Selecting agents to respond to user queries

1 . A method for routing user requests from an automated assistant associated with an online resource, the method performed by one or more processors of a computing system associated with the online resource and comprising:

receiving, from the user over a communications network coupled to the computing system, a request for an automated assistant;

initiating a conversation, over the communications network, between the user and the automated assistant in response to the request;

identifying a plurality of queries from the user during a portion of the conversation;

determining a context for each of the plurality of queries;

selecting, for each of the plurality of queries, one agent of a plurality of agents based on the determined context for the respective query by:

comparing the context for the respective query with agent descriptions of the plurality of agents;

determining a degree of similarity between the context and each of the agent descriptions of the plurality of agents based on the comparisons; and

selecting the agent associated with the highest degree of similarity to generate the response for a respective sub-query;

sending each of the plurality of queries to a respective agent of the selected agents; and

receiving, from each of the selected agents, a response to the respective query of the plurality of queries, wherein each of the plurality of agents is associated with a corresponding large language model (LLM) trained using query-and-response training data associated with a unique context or a unique group of contexts.

2 . The method of claim 1 , wherein the context is based at least in part on one or more previous portions of the conversation.

3 . The method of claim 1 , wherein the context includes a browsing history of the user within a user assistance page or web site associated with the online resource.

4 . The method of claim 1 , wherein the context is based at least in part on a type of application through which the user sends the request to the online resource.

5 . The method of claim 1 , wherein different agents of the plurality of agents are configured to generate responses to different queries associated with different contexts or different groups of contexts.

6 . The method of claim 1 , further comprising:

combining the responses from the selected agents into an answer responsive to the plurality of queries; and

transmitting the answer to the user over the communications network.

7 . The method of claim 6 , further comprising:

presenting the answer to the user as part of the conversation between the automated assistant and the user.

8 . A computing system associated with an online resource, the computing system comprising:

one or more processors; and

a memory communicatively coupled with the one or more processors and storing instructions that, when executed by the one or more processors, causes the computing system to:

receive, from the user over a communications network coupled to the computing system, a request for an automated assistant;

initiate a conversation, over the communications network, between the user and the automated assistant in response to the request;

identify a plurality of queries from the user during a portion of the conversation;

determine a context for each of the plurality of queries;

select, for each of the plurality of queries, one agent of a plurality of agents based on the determined context for the respective query by:

comparing the context for the respective query with agent descriptions of the plurality of agents;

determining a degree of similarity between the context and each of the agent descriptions of the plurality of agents based on the comparisons; and

selecting the agent associated with the highest degree of similarity to generate the response for a respective sub-query;

send each of the plurality of queries to a respective agent of the selected agents; and

receive, from each of the selected agents, a response to the respective query of the plurality of queries, wherein each of the plurality of agents is associated with a corresponding large language model (LLM) trained using query-and-response training data associated with a unique context or a unique group of contexts.

9 . The computing system of claim 8 , wherein the context is based at least in part on one or more previous portions of the conversation.

10 . The computing system of claim 8 , wherein the context includes a browsing history of the user within a user assistance page or web site associated with the online resource.

11 . The computing system of claim 8 , wherein the context is based at least in part on a type of application through which the user sends the request to the online resource.

12 . The computing system of claim 8 , wherein different agents of the plurality of agents are configured to generate responses to different queries associated with different contexts or different groups of contexts.

13 . The computing system of claim 8 , wherein execution of the instructions further causes the computing system to:

combine the responses from the selected agents into an answer responsive to the plurality of queries; and

transmit the answer to the user over the communications network.

14 . The computing system of claim 13 , wherein execution of the instructions further causes the computing system to:

present the answer to the user as part of the conversation between the automated assistant and the user.