Granted Patent B2

US 12,380,280 · App. 17/946,952 · Granted Aug 5, 2025

Inventors: Mitchell McNeill (Dallas, TX); Neil Brady (Indianapolis, IN); Nathan Smith (Denver, CO)

Assignee: Salesforce, Inc.

G06F40/35G10L25/30H04L51/02

Abstract

Claims (84)

1. A method, implemented at least in part by one or more computing devices of a communication platform, the method comprising:

receiving, via a virtual space and from a user associated with a user profile of the communication platform, a request to obtain information;

inputting the request into a trained machine learning model trained to determine a plurality of topics associated with the request and respective confidence levels associated with the plurality of topics, wherein the trained machine learning model is trained on content stored within the virtual space;

receiving, from the trained machine learning model, the plurality of topics and the respective confidence levels;

determining that an individual confidence level associated with an individual topic is a highest confidence level of the respective confidence levels;

determining, based at least in part on the individual confidence level being the highest confidence level, that the individual confidence level meets or exceeds a threshold confidence level;

based least in part on the individual confidence level meeting or exceeding the threshold confidence level, associating the individual topic with the request;

determining that the individual topic of the request corresponds to a graphical identifier;

causing to display, in response to determining that the individual topic corresponds to the graphical identifier and based at least in part on the request being associated with the individual topic, the graphical identifier on the virtual space; and

causing to display, based at least in part on the graphical identifier to be displayed, a workflow response on the virtual space.

2. The method of claim 1 , wherein the trained machine learning model is a first trained machine learning model and the virtual space is a first virtual space, further comprising:

receiving, via a second virtual space and from a second user associated with a second user profile, a second request to obtain information, wherein the second virtual space is different from the first virtual space;

inputting the second request into a second trained machine learning model trained to determine a second plurality of topics associated with the second request and second respective confidence levels associated with the second plurality of topics, wherein the second trained machine learning model is different than the first trained machine learning model;

receiving, from the second trained machine learning model, the second plurality of topics and the second respective confidence levels; and

causing to display, based at least in part on the second plurality of topics and the second respective confidence levels, a second workflow response to the second virtual space, wherein the second workflow response is different than the workflow response.

3. The method of claim 1 , wherein the individual topic is a first individual topic and the individual confidence level is a first individual confidence level, further comprising:

determining, based at least in part on the request, that the first individual topic and a second individual topic are associated with the request;

determining that the first individual topic is associated with the first individual confidence level and the second individual topic is associated with a second individual confidence level; and

selecting, based at least in part on the first individual confidence level exceeding the second individual confidence level, the first individual topic.

4. The method of claim 1 , wherein the graphical identifier is a first graphical identifier, further comprising:

receiving, via the virtual space and from the user, an indication of performance of the user; and

causing to display, based at least in part on the indication of performance, a second graphical identifier to the virtual space, wherein the second graphical identifier is different than the first graphical identifier.

5. The method of claim 1 , wherein the request comprises a question associated with the virtual space and the trained machine learning model comprises a natural language processing component.

6. The method of claim 1 , wherein the trained machine learning model is trained based at least in part on:

first interaction data comprising prior topics associated with prior user requests; and

second interaction data comprising prior confidence levels associated with the prior topics, to learn relationships between the first interaction data and the second interaction data, wherein the trained machine learning model is configured to use the learned relationships to generate the plurality of topics associated with the request and the respective confidence levels associated with the plurality of topics.

7. The method of claim 1 , wherein the virtual space comprises a channel, a thread, or a board.

8. The method of claim 1 , wherein the workflow response comprises a second request to access a first party resource associated with the communication platform or a remote third-party resource.

9. A system comprising:

one or more processors; and

one or more non-transitory computer-readable media storing instructions that, when executed by the one or more processors, cause the system to perform operations comprising:

receiving, via a virtual space and from a user associated with a user profile, a request to obtain information;

inputting the request into a trained machine learning model trained to determine that a topic is associated with the request, wherein the trained machine learning model is trained on content stored within the virtual space;

receiving, from the trained machine learning model, an output that includes a plurality of confidence levels;

determining that a confidence level of the plurality of confidence levels is a highest confidence level;

determining, based at least in part on the confidence level being the highest confidence level, that the confidence level meets or exceeds a threshold confidence level;

based at least in part on the output and the confidence level meeting or exceeding the threshold confidence level, associating the request with the topic;

determining that the topic of the request corresponds to a graphical identifier;

causing to display, in response to determining that the topic corresponds to the graphical identifier and based at least in part on the request being associated with the topic, the graphical identifier to the virtual space; and

causing to display, based at least in part on the graphical identifier to be displayed, a response to the virtual space.

