Knowledgebases provide the tools to enable research – allowing you to find content or documents related to a given subject. The problem for the end user is that they are not out to conduct research– they have a specific issue that needs resolution. This simple fact renders the associated knowledgebase articles an archaic means of discovery that places the burden of effort on the user to identify an answer, rather than placing it on the software tool itself. Associated with this burden are four primary constraints that call the efficacy of knowledgebase solutions in to question:

1. Volume constraints  – the multitude of potentially relevant articles creates the proverbial needle in a haystack situation
2. Omissions – the necessary content does not exist or is not well represented
3. Assumed understanding of the domain – assumption that the user has the foundation necessary to identify what it is they are looking for using appropriate domain terminology
4. Measurable outcomes– just because a page is frequented often does not mean that it is successfully addressing any user issues

Virtual assistants resolve this on all levels. Volume constraints are eliminated because there is no hierarchy of navigation or competing documents that potentially muddy the waters – virtual assistants codify the one best answer or the most efficient path to resolution either in direct response or by pointing you to a very specific piece of content. While not entirely eliminated, omissions become much less of a concern through what we refer to as the ‘voice of the customer’ – the ability to identify very specifically what is important (topic and frequency) to end users and, based on business priority, ensure that all future occurrences are addressed. Inversely, having users searching and browsing content where you don’t establish any sense of what they were really after leaves you subject to guessing about what is important to them – and no amount of search tuning, keyword tagging or meta data can make up for the fact that your knowledgebase is crafted with fundamental lack of understanding of a user’s intent. This issue of findability is only compounded when you consider users who do not know exactly what they need or how to frame queries to identify the necessary content.

Let’s face it – knowledgebases are much more effective for internal employees – experts who have learned and understand the nuances of the application and the multitude of ways to form a query to improve their results. They also have the foundational understanding to know when they have identified the appropriate material, while less familiar users may walk away still in question. The result is that, according to the Customer Contact Council, 57% of inbound calls come from customers who were already on your website navigating content and searching through online tools such as a knowledgebase.

Hi, Sam, and welcome to chatbots.org!

You make some valid points in your post, and the article that you’ve linked to makes for an interesting read, even if the article is nearly a half decade old. For those of us who are working on developing a chatbot whose main role is to assist customers in various ways, you’ve provided some insights that should be carefully considered. Thanks for the information, and I apologize for the delay in responding.

Thanks Sam for the article and welcome!

I am in the process of developing a commercial virtual assistant and while evaluating the market for such a product, I had come to many of the same conclusions. First time users to websites are much like first time visitors to a large department store. Possibly they may have come to purchase a specific item or even just to browse. Being unfamiliar with the store’s layout and range of products, the customer would quickly become frustrated and possibly leave the store without making a purchase. This is why there are sales associates ready to answer questions, direct the shopper to the right department or suggest a product. This is true on websites as well. Sure there are search engines, knowledge bases and comprehensive nested menus. But still if the visitor does not find the information that they were seeking quickly they leave the site unsatisfied.

The role of a well designed virtual assist is to have the ability to evaluate the context of the visitors questions, reply with a helpful answer and or provide navigation to associated resources. Virtual assistants can play a valuable role in customer retention while being a cost effective solution over live agents. Sadly many VA’s that I have seen on websites have missed the mark, not by technology but in the knowledge base design itself.

You can have access to great tools but if not utilized in the right way they can make matters much worse.