2 Building a conversational AI

 

This chapter covers

  • Building an FAQ conversational AI
  • Building a process-oriented conversational AI
  • Using generative AI inside of your conversational AI

In production, conversational AI can be quite complex, and throughout this book, we’ll cover many techniques that address the real-world problems you’ll face as you build and deploy your own solutions. In this chapter, we’ll build Cake Bot, a conversational AI solution with elements from several different kinds of conversational AIs. This will give us a solid foundation for understanding conversational AI structure.

We’ll follow a fictitious small American bakery from Ohio called Cake Shop. The company makes custom cakes and takes orders for delivery or pickup. They want to add a conversational AI solution to their website to help their customers. Since they have never built a bot before, they intend to start small but hope to quickly expand the scope and capability of their solution. They decide to start with an AI solution that answers their most frequently asked questions.

Many of the tasks in this chapter could be done with large language models. However, this bakery is cautious. They especially want to control the wording of responses given for several question types. Thus, their solution will blend traditional and generative techniques.

2.1 Building an FAQ bot

2.1.1 FAQ bot foundations

2.1.2 Static question and answering

2.1.3 Dynamic question and answering

2.2 Routing agents and process-oriented bots

2.2.1 Routing agents

2.2.2 Transitioning from a routing agent to a process-oriented bot

2.3 Responding to the user with generative AI

2.3.1 Integrating with an LLM

2.3.2 Routing requests to an LLM

Summary