Extraordinary Journey or Navigation Nightmare?
Who doesn’t like the idea of a road trip? The anticipation of new experiences and attractions is always exciting. But if you miss a turn, take the wrong exit or receive bad directions along the way, you could find yourself completely lost. Ultimately, the experience will be frustrating and fall short of your expectations.
This is an apt analogy for examining critical issues in knowledge management (KM). When a user sets out on a journey to find information, KM should act as a sophisticated navigation system that quickly and enjoyably guides them to their destination. It should also help the user find relevant points of interest along the way, and provide a detour when they encounter a dead end.
"Will you guide me to Springfield?"
A little clarification goes a long way. If you were asked this question as the designated "navigator" on a road trip, you wouldn’t hand the driver a world atlas or a stack of travel guides for every Springfield in existence. But in a way, this is how many conventional knowledge management solutions operate. They return excessive links and documents without truly understanding what the user needs.
To be effective, the KM system must first understand that "Guide me to Springfield" means something different than "Find me a guide in Springfield" or "Find me a Springfield guide." It must also understand that the intent behind "Take me to Springfield" and "Drive me to Springfield" is the same. Then, because there are many "Springfields," the system must clarify which "Springfield" the user wishes to visit (in this case, Springfield, IL). This enables it to zero in on a specific map.
Challenges in these areas stem from the fact that most conventional KM solutions are based on keyword search and pattern-matching technologies that return results based on matches among words and phrases. They do not understand the meanings of words, syntax or the intent behind expressions. Moreover, they cannot help the user clarify their specific needs.
As a result, two users could receive surprisingly different answers to the same basic question. Conversely, two dissimilar questions containing identical keywords could receive the same response. This challenge is a constant because these systems cannot learn to recognize and accommodate the myriad ways users ask questions. Consequently, they deliver a high rate of "false positives" and inaccurate answers.
KM solutions utilizing natural language dialog-based technology overcome these challenges. This type of solution is engineered to understand the meaning of words (including synonyms, pronouns, sentence fragments and slang), and the intent behind questions and expressions, to deliver the most accurate and consistent responses. It can engage the user in a "human-like" conversation, asking questions to clarify needs.
Avoiding Delays
Building context expedites the process. As the journey continues, the driver asks you "What exit should I take?" You shouldn’t need to ask where they’re going to provide the right answer—you already know the destination is Springfield, IL. Subsequent conversations should build on that knowledge.
Keyword search or pattern-matching-based KM solutions do not typically "learn" during an interactive session. Each subsequent inquiry is treated as a new search.
A natural language dialog-based KM solution can learn during a "conversation" and contextualize responses around user-specific information retrieved from other systems. It delivers more precise and relevant responses. In fact, it can deliver an answer as a direct statement, present a specific link or document, and/or point to a specific paragraph within a document. When the answer is unclear, it can provide links to related topics, ask the user to rephrase the question, or escalate the interaction to a live agent along with the interaction transcript.
Dealing with Roadblocks and Dead Ends
It is impossible to predict every scenario. Companies receive tens of thousands of questions every month. They can’t be prepared for every one. To ensure the success of your KM solution, you must be able to identify and overcome the roadblocks and dead ends: gaps in the knowledgebase. With many conventional KM solutions, identifying and filling all of the gaps is impossible. They can generate reports indicating the number of unanswered questions, but they cannot identify or categorize them in a meaningful way.
A natural language dialog-based KM solution can analyze all users’ questions and categorize them under meaningful topics. This capability enables organizations to quickly identify and address inaccuracies and knowledge gaps, while helping them to accurately assess the system’s overall performance. Additionally, it provides organizations valuable insight into areas of opportunity.
Introducing New Sources of Information
Change is certain. Just as the printed map to Springfield could be out-of-date, knowledgebase information becomes obsolete. Unfortunately, too many organizations rely extensively on IT—or a third party—to update their knowledgebase when the information should be input directly from expert source (i.e. product managers, marketers, developers). This process is inefficient and expensive.
Choosing a KM solution that enables subject-matter experts to easily author knowledgebase content eliminates the strain to on IT and mitigates risks. When this KM system acts as the centralized knowledgebase for all channels, the result is more consistent, accurate and timely information and better service across all channels. Web users, contact center agents, employees and partners can immediately access up-to-the-minute information—getting to the heart of KM best practices.
Astute Solutions’ RealDialog™ is a KM and Web self-service solution that ensures consistent, accurate and timely responses across all channels. Through advanced linguistic processing and an understanding of natural language, RealDialog grasps the intent of user expressions and engages them in dialogs to deliver exceptionally precise responses. Additionally, it captures direct and unfiltered insight into user needs and concerns. Please visit
www.astutesolutions.com/kmbestpractices.