# US 2021/0342547 A1 · System for focused conversation context management in a reasoning agent/behavior engine of an agent automation system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0342547 A1
- Document ID: US20210342547A1
- Status: Published application
- Kind: A1
- Application number: 17/304,918
- Filed: 2021-06-28
- Published: 2021-11-04
- Earliest claimed: 2018-03-23
- Assignees: ServiceNow, Inc.
- Applicants: ServiceNow, Inc.
- Inventors: Edwin Sapugay; Anil Kumar Madamala; Maxim Naboka; Srinivas SatyaSai Sunkara; Lewis Savio Landry Santos; Murali B. Subbarao
- Web page: https://patentgrep.com/patent/US20210342547A1

## CPC

- G06F 40/30: Semantic analysis
- G06F 40/295: Named entity recognition
- G06F 40/35: Discourse or dialogue representation
- G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L 15/26: Speech to text systems (G10L15/08 takes precedence)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

An agent automation system includes a memory configured to store a reasoning agent/behavior engine (RA/BE) including a first persona and a current context and a processor configured to execute instructions of the RA/BE to cause the first persona to perform actions comprising: receiving intents/entities of a first user utterance; recognizing a context overlay cue in the intents/entities of the first user utterance, wherein the context overlay cue defines a time period; updating the current context of the RA/BE by overlaying context information from at least one stored episode associated with the time period; and performing at least one action based on the intents/entities of the first user utterance and the current context of the RA/BE.

## Claims

### Claim 1

An agent automation system, comprising:
at least one memory configured to store a reasoning agent/behavior engine (RA/BE) comprising a plurality of personas, each configured to manage episodic context information using a respective plurality of frames; and
at least one processor configured to execute stored instructions to cause the RA/BE to perform actions comprising:
receiving intents or entities of a first user utterance of a user;
determining a start of an episode in response to receiving the intents or the entities of the first user utterance;
selecting a first persona of the plurality of personas to respond to the first user utterance;
generating, via the first persona, a first response to the first user utterance based at least in part on the intents or the entities;
providing, via the first persona, the first response to the user; and
storing, via the first persona, conversational context information in one or more local frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the conversational context information stored in the one or more local frames when generating a subsequent response to a subsequent user utterance of the user within the episode.

### Claim 2 (depends on 1)

The system of claim 1, wherein the conversational context information comprises a user information, topic information, the first user utterance, the first response, or any combination thereof.

### Claim 3 (depends on 2)

The system of claim 2, wherein the at least one processor is configured to execute the stored instructions to cause the RA/BE to perform actions comprising:
before generating the first response, performing, via the first persona, at least one action based on the intents or the entities of the first user utterance, wherein the conversational context information comprises a description of the at least one action.

### Claim 4 (depends on 1)

The system of claim 1, wherein the at least one processor is configured to execute the stored instructions to cause the RA/BE to perform actions comprising:
requesting and receiving, via the first persona, service-based context information that is contemporaneous with the episode from one or more external services; and
storing, via the first persona, the service-based context information in one or more service-based frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the service-based context information stored in the one or more service-based frames when generating the subsequent response to the subsequent user utterance of the user within the episode.

### Claim 5 (depends on 4)

The system of claim 4, wherein the one or more external services comprise an external weather service, an external stock service, an external news service, an external retail service, or any combination thereof.

### Claim 6 (depends on 1)

The system of claim 1, wherein the at least one processor is configured to execute the stored instructions to cause the RA/BE to perform actions comprising:
requesting and receiving, via the first persona, cross-persona context information from a second persona of the plurality of personas; and
storing, via the first persona, the cross-persona context information in one or more cross-persona frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the cross-persona context information stored in the one or more cross-persona frames when generating the subsequent response to the subsequent user utterance of the user within the episode.

### Claim 7 (depends on 6)

The system of claim 6, wherein the cross-personal context information is retrieved by the second persona from one or more local frames of the second persona for the episode and provided to the first persona via one or more cross-persona scripts.

### Claim 8 (depends on 1)

The system of claim 1, wherein the at least one memory is configured to store a respective persona context database that stores the respective plurality of frames of episodic context information for each of the plurality of personas.

### Claim 9 (depends on 1)

The system of claim 1, wherein the at least one processor is configured to execute the stored instructions to cause the RA/BE to perform actions comprising:
receiving intents or entities of a second user utterance of the user;
determining an end of the episode and a start of a new episode in response to receiving the intents or the entities of the second user utterance;
selecting a second persona of the plurality of personas to respond to the second user utterance;
generating, via the second persona, a second response to the second user utterance based at least in part on the intents or the entities of the second user utterance;
providing, via the second persona, the second response to the user; and
storing, via the second persona, additional conversational context information in one or more additional local frames of the second persona for the new episode, wherein the first persona is configured to use at least a portion of the additional conversational context information stored in the additional one or more local frames when generating another subsequent response to another subsequent user utterance of the user within the new episode.

### Claim 10 (depends on 9)

The system of claim 9, wherein the end of the episode and the start of the new episode is determined based on an amount of time that has passed between receiving the intents or the entities of the first user utterance and receiving the intents or the entities of the second user utterance, based on a change in topic as determined from the intents or the entities of the first user utterance and the intents or the entities of the second user utterance, or any combination thereof.

