• whitepaper

    Automating Technical Support Knowledge



    1.0 The Challenge of Delivering Expert Technical Support

    The increasing complexity and variety of IT, Internet, Telecom and technology products and services being offered to customers and the rapid pace of change of these products and services is increasing the demand for expert customer service. This involves the timely and effective diagnosis and resolution of customer queries / problems that is vital to customer satisfaction and retention. Delivering such support requires an increasing number of highly trained customer service agents. In a commercial environment of increased competition and reduced profit margins, the challenge is how to reconcile the seemingly conflicting objectives of reducing customer support costs while maintaining / improving support quality. The challenges can be summarised as follows:

    • Products and services are getting more complex, have more variants and change very rapidly over time.

    • Customers are demanding high quality service over multiple contact channels (telephone, email, chat and web self-service).

    • Customer Service agents need extensive and regular training and are difficult to retain.

    • Pressure to improve customer service and at the same time to reduce its cost.

    1.1 CRM / Call Tracking Systems Do Not Improve Customer Service

    Investment in CRM (customer relationship management) software has done little to improve customer loyalty according to the latest research from IT services company Accenture. A survey of 2,000 customers in both the UK and US found that 61 percent identified poor service or product quality as the main reason for moving between companies. Almost one-fifth of respondents cited technology problems as a reason for their dissatisfaction.

    Although CRM / call tracking systems are intended to improve customer relations, over one-third of customers complained of being forwarded through multiple company representatives before problems were adequately resolved. “Customers expect the first customer service representative they talk with to have the knowledge, tools and capabilities required to address their needs,” according to Robert Wollan, managing partner of Accenture’s customer contact unit. Customers spend an average of six minutes on hold while waiting for help on the telephone and speak to an average of 2.6 service representatives before their query is resolved.

    1.2 Knowledge Management Delivers Limited Benefits to Customer Service

    CRM systems are being supplemented by Knowledge Management Systems in an effort to improve customer service. The majority of the current generation of Knowledge Management solutions for customer service involve the managing of unstructured knowledge. By managing unstructured knowledge we mean managing content such as documents of various formats, web pages, and similar (emails, textual description of solutions to problems) using search engine like technology. The knowledge management technology normally involves the indexing of these documents (content) and allowing the subsequent search for the relevant documents through keywords search, natural language query or browsing simple hierarchical categories. Other extensions to managing unstructured knowledge include:

    • Showing a list of the current most searched for documents.

    • Allowing the users of the search system to give their feedback about the relevance of the found documents to their query. This feedback is then used to improve future search results.

    The limitations of using this type of technology are firstly that the solutions found depend very much on the way the query is phrased and secondly that ultimately the customer service agent is interpreting the recommendation, advice and problem resolution knowledge in the found documents. For this reason, if the knowledge becomes moderately complex, the quality of customer support will start to depend critically on the interrogation and interpretative skills and experience of the agent thereby eroding most of the benefits expected from the introduction of knowledge management. Similarly, this approach to knowledge management can be of little benefit in offering effective technical support over email, chat or web self-help. Therefore knowledge management can, in effect, increase the skill requirements (and therefore cost) of an agent because the job becomes akin to a technical help desk.

    1.3 Automation of Support Knowledge Improves Customer Service

    Support Knowledge Automation (also called Service Resolution Management) captures structured support knowledge such as diagnostic trees, trouble-shoot flows and related service resolution procedures. Once captured, the deployment of such automated knowledge almost eliminates the reliance on the skills of agents in interpreting the knowledge. The agent is guided by the system to ask the right sequence of questions in order to achieve problem resolution in the most efficient and direct way. Support Knowledge Automation can therefore enable agents with minimal training to answer complex queries and problems. The resulting benefits are:

    • Increase first call resolution and reduce call duration. Agents assisted by Automated Support Knowledge can resolve queries during the first call and can do so in the most direct and efficient manner.

    • Deliver consistent and high quality customer support through every agent. Every agent when supported by Automated Support Knowledge becomes more effective and knowledgeable.

    • Empowering agents with knowledge reduces the cost of training agents and mitigates agent turnover.

    • Overcome the issues relating to outsourcing support to places like India. Support knowledge is deployed consistently regardless of agent or location and changes in product knowledge can be propagated immediately.

    • Increased customer satisfaction resulting from receiving fast, efficient and consistent support.

    The captured customer service knowledge can also be deployed to empower customer service agents to answer email and chat queries. Finally the same customer service knowledge can be delivered as a web self-help solution allowing customers to get the same quality of service while reducing the load on contact centres. Such technology is no longer “nice to have” but is a critical enabling technology for customer service.

