DETAILED ACTION
This communication is in response to the Application filed on 02/28/2025. Claims 1-17 are pending and have been examined. Claims 1, 9 and 13 are independent. This Application was published as US20250307557A1.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 02/28/2025 was filed. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Priority
Acknowledgment is made of applicant’s claim for foreign priority based on application KR10-2024-0044188, KR10-2024-0044217 and KR10-2024-0138312 filed in Korean Intellectual Property Office (KIPO) on 04/01/2024, 04/01/2024, and 10/11/2024, respectively and receipt of a certified copy thereof.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 2, 13 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Whitmore et al. (US Pub 2024/0296275) in view of Cunningham et al. (US Pub 2024/0406124).
Regarding Claim 1,
Whitmore discloses a method for generating a response message using a generative Al model, performed by a computing system (Whitmore, Abstract, "…Systems and methods for using a generative artificial intelligence (AI) model to generate a suggested draft reply to a selected message..."), the method comprising:
acquiring a first message with a first user as a recipient (Fig.2, par [034], "…the message generator 110 and an example data flow 200...A user may use the messaging application 112 to receive messages 222 and to generate content for messages 222...");
automatically generating a second prompt for generating a response to the first message by using the inquired first task-related information and transmitting the second prompt to the generative artificial intelligence-based query service (Fig.2, par [038], "…the message selection causes the message generator 110 to perform a multi-turn process with the generative AI model 108 to generate a suggested draft reply 233 to the selected message 222..."; Fig.4, par [075], "…At decision operation 430, a determination is made as to whether to perform a subsequent query with the generative AI model 108…the message generator 110 generates a subsequent prompt for the generative AI model 108 including the selected editing option(s) 324...
results from the subsequent query are included in a next suggested draft reply 233 that is presented to the user in the application UI 106..."); and
generating a response message to the first message by using a second answer received from the generative artificial intelligence-based query service in response to the transmission of the second prompt (par [076], "…when a selection is made by the user to continue with the displayed suggested draft reply 233, the suggested draft reply 233 is included in a reply message 244 at operation 434...The user views the reply message 244 or edit the reply message 244 until it correctly matches the user's intent and sentiment...The reply message 244 is sent to the recipient(s) at operation 438...").
Whitmore does not explicitly disclose generating dedicated prompt asking a generative AI model to identify the sender's indention nor retrieve the task-related information of the user.
However, Cunningham, in the analogous field of electronic message response generation using an AI system with LLM, discloses automatically generating a first prompt for identifying an intention of the first message and transmitting the first prompt to a generative artificial intelligence-based query service (Cunningham, Fig.3, par [060], "…At the AI system 304d, data from the e-mail...is passed through the LLM with one or more suitable prompts to identify: 1) the "intent" of the received e-mail (i.e., the presumed purpose, goal, or aim that the sender of the e-mail had in mind...2) one or more predetermined data identifiers that identify data stored in the ERP system 306d which is necessary to reply to the received e-mail...");
receiving a first answer from the generative artificial intelligence-based query service in response to the transmission of the first prompt (it is construed that the intent and data-identifier output generated by the AI system 304d is passed back for the data identifier verification/ERP system downstream process as par [060] discloses "...At the AI system 304d, data from the e-mail...is passed through the LLM with one or more suitable prompts...");
inquiring information related to the first answer in an internal database that includes first task-related information of the first user (Fig.5, paras [077-079], "…S505 comprises querying an ERP system with the verified data-identifier data to extract ERP data corresponding to the predetermined data identifiers of which the data-identifier data is indicative..."; Fig.3, paras [056-058], "…Once, the sender and recipient pair have been identified, the client data retrieval module 303d correlates this sender/recipient pair with the stored client data, stored in the client data database 303ad, to identify client data associated with this sender/recipient pair...
The client data typically includes data records relating to previous exchanges between the sender of the received e-mail and the recipient of the received e-mail...").
Therefore, it would have been obvious to one of ordinary skill in the art, before effective filing date of the claimed invention, to have applied generative AI intent-identification and internal task data retrieval technique of Cunningham to a known multi-turn AP replay-drafting system of Whitmore, with reasonable expectation, to have yielded the AI-assisted electronic message response system which can sort, prioritize, and address various inquiries and request from client in order to ensure the transmission of correct information while maintaining efficient communication (Cunningham, Background, paras [003-004]).
