Automated calling system

1. A method implemented by one or more processors, the method comprising:

receiving, by a computing device, audio data of an utterance spoken by a user who is having a conversation with a bot;

determining, by the computing device, a user intent of a first previous portion of the conversation by the user and a bot intent of a second previous portion of the conversation by the bot, wherein the first previous portion of the conversation occurred prior to receiving the audio data of the utterance, and wherein the second previous portion of the conversation also occurred prior to receiving the audio data of the utterance;

generating, by the computing device, and based on at least the audio data, the user intent; and the bot intent, synthesized speech of a reply by the bot to the utterance; and

causing, by the computing device, the synthesized speech to be provided for audible presentation to the user.

2. The method of claim 1 , further comprising:

determining, by the computing device, a context of the conversation,

wherein generating the synthesized speech of the reply by the bot to the utterance is further based on the context of the conversation.

3. The method of claim 2 , wherein the context of the conversation comprises at least a task associated with the conversation, a time the conversation is initiated, or a location associated with the user.

4. The method of claim 1 , wherein generating the synthesized speech of the reply by the bot to the utterance comprises:

determining, by the computing device, an additional bot intent of the reply by the bot to the utterance;

generating, by the computing device, and based on the additional bot intent, a transcription of the reply by the bot to the utterance; and

generating, by the computing device, and using a speech synthesizer, the synthesized speech of the reply by the bot to the utterance.

5. The method of claim 1 , further comprising:

bypassing, by the computing device, performance of speech recognition on the utterance spoken by the user.

6. The method of claim 1 , further comprising:

accessing, by the computing device, historical data for previous conversations, wherein the historical data includes, for each previous conversation, (i) a previous context of the previous conversation, (ii) previous first speaker intents of portions of the previous conversation spoken by a first speaker, (iii) previous second speaker intents of portions of the previous conversation spoken by a second speaker, (iv) previous audio data of a most recent utterance of the first speaker or the second speaker during the previous conversation, and (v) a previous intent of a previous reply to the most recent utterance; and

training, by the computing device, and based on the historical data, a machine learning model that is configured to receive (i) audio data of a most recent given utterance of a given conversation, (ii) a given user intent of a first portion of the given conversation spoken by a given user, (iii) a given bot intent of a second portion of the given conversation outputted by the bot, and (iv) a given context of the given conversation and output a given intent for a given reply to the most recent given utterance.

7. The method of claim 6 , wherein one or more of the previous conversations comprise previous telephone conversations.

8. The method of claim 6 , wherein generating the synthesized speech of the reply by the bot to the utterance comprises:

processing, by the computing device, and using the machine learning model, at least the audio data, the user intent; and the bot intent to generate output;

determining, by the computing device, and based on the output, an additional bot intent of the reply by the bot to the utterance;

generating, by the computing device, and based on the additional bot intent, a transcription of the reply by the bot to the utterance; and

generating, by the computing device, and using a speech synthesizer, the synthesized speech of the reply by the bot to the utterance.

9. The method of claim 1 , wherein an intent of a portion of the conversation identifies a type of information conveyed in the portion of the conversation.

10. The method of claim 1 , wherein the utterance includes a request to perform a task.

11. The method of claim 10 , further comprising:

based on the conversation:

determining, by the computing device, whether the task has been completed; and

in response to determining that the task has been completed:

causing, by the computing device, the bot to terminate the conversation.

12. The method of claim 11 , further comprising:

in response to determining that the task has not been completed:

causing, by the computing device, the bot to continue the conversation.

13. A system comprising:

at least one processor; and

memory storing instructions that, when executed, cause the at least one processor to perform operations, the operations comprising:

receiving, by a computing device, audio data of an utterance spoken by a user who is having a conversation with a bot;

determining, by the computing device, a user intent of a first previous portion of the conversation by the user and a bot intent of a second previous portion of the conversation by the bot, wherein the first previous portion of the conversation occurred prior to receiving the audio data of the utterance, and wherein the second previous portion of the conversation also occurred prior to receiving the audio data of the utterance;

generating, by the computing device, and based on at least the audio data, the user intent; and the bot intent, synthesized speech of a reply by the bot to the utterance; and

causing, by the computing device, the synthesized speech to be provided for audible presentation to the user.

14. The system of claim 13 , the operations further comprising:

determining, by the computing device, a context of the conversation,

wherein generating the synthesized speech of the reply by the bot to the utterance is further based on the context of the conversation.

15. The system of claim 14 , wherein the context of the conversation comprises at least a task associated with the conversation, a time the conversation is initiated, or a location associated with the user.

16. The system of claim 13 , wherein generating the synthesized speech of the reply by the bot to the utterance comprises:

determining, by the computing device, an additional bot intent of the reply by the bot to the utterance;

generating, by the computing device, and based on the additional bot intent, a transcription of the reply by the bot to the utterance; and

generating, by the computing device, and using a speech synthesizer, the synthesized speech of the reply by the bot to the utterance.

17. The system of claim 13 , the operations further comprising:

bypassing, by the computing device, performance of speech recognition on the utterance spoken by the user.

18. The system of claim 13 , the operations further comprising:

accessing, by the computing device, historical data for previous conversations, wherein the historical data includes, for each previous conversation, (i) a previous context of the previous conversation, (ii) previous first speaker intents of portions of the previous conversation spoken by a first speaker, (iii) previous second speaker intents of portions of the previous conversation spoken by a second speaker, (iv) previous audio data of a most recent utterance of the first speaker or the second speaker during the previous conversation, and (v) a previous intent of a previous reply to the most recent utterance; and

training, by the computing device, and based on the historical data, a machine learning model that is configured to receive (i) audio data of a most recent given utterance of a given conversation, (ii) a given user intent of a first portion of the given conversation spoken by a given user, (iii) a given bot intent of a second portion of the given conversation outputted by the bot, and (iv) a given context of the given conversation and output a given intent for a given reply to the most recent given utterance.

19. The system of claim 18 , wherein generating the synthesized speech of the reply by the bot to the utterance comprises:

processing, by the computing device, and using the machine learning model, at least the audio data, the user intent; and the bot intent to generate output;

determining, by the computing device, and based on the output, an additional bot intent of the reply by the bot to the utterance;

generating, by the computing device, and based on the additional bot intent, a transcription of the reply by the bot to the utterance; and

generating, by the computing device, and using a speech synthesizer, the synthesized speech of the reply by the bot to the utterance.

20. A non-transitory computer-readable storage medium storing instructions that, when executed cause at least one processor to perform operations, the operations comprising:

receiving, by a computing device, audio data of an utterance spoken by a user who is having a conversation with a bot;

determining, by the computing device, a user intent of a first previous portion of the conversation by the user and a bot intent of a second previous portion of the conversation by the bot, wherein the first previous portion of the conversation occurred prior to receiving the audio data of the utterance, and wherein the second previous portion of the conversation also occurred prior to receiving the audio data of the utterance;

generating, by the computing device, and based on at least the audio data, the user intent; and the bot intent, synthesized speech of a reply by the bot to the utterance; and

causing, by the computing device, the synthesized speech to be provided for audible presentation to the user.