    2.0 XpertRule eService – Integrating Structured and Unstructured Knowledge

    In practice, customer service knowledge is a hybrid combination of structured, unstructured knowledge and core customer service interaction skills. The solution to a customer problem/query can be found in a document or in a diagnostic / trouble-shoot tree or procedure. A Customer Service knowledge base consists of tens of thousands of such solutions. The technologies required for a hybrid (structured and unstructured) knowledge management system are illustrated below:




    Note that unstructured search results in finding a document or URL while structured search using diagnostic / classification trees can lead to a document or other structured knowledge solutions such as trouble-shoot flows. In other words structured customer support knowledge can be split into two stages; the first step is to classify / diagnose the nature of the query and the second step is the resolution of the query.

    XpertRule eService is a software environment for the capture and maintenance of structured knowledge and for the seamless integration of this knowledge with unstructured knowledge (content). XpertRule eService also delivers the most scalable deployment of structured knowledge on the market today. These breakthrough capabilities are achieved as described in the sections below.

    2.1 Knowledge Authoring Environment

    XpertRule allows the capture of customer service / support knowledge using graphical process flows (trees) representing recommendation, advisory, diagnostic, trouble-shooting and problem resolution knowledge. The Enterprise strength knowledge-authoring environment allows authors to collaborate on capturing and maintaining the customer service knowledge working over a network.

    The development environment maintains thousands of possible questions and flows (trees). Each knowledge author can work on his/her area of expertise whilst sharing questions and trees with other knowledge authors. The graphical and structured knowledge representation speeds the process of capturing the knowledge and subsequently maintaining the knowledge as the products and services evolve. Extensive technical support knowledge can typically be captured in a few weeks.

    Integration between XpertRule and a Content Management System (CMS) is achieved by means of an intermediary API to allow seamless integration between XpertRule resources and resources managed by the CMS.

    Customer Service Knowledge centres around the solutions base. Each solution can have an optional set of attributes. A solution points to a document, or to a knowledge flow / tree. The knowledge-scripting environment allows the maintenance of a large number of questions, advisory, diagnostic and classification trees which are used to narrow down the list of required solutions.

    The trouble-shooting knowledge consists of flows and trouble-shoot steps. Flows can also contain sub-flows representing common trouble-shoot sequences. The knowledge-authoring environment allows the maintenance of thousands of questions, steps, sub-flows and flows. Each knowledge author can work on his/her area of expertise whilst sharing questions, steps and sub-flows with other knowledge authors. Thegraphical and structured knowledge representation speeds up the process of capturing the knowledge and subsequently maintaining the knowledge as the products and services evolve.

    The task of finding the relevant solution(s) that can address the customer question or query is normally called a classification or diagnostic task. By asking a sequence of focussed questions, the knowledge flow narrows down the list of suitable solutions. Below is an example of such knowledge which narrows downs the solution to one of five possible trouble-shoot sequences.




    The knowledge for one of the five trouble-shoot solutions is shown below:




    Solutions can also be assigned attributes (e.g. Windows Operating System and Internet connection method) that can act as selection filters that further narrow down the relevant solutions to a query.

    The deployment engine logs all the data from agent sessions to a database. The data captured includes answers to questions, the time taken to capture each answer, trouble-shoot fixes tried and whether problem was resolved. All this logged data is processed by the authoring environment to reveal the statistics for each diagnostic and trouble-shoot flow. This is an important catalyst for improving the knowledge by highlighting “problems” with poor record of resolution rates and identifying which fixes are more effective than others at resolving problems.

    2.2 Multi-Channel, Integrated and Scalable Knowledge Deployment

    XpertRule offers a unique technology for deploying customer service knowledge across all customer contact channels; telephone, web self-help, email and chat. This is achieved through XpertRule’s pure HTML deployment that does not require a rules engine on the server or the client. The classification, diagnostic, advisory and trouble-shoot knowledge scripted in XpertRule is transformed into a set of intelligently linked HTML pages which are stored on a web HTTP server. This unique approach has very many advantages:

    1. It is extremely scalable whereby hundreds of thousands of agents / customers can access the knowledge base at the same time. This is particularly important when supporting large contact centres and web self-service.

    2. The Knowledge server requires very low maintenance since there is no knowledge engine to maintain. It is as easy as maintaining a web site.