Regarding Claim 2,
Whitmore in view of Cunningham discloses the method of claim 1, wherein the generating the response message to the first message includes:
automatically generating a third prompt for generating the response message to the first message by using the second answer and second task-related information of the first user and transmitting the third prompt to the generative artificial intelligence-based query service (Whitmore, Fig.4, par [075], "…the message generator 110 generates a subsequent prompt for the generative AI model 108 including the selected editing option(s) 324...the subsequent prompt is included in a subsequent query provided to the generative AI model 108..."); and
generating the response message to the first message by using a third answer received from the generative artificial intelligence-based query service in response to the transmission of the third prompt (Whitmore, Fig.2, par [038], "…the message selection causes the message generator 110 to perform a multi-turn process with the generative AI model 108 to generate a suggested draft reply 233 to the selected message 222...").
Regarding Claim 13 is a system claim with limitations similar to the limitations of Claim 1 and is rejected under similar rationale. Additionally,
Whitmore discloses a system for generating a response message using a generative Al model, the system comprising: a communication interface; a memory into which a computer program is loaded; and one or more processors in which the computer program is executed, wherein the computer program includes: an operation (Whitmore, Fig.8, paras [107-111], "…the computing device 800 includes at least one processing unit 802...the system memory 804...The system memory 804 may include an operating system 805 and one or more program modules 806...one or more communication connections 816 allowing communications with other computing devices 818...")
…
Rationale for combination is similar to that provided for Claim 1.
Claim 14 is a system claim with limitations similar to the limitations of Claim 2 and is rejected under similar rationale.
Claims 3 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Whitmore in view of Cunningham further in view of Robins (US Pub 2011/0258276).
Regarding Claim 3,
Whitmore in view of Cunningham discloses the method of claim 1 but does not explicitly teach the limitation of "determining an absent state of a messenger of the first user."
Robins, in the analogous field of automatic mail reply, discloses determining an absent state of a messenger of the first user (Robins, Fig.1, paras [022, 025], "…a presence information obtaining unit 102, configured to obtain presence information of the user..."); and
providing the generated response message to a sender of the first message (Robins. Fig.1, par [022], "…an automatic reply message sending unit 104, configured to send the corresponding mail automatic reply message to a sender of the mail..."; Whitmore, Fig.4, par [076], "…the content included in the suggested draft reply 233 is inserted into the body 302 of the reply message 244…The reply message 244 is sent to the recipient(s) at operation 438...").
Therefore, it would have been obvious to one of ordinary skill in the art, before effective filing date of the claimed invention, to have applied generative a presence information obtaining unit of Robins to a known AI-assisted multi-turn electronic message response system of Whitmore in view of Cunningham, with reasonable expectation, to have yielded the AI-assisted electronic message response system withholds or auto-sends the reply based on the user's absence/connected state to streamline user's mail processing (Robins, Background, paras [004]).
Claim 15 is a system claim with limitations similar to the limitations of Claim 3 and is rejected under similar rationale
Claims 4, 5, 7, 16, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Whitmore in view of Cunningham further in view of Robins further in view of Brdiczka et al. (US Pub 2021/0303784).
Regarding Claim 4,
Whitmore in view of Cunningham discloses the method of claim 1, but does not explicitly teach the limitations, "waiting for the first user to provide the response message to the first message for a preset time," and "when the first user does not provide the response message to the first message for the preset time."
Brdiczka, in the analogous field of a natural language input classification from messages, discloses waiting for the first user to provide the response message to the first message for a preset time (Brdiczka, Fig.2, par [037], "…the natural language input classification system 102 detects user input by receiving a user input data stream from the client device 112..."; Fig.4, par [066], "…The natural language input classification system 102 can further assign an "active" chat state to the client device 112 in response to detecting a pause in the user input of sufficient length and an activated input trigger associated with the user input..."); and
Whitmore discloses providing the generated response message to the first message to a sender of the first message (Whitmore, par [076], "… the suggested draft reply 233 is included in a reply message 244 at operation 434...The reply message 244 is sent to the recipient(s) at operation 438...") when the first user does not provide the response message to the first message for the preset time.
Robins teaches a presence-gated automatic message delivery system (Robins, Fig.1, par [023], "…The automatic reply message sending unit 104 is further configured to send the mail automatic reply message corresponding to the presence information of the user to the sender of the mail according to the mail automatic reply time interval...")
