Conversational agent response determined using a sentiment
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for handing off a user conversation between computer-implemented agents. One of the methods includes receiving, by a computer-implemented agent specific to a user device, a digital representation of speech encoding an utterance, determining, by the computer-implemented agent, that the utterance specifies a requirement to establish a communication with another computer-implemented agent, and establishing, by the computer-implemented agent, a communication between the other computer-implemented agent and the user device.
1. A computer-implemented method comprising:
receiving, by a first computer-implemented agent for a user device, a digital representation of speech encoding an utterance that includes a command; and
in response to processing the digital representation of the speech to determine words included in the utterance:
determining, using the words of the utterance, to request input identifying a sentiment of a user who spoke the utterance;
in response to determining to request input identifying a sentiment of the user who spoke the utterance, providing a request for user input identifying a sentiment;
as a result of providing the request for user input identifying a sentiment, receiving user input identifying a sentiment of a speaker of the utterance;
identifying, by the first computer-implemented agent using the words included in the utterance, a second computer-implemented agent that is most likely able to perform the command, the second computer-implemented agent being a different computer-implemented agent from the first computer-implemented agent; and
in response to identifying the second computer-implemented agent that is most likely able to perform the command, providing, by the first computer-implemented agent to the second computer-implemented agent, data for the utterance and the sentiment of the speaker of the utterance to cause the second computer-implemented agent to provide a response to the command based on the sentiment of the speaker.
2. The computer-implemented method of claim 1 , wherein:
the first computer-implemented agent executes on a first system;
the second computer-implemented agent executes on a second system that is a different system from the first system; and
providing, by the first computer-implemented agent to the second computer-implemented agent, data for the utterance and the sentiment of the speaker of the utterance to cause the second computer-implemented agent to provide a response to the command based on the sentiment of the speaker comprises providing, by the first computer-implemented agent to the second computer-implemented agent using a network, the data for the utterance and the sentiment of the speaker.
3. The computer-implemented method of claim 1 , wherein identifying the second computer-implemented agent that is most likely able to perform the command comprises identifying the second computer-implemented agent that is most likely able to perform the command using the words of the utterance and a purpose for the second computer-implemented agent.
4. The computer-implemented method of claim 3 , wherein identifying the second computer-implemented agent that is most likely able to perform the command comprises:
determining, using the words of the utterance, a type of the command;
in response to determining the type of the command, determining, using the type of the command, that two or more other computer-implemented agents are able to perform the type of the command, the two or more other computer-implemented agents including the second computer-implemented agent;
in response to determining that two or more agents are able to perform the type of the command, determining, for each of the two or more other computer-implemented agents, a likelihood that the respective computer-implemented agent is able to perform the command using a purpose of the respective computer-implemented agent; and
in response to determining the likelihoods for each of the two or more other computer-implemented agents, selecting, as the second computer-implemented agent, the computer-implemented agent from the two or more other computer-implemented agents that has the highest respective likelihood that the computer-implemented agent is able to perform the command.
5. The computer-implemented method of claim 1 , wherein the command comprises a request for results responsive to the utterance.
6. The computer-implemented method of claim 1 , wherein providing the request for user input identifying a sentiment comprises requesting user input identifying a yes or a no response to a question.
7. The computer-implemented method of claim 1 , wherein providing the request for user input identifying a sentiment comprises requesting user input identifying an option from a set of options.
8. The computer-implemented method of claim 1 , wherein identifying the second computer-implemented agent that is most likely able to perform the command comprises identifying the second computer-implemented agent that is most likely able to perform the command using the words of the utterance and profile data for the user device.
9. The computer-implemented method of claim 8 , wherein the profile data includes demographic information for the user who spoke the utterance.
10. A system comprising one or more computers and one or more storage devices on which are stored instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
receiving, by a first computer-implemented agent for a user device, a digital representation of speech encoding an utterance that includes a command; and
in response to processing the digital representation of the speech to determine words included in the utterance:
determining, using the words of the utterance, to request input identifying a sentiment of a user who spoke the utterance;
in response to determining to request input identifying a sentiment of the user who spoke the utterance, providing a request for user input identifying a sentiment;
as a result of providing the request for user input identifying a sentiment, receiving user input identifying a sentiment of a speaker of the utterance;
identifying, by the first computer-implemented agent using the words included in the utterance, a second computer-implemented agent that is most likely able to perform the command, the second computer-implemented agent being a different computer-implemented agent from the first computer-implemented agent; and
in response to identifying the second computer-implemented agent that is most likely able to perform the command, providing, by the first computer-implemented agent to the second computer-implemented agent, data for the utterance and the sentiment of the speaker of the utterance to cause the second computer-implemented agent to provide a response to the command based on the sentiment of the speaker.