10. The system of claim 9 , wherein the trained machine learning model is a first trained machine learning model, the virtual space is a first virtual space, and the response is a first workflow response, the operations further comprising:

receiving, via a second virtual space and from a second user associated with a second user profile, a second request to obtain information, wherein the second virtual space is different from the first virtual space;

inputting the second request into a second trained machine learning model trained to determine a second plurality of topics associated with the second request, wherein the second trained machine learning model is different than the first trained machine learning model;

receiving, from the second trained machine learning model, the second plurality of topics; and

causing to display, based at least in part on the second plurality of topics, a second workflow response to be displayed to the second virtual space, wherein the second workflow response is different than the first workflow response.

11. The system of claim 9 , wherein the topic is a first topic, the operations further comprising:

determining, based at least in part on the output, a first confidence level associated with the first topic;

determining, based at least in part on the request, that the first topic and a second topic are associated with the request;

determining that the first topic is associated with the first confidence level and the second topic is associated with a second confidence level; and

selecting, based at least in part on the first confidence level exceeding the second confidence level, the first topic.

12. The system of claim 9 , wherein the graphical identifier is a first graphical identifier, the operations further comprising:

receiving, via the virtual space and from the user, an indication of performance of the user; and

causing to display, based at least in part on the indication of performance, a second graphical identifier to the virtual space, wherein the second graphical identifier is different than the first graphical identifier.

13. The system of claim 9 , wherein the trained machine learning model is trained based at least in part on:

first interaction data comprising prior topics associated with prior user requests; and

second interaction data comprising prior confidence levels associated with the prior topics, to learn relationships between the first interaction data and the second interaction data, wherein the trained machine learning model is configured to use the learned relationships to generate the topic associated with the request and a respective confidence level associated with the topic.

14. The system of claim 9 , wherein the virtual space comprises a channel, a thread, or a board.

15. One or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to:

receiving, via a virtual space and from a user associated with a user profile, a request to obtain information;

inputting the request into a trained machine learning model trained to determine that a topic is associated with the request, wherein the trained machine learning model is trained on content stored within the virtual space;

receiving, from the trained machine learning model, an output that includes a plurality of confidence levels;

determining that a confidence level of the plurality of confidence levels is a highest confidence level;

determining, based at least in part on the confidence level being the highest confidence level, that the confidence level meets or exceeds a threshold confidence level;

based at least in part on the output and the confidence level meeting or exceeding the threshold confidence level, associating the request with the topic;

determining that the topic of the request corresponds to a graphical identifier;

causing to display, in response to determining that the topic corresponds to the graphical identifier and based at least in part on the request being associated with the topic, the graphical identifier to the virtual space; and

causing to display, in response to displaying the graphical identifier and in response to identifying a response associated with the graphical identifier, the response to the virtual space.

16. The one or more non-transitory computer-readable media of claim 15 , wherein the trained machine learning model is a first trained machine learning model, the virtual space is a first virtual space, and the response is a first workflow response, further comprising:

receiving, via a second virtual space and from a second user associated with a second user profile, a second request to obtain information, wherein the second virtual space is different from the first virtual space;

inputting the second request into a second trained machine learning model trained to determine a second plurality of topics associated with the second request, wherein the second trained machine learning model is different than the first trained machine learning model;

receiving, from the second trained machine learning model, the second plurality of topics; and

causing to display, based at least in part on the second plurality of topics, a second workflow response to be displayed to the second virtual space, wherein the second workflow response is different than the first workflow response.

17. The one or more non-transitory computer-readable media of claim 15 , wherein the topic is a first topic, further comprising:

determining, based at least in part on the output, a first confidence level associated with the first topic;

determining, based at least in part on the request, that the first topic and a second topic are associated with the request;

determining that the first topic is associated with the first confidence level and the second topic is associated with a second confidence level; and

selecting, based at least in part on the first confidence level exceeding the second confidence level, the first topic.

18. The one or more non-transitory computer-readable media of claim 15 , wherein the graphical identifier is a first graphical identifier, further comprising:

receiving, via the virtual space and from the user, an indication of performance of the user; and

causing to display, based at least in part on the indication of performance, a second graphical identifier to the virtual space, wherein the second graphical identifier is different than the first graphical identifier.

19. The one or more non-transitory computer-readable media of claim 15 , wherein the trained machine learning model is trained based at least in part on:

first interaction data comprising prior topics associated with prior user requests; and

second interaction data comprising prior confidence levels associated with the prior topics, to learn relationships between the first interaction data and the second interaction data, wherein the trained machine learning model is configured to use the learned relationships to generate the topic associated with the request and a respective confidence level associated with the topic.

20. The one or more non-transitory computer-readable media of claim 15 , wherein the virtual space comprises a channel, a thread, or a board.

Assignments (1)

Continuity (1)

References Cited (20)