge Studio, the administrator may set up workflows for authors to use in Knowledge Advisor. For example, when an author enters a new solution, the system may require a review and authorization process by several “subject matter experts”, whose role is to ensure that the information entered into the knowledge base is accurate, complete, and up-to-date with the latest support practices being used.
In some cases the reviewers might include product managers, marketing managers, or regulatory and legal personnel. The workflow process ensures that each solution is reviewed automatically, approved in a timely fashion and each step of the process is time- and date-stamped with version control.
Last but not least, the structured approach to knowledge provides advanced reporting and analytic capabilities, since the Servigistics Knowledge Analytics module has tools that report on supply/demand for both metadata and content (that can be segmented on the data’s structure). These capabilities provide administrators with decision support for maintaining retrievable and relevant content.
Knowledge Analytics generates reports that reflect how people are using the system, what types of searches they are making, and captures feedback on the user’s impressions and rating their success using the system. For example, a “gap” analysis report would identify “holes” in the product coverage of possible solutions. The Analytics module also provides administrators with reporting views to create custom reports on user, content, tagging, or session data.
Any Potential Benefits from These Whizbang Thingies?
As mentioned in Part 1, some Servigistics case studies have proven that SKM can reduce troubleshooting time by solving issues once and making the resolutions available across the service organization. It is sort of like a collective unconscious aligning everyone in the service operation.
In addition, the solution can decrease the number of service calls by empowering customers, technicians, partners, and dealers to more reliably and accurately find their solutions through self-service. Furthermore, it can improve product reliability by identifying and tracking quality issues (e.g., the number of uses per solution).
For example, a France-based home appliance manufacturer found that in less than six months, SKM improved first call resolution from less than 20 percent to over 50 percent. In addition, Level 1 agent training was reportedly reduced from eight weeks to four weeks, while operating costs were cut by five percent.
For global companies preparing for the loss of valuable workers due to retirement, it should be encouraging to know that their collective knowledge and experience won’t be lost. Instead, this know-how can be collected, automated and accessed by everyone in the service operation.
In other words, in operations that face the potential loss of intellectual capital through retirement, the SKM can reduce service employee ramp-up time by providing a diagnosis framework. So, instead of viewing this upcoming workforce transition as a problem, best-in-class companies can use it as an opportunity to manage service as a profit center rather than a cost of doing business.
The third and final part of this blog series will conclude with the potential benefits of the SKM solution and its competitive landscape. In the meantime, what are your views, comments, opinions, etc. about KM deployments in general, and about Servigistcs’ SKM per se?
knowledge Management is essential for an employee