Overview of generative AI
Last updated: 24 September 2024
Generative AI refers to a type of artificial intelligence that can create new content (text but also images, music, or other) based on patterns it has learned from existing data.
Unlike traditional AI, which focuses on analysing data and making predictions, generative AI is capable of producing original outputs.
In the context of our platform, our generative AI models are trained on vast amounts of written material to understand language structure, grammar, and context. As a result, our AI can generate human-like text in response to questions from your users.
Generative AI makes it easier and faster to produce high-quality chatbots, helping our customers (and their users) save time and effort.
While generative AI is powerful, it is important to remind users to review the generated text to ensure accuracy. AI is prone to hallucination.
Use cases
Our customers leverage our generative AI functionality for a variety of use cases such as:
Automatically generate responses to customer queries.
Automatically generate responses to employee queries based on the company’s handbook.
Explain complex financial context to users as they go through an application (e.g. loan).
And more.
Key functionalities
Our generative AI functionality relies on four key functionalities.
Large language models. Our models are trained to understand language and context, allowing them to respond in a way that feels natural and relevant.
Text generation. Our generative AI can create sentences and paragraphs based on user input, prompt customisation, and ground data.
Ground data. Our generative AI can use specific, customer-uploaded documents to respond to user questions. For example, a customer may upload an HR handbook to a chatbot’s knowledge base, forcing it to generate responses to employee queries using data it finds in the handbook.
Prompt. Our customers guide their generative AI’s output by providing a specific prompt or set of instructions. For example, a customer may enter the following prompt: “You are an HR chatbot. You answer using formal language”.