In association with

Check out the Program 

Smart and Smarter: Using Knowledge Maps, User’s Context, and Structured Content to Make Chatbots Really Helpful

Alex Masycheff | Room 2, 2. day, Workshop

To provide accurate and relevant answers to users, most chatbots require explicit definition of question-answer pairs. While this approach may work for simple scenarios, for real-life use cases (e.g., customer support, marketing, sales, etc.), it poses several challenges: In the long run, when the amount of content is large, content is frequently updated, and contains multiple content variations (e.g., by product model or release), developing and maintaining such chatbots is expensive, timeconsuming, and error-prone.

One the challenges of the information age is that we don't know what we don't know. The user isn't necessarily aware of all available opportunities. If the user doesn't know that an opportunity exists, the user won't ask a question, and thus, won't get an answer, and will probably miss information that might be useful and relevant.

At this workshop, we will discuss how a combined power of ontologies, structured content, and user's context can help us build a chatbot that will be both scalable and smart.

You'll learn:

  • What makes most of chatbots non-scalable and expensive to maintain
  • Why a chatbot requires an ontology to become smart
  • How a chatbot can navigate the user through the knowledge by applying the user's context to the ontology
  • How an ontology can be used to infer semantic relationships that were not  explicitly defined, and how it can help the chatbot provide the user with useful and relevant information
  • Where ontologies are coming from and how to build them
  • What tools are available for building and managing ontologies