Providing variable responses in a virtual-assistant environment
Inventors
Brown, Fred A • Miller, Tanya M • Zartler, Mark
Assignees
Verint Americas Inc • Next IT Innovation Labs LLC
Publication Number
US-9836177-B2
Publication Date
2017-12-05
Expiration Date
2031-12-30
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
Virtual assistants intelligently emulate a representative of a service provider by providing variable responses to user queries received via the virtual assistants. These variable responses may take the context of a user's query into account both when identifying an intent of a user's query and when identifying an appropriate response to the user's query.
Core Innovation
The invention provides techniques for virtual assistants to intelligently emulate a representative of a service provider by providing variable responses to user queries received via the virtual assistants. These variable responses take into account the context of a user's query both when identifying an intent of the query and when selecting an appropriate response. The techniques parse a user's query to identify expressed concepts and the context associated with the query, then map these to one of multiple different intents. Subsequently, the intent is mapped to a response based at least in part on the context, and the response is provided to the user via the virtual assistant.
The problem being solved addresses the difficulty users face when seeking specific information on service providers' websites or client applications due to the overwhelming amount of content and navigation paths. Although users access information online to avoid cost and inconvenience associated with human representatives, the complexity of site content can frustrate users, causing them to abandon their queries or resort to contacting human support, eliminating the cost-saving advantages. Increasing the ability of virtual assistants to emulate human representatives and respond more intelligently to user queries remains a priority.
Claims Coverage
The patent includes two independent claims directed to computer-readable media, computing devices, and methods for providing variable responses via a virtual assistant.
Mapping user query to intent based on concepts and context
The invention captures a single query comprising a fixed word or phrase via a virtual assistant. It parses the query to identify one or more concepts expressed therein. It then identifies a context associated with the user or session (such as geolocation, time of day, device type, language type, sign-in status, or user status). Using the identified concepts and a first portion of the context, the query maps to one of multiple intents.
Determining response from intent and second portion of context including actions
After mapping the query to an intent, the system determines a response to the query from among a plurality of responses associated with that intent, based on a second, different portion of the context. The response specifies content for output via the virtual assistant and an action to perform on behalf of the user.
Switching communication channels to provide responses
The invention determines to switch from the first communication channel to a second communication channel to provide at least a portion of the response to the user. The second communication channel is selected from multiple options based on a communication channel previously used by the user to communicate with the virtual assistant.
Providing response content with various media and performing actions automatically
The response provided can include plain text, hyperlinks, audio output by the avatar representing the virtual assistant, and actions such as navigating to different pages, purchasing items, altering reservations, or initiating communications on behalf of the user. The virtual assistant is configured to emulate human-to-human interaction and may resemble a human when displayed.
The independent claims cover systems and methods that receive a single user query via a virtual assistant, identify concepts and context to map the query to intents, determine responses from intents and context, switch communication channels to output the response, and perform actions on behalf of the user. These features allow the virtual assistant to provide variable, context-aware, and action-enabled responses that emulate human representatives.
Stated Advantages
Enables virtual assistants to more closely emulate human-to-human interactions by taking into account the context of user queries both when identifying intent and when selecting responses.
Allows virtual assistants to provide variable responses tailored to the user's context, improving user experience and relevance of interactions.
Supports multiple modes of communication and dynamic switching between communication channels based on user behavior or schedule, enhancing flexibility and responsiveness.
Automatic performance of actions on behalf of the user streamlines user interactions and improves service efficiency.
Documented Applications
Used by service providers on websites or client applications to provide information to users without human representatives.
Applications include merchants, news organizations, insurance agencies, or any entity providing products, digital content, or services.
Virtual assistants display avatars resembling human representatives and provide interfaces (e.g., text boxes) for users to input queries.
Responses may include textual, audible content, hyperlinks, and automatic actions related to reservations, orders, or requests.
Suitable for use on various electronic devices, including mobile phones and desktop computers, over diverse communication networks.
Interested in licensing this patent?