Getting Started

Introduction

With Pulse, you can quickly launch a smart virtual colleague in the form of an AI chatbot. This guide takes you through the process of building your Pulse chatbot, from start to activation, and beyond activation, to further optimize the chatbot.

Preparation

Before diving into building, it's wise to determine the role your chatbot will play within your organization and the exact scope of your chatbot. Think of the scope as a compass that guides the route. It indicates what you want the chatbot to answer and which processes and tasks you want the chatbot to automate. Ask yourself questions like 'What questions should my chatbot be able to answer?' and 'Where should my chatbot not delve too deeply?' By thinking about this in advance, you can focus on implementing valuable knowledge during the building process of your chatbot, resulting in more efficient development and an improved experience for the users of the chatbot.

The technology

Pulse utilizes OpenAI's language models: GPT 3.5 and GPT 4o. Two key components of this AI are Natural Language Understanding (NLU) and Natural Language Generation (NLG). NLU helps the chatbot understand what people are saying by analyzing sentences and determining their meaning. NLG ensures that the chatbot can provide human-like responses by converting the gathered information into understandable language. Because a chatbot based on GPT-3/4 is an advanced model, each response will be slightly phrased differently. This is because the model has a lot of knowledge and can apply variation in how it formulates responses. This makes conversations with the chatbot more natural and enjoyable!