11. The system of claim 10 , wherein:
the first computer-implemented agent executes on a first system;
the second computer-implemented agent executes on a second system that is a different system from the first system; and
providing, by the first computer-implemented agent to the second computer-implemented agent, data for the utterance and the sentiment of the speaker of the utterance to cause the second computer-implemented agent to provide a response to the command based on the sentiment of the speaker comprises providing, by the first computer-implemented agent to the second computer-implemented agent using a network, the data for the utterance and the sentiment of the speaker.
12. The system of claim 10 , wherein identifying the second computer-implemented agent that is most likely able to perform the command comprises identifying the second computer-implemented agent that is most likely able to perform the command using the words of the utterance and a purpose for the second computer-implemented agent.
13. The system of claim 12 , wherein identifying the second computer-implemented agent that is most likely able to perform the command comprises:
determining, using the words of the utterance, a type of the command;
in response to determining the type of the command, determining, using the type of the command, that two or more other computer-implemented agents are able to perform the type of the command, the two or more other computer-implemented agents including the second computer-implemented agent;
in response to determining that two or more agents are able to perform the type of the command, determining, for each of the two or more other computer-implemented agents, a likelihood that the respective computer-implemented agent is able to perform the command using a purpose of the respective computer-implemented agent; and
in response to determining the likelihoods for each of the two or more other computer-implemented agents, selecting, as the second computer-implemented agent, the computer-implemented agent from the two or more other computer-implemented agents that has the highest respective likelihood that the computer-implemented agent is able to perform the command.
14. The system of claim 10 , wherein the command comprises a request for results responsive to the utterance.
15. The system of claim 10 , wherein providing the request for user input identifying a sentiment comprises requesting user input identifying a yes or a no response to a question.
16. The system of claim 10 , wherein providing the request for user input identifying a sentiment comprises requesting user input identifying an option from a set of options.
17. The system of claim 10 , wherein identifying the second computer-implemented agent that is most likely able to perform the command comprises identifying the second computer-implemented agent that is most likely able to perform the command using the words of the utterance and profile data for the user device.
18. The system of claim 17 , wherein the profile data includes demographic information for the user who spoke the utterance.
19. A non-transitory computer readable storage medium storing instructions executable by a data processing apparatus and upon such execution cause the data processing apparatus to perform operations comprising:
receiving, by a first computer-implemented agent for a user device, a digital representation of speech encoding an utterance that includes a command; and
in response to processing the digital representation of the speech to determine words included in the utterance:
determining, using the words of the utterance, to request input identifying a sentiment of a user who spoke the utterance;
in response to determining to request input identifying a sentiment of the user who spoke the utterance, providing a request for user input identifying a sentiment;
as a result of providing the request for user input identifying a sentiment, receiving user input identifying a sentiment of a speaker of the utterance;
identifying, by the first computer-implemented agent using the words included in the utterance, a second computer-implemented agent that is most likely able to perform the command, the second computer-implemented agent being a different computer-implemented agent from the first computer-implemented agent; and
in response to identifying the second computer-implemented agent that is most likely able to perform the command, providing, by the first computer-implemented agent to the second computer-implemented agent, data for the utterance and the sentiment of the speaker of the utterance to cause the second computer-implemented agent to provide a response to the command based on the sentiment of the speaker.
20. The computer readable storage medium of claim 19 , wherein:
the first computer-implemented agent executes on a first system;
the second computer-implemented agent executes on a second system that is a different system from the first system; and
providing, by the first computer-implemented agent to the second computer-implemented agent, data for the utterance and the sentiment of the speaker of the utterance to cause the second computer-implemented agent to provide a response to the command based on the sentiment of the speaker comprises providing, by the first computer-implemented agent to the second computer-implemented agent using a network, the data for the utterance and the sentiment of the speaker.