It would have been obvious to a person of ordinary skill in the art to apply a natural language input state classification (e.g., composing/active) of Brdiczka to the presence-gated auto-replying system of Whitmore in view of Cunningham further in view of Robins (see claim 3), with the reasonable expectation that the combination would have yielded a system that waits before sending when no input is sensed and withholds the automatic reply when the user is actively typing. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Whitmore, Cunningham, Robins, and Brdiczka to obtain the invention as specified in claim 4.
Regarding Claim 5,
Whitmore in view of Cunningham further in view of Robins further in view of Brdiczka discloses the method of claim 4, further comprising: additionally receiving a second message from the sender of the first message (Robins, Fig.4, par [072], "…Step S408: Determine whether a subsequent mail from the same sender of the mail is received after the new mail is received and before the automatic reply time of the new mail is reached...if the subsequent mail is received, step S407 is performed to...");
generating a response message to the second message; and immediately providing the response message to the second message to the sender of the first message (Robins, par [022], "…an automatic reply message matching unit 103...an automatic reply message sending unit 104, configured to send the corresponding mail automatic reply message to a sender of the mail…").
Regarding Claim 7,
Whitmore in view of Cunningham discloses the method of claim 1 but does not explicitly disclose the limitations of the claim.
However, Brdiczka discloses sensing an input of the response message to the first message of the first user (Brdiczka, par [066], "…the natural language input classification system 102 can assign a "composing" chat state to the client device 112 in response to detecting continuously provided user input (e.g., while the user is typing)..."); and
Robins discloses providing the generated response message to a sender of the first message (Robins, par [022], "…an automatic reply message sending unit 104, configured to send the corresponding mail automatic reply message to a sender of the mail...") when the input of the response message is not sensed (Brdiczka, Fig.4, par [066], "…The natural language input classification system 102 can further assign an "active" chat state to the client device 112 in response to detecting a pause in the user input of sufficient length and an activated input trigger associated with the user input...").
Rationale for combination is similar to that provided for Claim 4.
Claim 16 is a system claim with limitations similar to the limitations of Claim 4 and is rejected under similar rationale.
Claim 17 is a system claim with limitations similar to the limitations of Claim 7 and is rejected under similar rationale.
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Whitmore in view of Cunningham further in view of Robins further in view of Brdiczka further in view of Beaver et al. (US Pub 2024/0040346)
Regarding Claim 6,
Whitmore in view of Cunningham further in view of Robins further in view of Brdiczka discloses the method of claim 4, further comprising: additionally receiving a second message from the sender of the first message (Robins, Fig.4, par [072], "…Step S408: Determine whether a subsequent mail from the same sender of the mail is received after the new mail is received and before the automatic reply time of the new mail is reached...if the subsequent mail is received, step S407 is performed to...");
determining whether it is possible to generate a response message to the second message (Cunningham, Fig.3, par [065], "…the verification module 305d which is configured to perform a verification operation whereby the data identifiers identified by the AI system 304d are compared against the client data identified by the client data retrieval module 303d to verify that the data identifiers do indeed pertain to relevant client data and have not, for example, been "hallucinated" by the AI system 304d….") but is silent on the subsequent steps when the verification module determines that client data is not relevant.
Beaver, in the analogous field of a customer facing intelligent virtual assistant service and user interface, discloses immediately providing a message, which indicates that it is not possible to generate the response message to the second message, to the sender of the first message when it is not possible to generate the response message to the second message (Beaver, Fig.4B par [075], "…an example of how the customer IVA service component 124 may provide numerous communications paths...The customer conversational artificial intelligence component 131 accesses an application database 126 of known entities 416 to determine if an entity is found 418 that matches the first task request data 117 A. If the entity name from the task request data is not found, then the virtual assistant service 103 will ask the user to verify information regarding the entity name 428 and/or advise the user that the entity is not available in the application database 126...").
It would have been obvious to a person of ordinary skill in the art to apply a known lookup-failure notification technique of Beaver to the know database-verification step of Cunningham, with the reasonable expectation of yielding a system immediately notifying when a response cannot be generated because the AI-extracted information wasn't verified against internal data. Therefore, it would have been obvious to person of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Whitmore, Cunningham, Robins, Brdiczka, and Beaver to obtain the invention as specified in claim 6.
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Whitmore in view of Cunningham further in view of Brdiczka.
Regarding Claim 8,
Whitmore in view of Cunningham discloses the method of claim 1 but does not explicitly disclose the limitations of the claim.
