# US 12,224,968 B2 · Systems and methods for analyzing chatbot communication sessions to reduce escalation > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,224,968 B2 - Document ID: US12224968B2 - Status: Granted - Kind: B2 - Application number: 17/456,386 - Filed: 2021-11-24 - Published: 2023-05-25 - Granted: 2025-02-11 - Earliest claimed: 2021-11-24 - Assignees: Verizon Patent and Licensing Inc. - Applicants: Verizon Patent and Licensing Inc. - Inventors: Lakshmi Ramanathan; Johncy Arul Anburose - Web page: https://patentgrep.com/patent/US12224968B2 ## CPC - G06F 40/30: Semantic analysis - H04L 51/02: using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages - G06F 40/35: Discourse or dialogue representation - G06N 20/00: Machine learning - H04L 51/216: Handling conversation history, e.g. grouping of messages in sessions or threads - G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract A device may receive session data of a communication session between an artificial intelligence (AI) communication device and a first user device. The device may analyze the session data to determine one or more portions of the session data and may identify a portion, of the one or more portions, for processing based on one or more criteria associated with the portion. The portion may include a plurality of entries including communications from the first user device and from the AI communication device. The device may analyze the plurality of entries to identify an entry including information regarding an escalation and may analyze one or more additional entries, of the plurality of entries, to determine a category associated with a cause of the escalation. The device may cause the AI communication device to be configured to address the cause based on the one or more additional entries and the category. ## Claims ### Claim 1 A method performed by a device, the method comprising: receiving session data of a communication session between an artificial intelligence (AI) communication device and a first user device; analyzing the session data to determine one or more portions of the session data; identifying a portion, of the one or more portions, for processing based on one or more criteria associated with the portion, wherein the portion includes a plurality of entries including communications from the first user device and communications from the AI communication device; analyzing the plurality of entries to identify an entry including information regarding an escalation, wherein the escalation indicates that the first user device has been connected to a second user device; analyzing one or more additional entries of the plurality of entries, in a reverse chronological order, to determine a category associated with a cause of the escalation, wherein analyzing the one or more additional entries comprises: analyzing a first next entry. preceding the entry, when the first next entry includes a first communication from the first user device and a second communication from the Al communication device and identifying a second next entry, preceding the first next entry, when the first next entry does not include the first communication and the second communication; and causing the AI communication device to be configured to address the cause of the escalation based on information regarding the one or more additional entries and information identifying the category. ### Claim 2 (depends on 1) The method of claim 1, wherein the AI communication device executes a chatbot, wherein receiving the session data comprises: receiving messages exchanged between the chatbot and the first user device, wherein the plurality of entries includes a subset of messages of the messages exchanged between the chatbot and the first user device, and wherein analyzing the plurality of entries comprises: analyzing the subset of messages to identify a message including the information indicating the escalation. ### Claim 3 (depends on 1) The method of claim 1, wherein the entry and the one or more additional entries are provided in a chronological order in the session data, wherein the one or more additional entries are provided prior to the entry, and wherein analyzing the one or more additional entries includes: analyzing the one or more additional entries in the reverse chronological order to determine the category associated with the cause of the escalation. ### Claim 4 (depends on 3) The method of claim 3, wherein the one or more criteria include a topic, wherein the first communication includes a request that includes information associated with the topic, wherein the second communication includes a response that includes information indicating that the AI communication device has identified a plurality of possible responses to the request, and wherein analyzing the one or more additional entries comprises determining the category based on the information associated with the topic. ### Claim 5 (depends on 1) The method of claim 1, wherein the portion is a first portion and the one or more criteria include a first topic, wherein identifying the portion for processing comprises: determining that the first topic is a subtopic of a second topic associated with a second portion of the one or more portions; and identifying the portion for processing based on determining that the first topic is a subtopic of the second topic. ### Claim 6 (depends on 1) The method of claim 1, wherein analyzing the session data comprises: analyzing the session data to identify at least one of: a period of time between a communication from the first user device and a communication from the AI communication device, a request to communicate with the second user device, or an indication of a graphical user interface being provided to the first user device after a threshold amount of time following initiation of the communication session. ### Claim 7 (depends on 1) The method of claim 1, wherein identifying the portion for processing comprises: identifying the portion based on the portion including: information indicating that the first user device has been connected to the second user device, information indicating that the first user device has provided a request to be connected to the second user device, or information indicating that the AI communication device has provided a suggestion, to the first user device, indicating that the first user device is to be connected to the second user device. ### Claim 8 A device, comprising: one or more processors configured to: receive session data regarding a communication session that includes a plurality of messages exchanged between a chatbot and a first user device; identify a portion of the session data, of one or more portions of the session data, for processing based on one or more criteria associated with the portion of the session data, wherein the portion of the session data includes a plurality of entries including messages from the first user device and messages from the chatbot; analyze the plurality of entries to identify a particular entry including escalation information regarding an escalation, wherein the escalation information indicates that the first user device has been connected to a second user device; analyze one or more additional entries, in a reverse chronological order, to determine a cause of the escalation, wherein, to analyze the one or more additional entries, the one or more processors are to: analyze a first next entry, preceding the particular entry, when the first next entry includes a first message from the first user device and a second message from the chatbot, and identify a second next entry, preceding the first next entry, when the first next entry does not include the first message and the second message; and cause the chatbot to be trained to address the cause and prevent another escalation, associated with the cause, during a subsequent communication session involving the chatbot, wherein the chatbot is trained based on first information regarding the