How to identify the best chatbot solution for your support issues?

This is the second article in the series State of support Chatbot . If you haven’t read the first one, do take a minute to read it.

Customers contact support with a variety of questions and issues and it is overwhelming for a person or a team to browse through all of them to draw issue categories out of them. Thanks to modern CRM, support organizations now have the following tools to categorize issues on the fly.

1. Disposition or Closure codes – These are combo box or a picklist present in the support case to capture the product, feature, issue type, resolution status..etc.
2. Labels/Tags – These are keywords assigned by the system automatically based on the case details using NLP (Natural Language Processing).
3. Case subject and description
4. Knowledge Base article attached (or Linked) – Organizations practicing Knowledge-Centered Service (KCS®) have article linking or attach as a standard practice in the case flow. These articles have primary and meta fields which contain rich information about the issue, environment, cause, and resolution.

Dispositions often provide shallow or no insight for this purpose, as they are kept very broad. NLP based labels capture the keywords, with a powerful clustering tool, one could easily list trending/frequent issues. This is considering the labeling/tag system is good, if not this should be passed too. A sample analysis of the case subject will provide better insights as long as they are in customer words.This is the source I prefer and luckily I had access to a good of sample cases with issue description from the customer. Knowledgebase article usage data is an excellent data point too, however, it should be used only when the *Attach Quality or Link Accuracy is *greater than 90%.

It is critical to get the issue as explained by the customer if not, the complexity and context of the issue will be lost. Example: A customer could contact support stating “I am unable to load …. Website” . However, it is very common to find a case where it could be captured as “Customer required help clearing cache and cookies” with an article for How to clear cache and cookies? Attached.

Analyzing the case subject or article attached from the above example, one might categorize it as a simple issue, however from the context of the customer it is not simple. With that in mind choose the data source that can help you categorize the issues between the issue complexity levels below.

Depending on where most of your support issues fall in the complexity matrix, the type of solution, effort and ROI (Return of Investment) of your chatbot will change drastically. For example, if most of your support issues fall into Level 1, a simple chatbot solution that can leverage existing help content might bring in the best result.

At the end of this analysis, you should be able to state the percentage breakdown of support volume by issue complexity.

If you found this article helpful, share it. If you have questions or feedback, comment or direct message me.

Knowledge-Centered Service (KCS®) is a registered Trademark of the Consortium for Service Innovation™.