### Claim 11 (depends on 1)

The system of claim 1, wherein the at least one processor is configured to execute the stored instructions to cause the RA/BE to perform actions comprising:
receiving intents or entities of a second user utterance of the user;
determining that the episode is still in progress in response to receiving the intents or the entities of the second user utterance;
selecting the first persona of the plurality of personas to respond to the second user utterance;
generating, via the first persona, a second response to the second user utterance based at least in part on the intents or the entities of the second user utterance and at least a portion of the conversational context information stored in the one or more local frames of the first persona;
providing, via the first persona, the second response to the user; and
storing, via the first persona, additional conversational context information in one or more additional local frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the additional conversational context information stored in the additional one or more local frames when generating another subsequent response to another subsequent user utterance of the user within the episode.

### Claim 12

A method of operating a reasoning agent/behavior engine (RA/BE) of an agent automation system, comprising:
receiving intents or entities of a first user utterance of a user;
determining a start of an episode in response to receiving the intents or the entities of the first user utterance;
selecting a first persona of a plurality of personas of the RA/BE to respond to the first user utterance;
generating, via the first persona, a first response to the first user utterance based at least in part on the intents or the entities;
providing, via the first persona, the first response to the user; and
storing, via the first persona, conversational context information in one or more local frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the conversational context information stored in the one or more local frames when generating a subsequent response to a subsequent user utterance of the user within the episode.

### Claim 13 (depends on 12)

The method of claim 12, wherein determining the start of the episode comprises:
determining the start of the episode based on an amount of time that has passed since receiving intents or entities of a previous user utterance, based on a change in topic as determined from the intents or the entities of the first user utterance, or any combination thereof.

### Claim 14 (depends on 12)

The method of claim 12, wherein selecting the first persona comprises:
selecting the first persona to respond to the first user utterance in response to determining that the first persona responded to a previous user utterance of the user.

### Claim 15 (depends on 12)

The method of claim 12, wherein selecting the first persona comprises:
selecting the first persona to respond to the first user utterance based on a time at which the intents or the entities of the first user utterance are received, based on determining that the intents or the entities of the first user utterance correspond to a topic that is managed by the first persona, or any combination thereof.

### Claim 16 (depends on 15)

The method of claim 15, wherein the topic is sales and the first persona is a sales persona, or the topic is marketing and the first persona is a marketing persona, or the topic is computer support and the first persona is a computer support persona.

### Claim 17 (depends on 12)

The method of claim 12, comprising:
receiving intents or entities of a second user utterance of the user;
determining that the episode is still in progress based on an amount of time that passed between receiving the intents or the entities of the first user utterance and receiving the intents or the entities of the second user utterance, based on a common topic being represented by the intents or the entities of the first user utterance and the intents or the entities of the second user utterance, or any combination thereof;
selecting the first persona of the plurality of personas to respond to the second user utterance;
generating, via the first persona, a second response to the second user utterance based at least in part on the intents or the entities of the second user utterance, at least a portion of the conversational context information stored in the one or more local frames of the first persona;
providing, via the first persona, the second response to the user; and
storing, via the first persona, additional conversational context information in one or more additional local frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the additional conversational context information stored in the additional one or more local frames when generating another subsequent response to another subsequent user utterance of the user within the episode.

### Claim 18

A non-transitory, computer-readable medium storing instructions of a reasoning agent/behavior engine (RA/BE) of an agent automation system executable by one or more processors of a computing system, wherein the instructions comprise instructions to:
receive intents or entities of a first user utterance of a user;
determine a start of an episode in response to receiving the intents or the entities of the first user utterance;
select a first persona of a plurality of personas of the RA/BE to respond to the first user utterance;
generate, via the first persona, a first response to the first user utterance based at least in part on the intents or the entities;
provide, via the first persona, the first response to the user; and
store, via the first persona, conversational context information in one or more local frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the conversational context information stored in the one or more local frames when generating a subsequent response to a subsequent user utterance of the user within the episode.

### Claim 19 (depends on 18)

The medium of claim 18, wherein the instructions comprising instructions to:
request and receive, via the first persona, service-based context information that is contemporaneous with the episode from one or more external services;
request and receive, via the first persona, cross-persona context information from a second persona of the plurality of personas; and
storing, via the first persona, the service-based context information in one or more service-based frames of the first persona for the episode and storing the cross-persona context information in one or more cross-persona frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the service-based context information stored in the one or more service-based frames and at least a portion of the cross-persona context information stored in the one or more cross-persona frames when generating the subsequent response to the subsequent user utterance of the user within the episode.

### Claim 20 (depends on 19)

The medium of claim 19, wherein the instructions comprising instructions to:
receive intents or entities of a second user utterance of the user;
determine that the episode is still in progress based on an amount of time that passed between receiving the intents or the entities of the first user utterance and receiving the intents or the entities of the second user utterance, based on a common topic being represented by the intents or the entities of the first user utterance and the intents or the entities of the second user utterance, or any combination thereof;
select the first persona of the plurality of personas to respond to the second user utterance;
generate, via the first persona, a second response to the second user utterance based at least in part on the intents or the entities of the second user utterance, at least a portion of the conversational context information stored in the one or more local frames of the first persona;
provide, via the first persona, the second response to the user; and
store, via the first persona, additional conversational context information in one or more additional local frames of the first persona for the episode, wherein the first persona is configured to use at least a portion of the additional conversational context information stored in the additional one or more local frames, at least a portion of the service-based context information stored in the one or more service-based frames, and at least a portion of the cross-persona context information stored in the one or more cross-persona frames when generating another subsequent response to another subsequent user utterance of the user within the episode.
