Virtual personal assistants
TEXT_INPUT:Ask the user to input their name;;CHATGPT:Prompt ChatGPT to greet the user by their name;;HTTP_REQUEST:Fetch current weather information based on user location;;DISPLAY_OUTPUT:Display the current weather information to the user;;DISPLAY_OUTPUT:Display a closing message to the user
Generative AI models like Language model (LLMs) and image generators are valuable resources for enhancing Virtual Personal Assistants. LLMs such as ChatGPT by OpenAI, adept at understanding and generating human-like text, can be used to create more naturalistic interactions, improving the overall user experience. For instance, integrating ChatGPT into voice assistants can allow for more dynamic, tailored conversations.
On the other hand, image generators, equipped with the ability to create a diverse range of images from text descriptions, can provide visual assistance. For example, when asked about a certain animal, a VP assistant utilizing such AI could provide a descriptive image along with the information. Similarly, when scheduling appointments, relevant visuals could be created to add context. Thus, by integrating these technologies, Virtual Personal Assistants can be made more interactive and efficient.
How to build with Clevis
This description provides an understanding of an example application you can build using Clevis, a tool for building AI applications. The application in question is designed to offer virtual personal assistant services to users, operationally streamlined via a series of carefully ordered steps.
In the first operation (Text Input), users are asked to input their names. This input data is then promptly passed on to ChatGPT, an advanced language AI model developed by OpenAI, in the following step (ChatGPT). The application uses the given data to have ChatGPT effectively greet users by their names, providing a personalised interaction.
The application then performs an Http Request to fetch current weather data based on the users' respective locations. The next step (Display Output) ensures that this weather information is correctly displayed to the users, offering useful and timely information.
In the final step of the operation (another instance of Display Output), the application displays a closing message, effectively ending the current user interaction session.
It's important to note that applications with the same or similar functionality can be built using Clevis, enhancing its versatility and adaptability in various AI operations.