Abstract
Conversational interface (UI) for artificial intelligence (AI) assistants typically uses text or speech and, as such, presents some cognitive load. The UI can be unintuitive for complex queries, leading to frustratingly long to-and-fro conversations. For sophisticated tasks, the UI can lack expressivity, forcing users to use long text prompts over multiple conversational turns. This disclosure describes techniques that enhance and simplify interactions with AI assistants using automatically generated graphical elements that enable a user to manipulate and weigh particular parameters in their input queries. Two modes of user input are defined - text mode and graphical mode - with easy mode switching. Even as a user enters part of a complex query, the mode appropriate to the context (given the thus-far entered query) is automatically surfaced. The user can modify the query via automatically surfaced graphical UI elements to converge faster to a target outcome. Repetitive, back-and-forth conversation with multiple follow-on questions presently needed to converge to a satisfactory outcome is reduced or eliminated.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Govindaraju, Raviteja and Sedouram, Ramprasad, "Automatic GUI generation Based on Layered Queries for Interaction with AI", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/8857