    3. Because the same knowledge base is used for all contact channels, it becomes possible to switch customers from telephone, email or chat channels to the web self-help whereby the customers are allowed to continue the diagnostic/ trouble-shoot session at their own pace without the costly and lengthy interactions with the customer service agents. The current state of a session is stored as a “bookmark” to a relevant HTML page within the knowledge base and can therefore be easily restarted.

    4. All knowledge flows (e.g. trouble-shoot flows) are generated as HTML pages which can be treated as and integrated seamlessly with other unstructured content. This means that knowledge flows can be searched for using the usual keyword, natural language and FAQ techniques of unstructured knowledge management.

    The XpertRule HTML Solution can be easily integrated within any existing CRM system. For telephone agents and for web self-help, the XpertRule Service Resolution Knowledge base can be invoked from within the CRM system. XpertRule will commence the Q&A for problem classification/ diagnosis followed by the trouble-shooting steps until resolution is achieved. XpertRule then returns a string to the main CRM system containing a list of all questions answered during the session and trouble-shoot steps invoked. An XpertRule session can be interrupted before the end and restarted by the CRM system at a later date at the same point where it was suspended.

    For email service agents, XpertRule can provide an AutoSuggest feature that prompts the agents for answers to questions that may be contained in the query email. XpertRule then suggests an email reply which can either be a request for more information or a required trouble-shoot flow which XpertRule can attach as either a graphical tree or a textual set of trouble-shoot steps. Email replies can also include a link to web self-help to continue trouble-shooting.

    For chat service agents, again XpertRule can provide an AutoSuggest capability for the agent to respond to the customer or alternatively the agent can divert the customer to web self-help.




    2.3 XpertRule eService – Technology Positioning

    The diagram below illustrates the positioning of XpertRule eService in relation to Content Management (CMS), Knowledge Management (KM), Service Resolution Management (SRM) and Customer Interactions Management (CIM) Systems. This can be summarised as follows:

    • XpertRule captures the scripted knowledge (problem resolution) flows.

    • XpertRule can deliver knowledge flows as HTML only solutions to SRM / KM systems thereby adding seamless scripted knowledge capability.

    • XpertRule can deploy the captured scripted knowledge flows directly to CIM or CRM systems as high performance and easy to integrate HTML knowledge.

    • All knowledge objects in XpertRule such as trees, reports and trouble-shoot flows can reference contents from the CMS.




    2.4 Summary of the Unique Features of XpertRule eService

    • A central Knowledge Object Dictionary for questions, trees, flows, sub-flows, trouble-shoot steps. This allows efficient multi-user knowledge scripting and maintenance whereby all objects are reusable across the knowledge base. For example the Question “Windows Version” with answers “2000, XP Home, XP Professional” can be defined once and used in many diagnostic, trouble-shoot and recommendation flows. It also allows different knowledge authors to maintain different scripts.

    • A highly graphical and intuitive environment for building a large hierarchy of trees flows and sub-flows. Flow between trees is very visible, as is the hierarchies of flows and sub-flows.

    • An advanced knowledge representation for trouble-shooting knowledge. The basic units are questions and trouble-shoot steps. Each trouble-shoot step can have an “include rule” which allows the deployed knowledge to include/exclude the step based on other captured data. A common sequence of steps can be defined as a reusable sub-flow that can be referenced by other flows.

    • A unique method of deploying scripted knowledge consisting of a set of linked HTML pages. XpertRule offers two deployment modes. The first mode is a Q&A mode whereby the user is prompted a question at a time and the flow of questioning flows in a way that follows the scripted trees. This applies to the flow in diagnostic trees that determine the nature of the problem and to the trouble-shooting trees that are designed to find a resolution. This is an ideal mode for novice customer service agents and for web self-help. A second mode displays the diagnostic and/or trouble-shoot flows to the user to allow the user the freedom and speed of knowledge navigation. This mode is ideal for the experienced customer service agents.

    • The HTML deployment strategy allows Scripted Knowledge to be integrated easily into any CRM (Customer Relationship Management) or CIM (Customer Interaction Management) system. The HTML deployment strategy delivers high performance across all customer contact channels (telephone, email, chat and web self-help) and allows for the easy switching of a support call between channels (e.g.from telephone to email).

    • A unique method of monitoring and reporting the effectiveness of the deployed support knowledge. By superimposing the logged agents sessions on the diagnostic and trouble-shoot flows, it is possible to identify graphically where the gaps in the support knowledge are and which trouble-shoot fixes are more effective than others. This allows for an effective knowledge improvement strategy.