cause and second information regarding one of the first next entry of the second next entry of the one or more additional entries. ### Claim 9 (depends on 8) The device of claim 8, wherein the one or more processors, to receive the session data regarding the communication session, are configured to: receive a transcript of the communication session, and wherein, prior to identifying the portion of the session data, the one or more processors are further configured to: perform data processing on the transcript of the communication session to at least one of: remove one or more characters, or identify a beginning of the communication session and an ending of the communication session. ### Claim 10 (depends on 8) The device of claim 8, wherein the portion of the session data is a first portion of the session data, and wherein the one or more processors, to identify the portion of the session data, are configured to: analyze the first portion of the session data, using a natural language processing technique, to determine a first topic associated with the first portion of the session data; analyze a second portion of the session data, of the one or more portions of the session data, using the natural language processing technique, to determine a second topic associated with the second portion of the session data; determine a first priority associated with the first topic and a second priority associated with the second topic; and identify the first portion of the session data for processing based on determining the first priority and the second priority. ### Claim 11 (depends on 10) The device of claim 10, wherein the one or more processors, to identify the first portion of the session data, are configured to: determine that the first priority is higher than the second priority; and identify the first portion of the session data for processing based on determining that the first priority exceeds the second priority. ### Claim 12 (depends on 8) The device of claim 8, wherein the first-particular entry and the one or more additional entries are provided in a chronological order in the session data, wherein the one or more additional entries are provided prior to the particular entry, wherein the one or more processors, to analyze the one or more additional entries to determine the cause of the escalation, are configured to: analyze the one or more additional entries, in the reverse chronological order, to determine the cause of the escalation after identifying the particular entry. ### Claim 13 (depends on 8) The device of claim 8, wherein the one or more processors, to analyze the one or more additional entries, are configured to: analyze the one of the first next entry of the second next entry based on determining that the one of the first next entry or the second next entry includes the first message from the first user device and the second message from the chatbot, wherein the chatbot is trained based on a machine learning model, used by the chatbot being trained to determine responses to request messages that caused escalations, and wherein the machine learning model is trained using historical data that includes the first message. ### Claim 14 (depends on 8) The device of claim 8, wherein the one or more criteria include at least one of a topic or a period of time; wherein the one or more processors are further configured to: receive information identifying the topic; and wherein the one or more processors, to identify the portion of the session data, are configured to: analyze the portion of the session data, using a natural language processing technique, to determine that the portion of the session data is associated with the topic. ### Claim 15 A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising: one or more instructions that, when executed by one or more processors of a device, cause the device to: receive session data regarding a communication session that includes a plurality of messages exchanged between a chatbot and a first user device; identify a portion of the session data, of one or more portions of the session data, for processing based on one or more criteria associated with the portion of the session data, wherein the portion of the session data includes a plurality of entries including messages from the first user device and messages from the chatbot; analyze the plurality of entries to identify an entry including escalation information regarding an escalation, wherein the escalation indicates that the first user device has been connected to a second user device; analyze one or more additional entries, of the plurality of entries, to determine a cause of the escalation, wherein the one or more additional entries are analyzed in a reverse chronological order, wherein the one or more instructions to analyze the one or more additional entries comprise one or more instructions to: analyze a first next entry, preceding the entry, when the first next entry includes a first message from the first user device and a second message from the chatbot, and identify a second next entry. g the first next entry, whe the first next entry does not include the first message and the second message; and cause the chatbot to be trained to address the cause during a subsequent communication session involving the chatbot, wherein the chatbot is trained based on information identifying the one or more additional entries and information regarding the cause of the escalation. ### Claim 16 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the one or more instructions, that cause the device to analyze the one or more additional entries, cause the device to: determine that the first next entry does not include the first message and the second message; and analyze the second next entry based on determining that the second next entry includes the first message device and the second message. ### Claim 17 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the one or more instructions, that cause the device to analyze the one or more additional entries, cause the device to: obtain one or more rules associated with identifying information regarding causes of escalations; and analyze the one or more additional entries to determine the cause of the escalation based on the one or more rules. ### Claim 18 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the one or more instructions, when executed by the one or more processors, further cause the device to: analyze the session data to identify at least one of: a period of time between a communication from the first user device and a communication from the chatbot, a request to communicate with the second user device, or an indication of a graphical user interface being provided to the first user device after a threshold amount of time following initiation of the communication session; and identify the one or more portions of the session data based on analyzing the session data. ### Claim 19 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the one or more instructions, that cause the device to analyze the plurality of entries, cause the device to: identify the entry based on the entry including: information indicating that the first user device has been connected to the second user device, information indicating that the first user device has provided a request to be connected to the second user device, or information indicating that the chatbot has provided a suggestion, to the first user device, indicating that the first user device is to be connected to the second user device. ### Claim 20 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the one or more instructions, when executed by the one or more processors, further cause the device to: receive a transcript of the communication session, and perform data processing on the transcript of the communication session to at least one of: remove one or more words or one or more characters, or identify a beginning of the communication session and an ending of the communication session.