Brdiczka discloses sensing an input of the response message to the first message of the first user (Brdiczka, par [066], "…the natural language input classification system 102 can assign a "composing" chat state to the client device 112 in response to detecting continuously provided user input (e.g., while the user is typing)..."); and
displaying a preview of the generated response message to the first user when the input of the response message is sensed (Brdiczka, par [067], "…If the natural language input classification system 102 determines that the current chat state is not "active" (e.g., "No" in the act 406), the natural language input classification system 102 can perform an act 408 of adding the instance of natural language input to a message queue based on a conversation identifier...").
It would have been obvious to a person of ordinary skill in the art to apply a natural language input state classification (e.g., composing/active) of Brdiczka to a known AI-assisted multi-turn electronic message response system of Whitmore in view of Cunningham, with the reasonable expectation that the combination would have yielded a system that waits before sending when no input is sensed and withholds the automatic reply when the user is actively typing. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Whitmore, Cunningham, and Brdiczka to obtain the invention as specified in claim 8.
Claims 9 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Robins further in view of Leiba et al. (US Pub 2018/0287981) further in view of Gardner et al. (US Pub 12,008,332).
Regarding Claim 9,
Robins discloses a method for generating a response message (Abstract, "…method for automatically replying to a mail...") using a generative Al model, performed by a computing system, the method comprising:
determining an absent state of a messenger of a first user (Fig.1, par [022], "…a presence information obtaining unit 102, configured to obtain presence information of the user...");
Robins does not explicitly disclose acquiring content generated during the absence or forming any AI prompt to summarize it.
However, Leiba, in the analogous field of personalized chat summarization, discloses acquiring a first task-related content related to the first user and generated after a time when the absent state of the first user is determined (Leiba, Fig.3, paras [037-038], "…The summary generation sub-system 302 may begin by receiving a set of group chat missed messages 304 that were generated while the user was away…After the missed messages 304 are collected over the time period the user was away, known algorithms may be used to cluster subsets of the missed messages 304 based on topic...");
Therefore, it would have been obvious to one of ordinary skill in the art, before effective filing date of the claimed invention, to have applied a known message acquisition technique of Leiba to a known presence/absence-detection system of Robins with a reasonable expectation of success that the combination would have yielded the predictable results of a system that, upon detecting absence, gathers and clusters the content generated during the period.
But Neither Leiba nor Robins discloses forming a generative-AI prompt to summarize the acquired content or using returned AI-answer to produce the summary.
Gardner, in the analogous field of AI-driven content summarization, discloses automatically generating a fourth prompt for summarizing the acquired task-related content when the messenger of the first user is changed to a connected state (Gardner, Fig.3, col.13:19-36, "…At operation 306, a prompt is automatically engineered for the LLM... the prompt engineering applies techniques like sentence reordering, entity replacement, keyword insertion, example output framing, and instructions guiding the LLM to hit the abstraction targets...");
transmitting the fourth prompt to a generative artificial intelligence-based query service (Gardner, Fig.3, col.13:37-42, "…At operation 308, the prompt is provided to the LLM. The LLM can be implemented using machine learning frameworks like TensorFlow or PyTorch…"); and
generating a summarized missed message by using a fourth answer received from the generative artificial intelligence-based query service in response to the transmission of the fourth prompt (Gardner, Fig.3, col.13:43-47, "…At operation 310, a response is received from the LLM. The response contains a second content item representing the first content item. The representation omits or simplifies sub-content items included in the first content item based on the abstraction level…"; col.48:61-63, "…Email summarization integrations help users digest long, complex email threads and messages by generating concise overviews...").
Therefore, it would have been obvious to one of ordinary skill in the art, before effective filing date of the claimed invention, to have applied a known summary generation technique using prompt engineering referencing a content item and LLM returned answer of Gardner to a known presence/absence-detection and content clustering system of Leiba in view of Robins with a reasonable expectation of success that the combination would have yielded the predictable results of a system that forms, transmits, and answers a summarization prompt upon the connected state trigger.
Regarding Claim 12,
Leiba in view of Robins further in view of Gardner discloses the method of claim 9, further comprising: checking a priority for the first task-related content set in advance (Leiba, Fig.3, par [039], "…At 306, the clustered messages are ranked by topics...the extracted message topics with the user's topics of interest 212 retrieved from the user profile sub-system 202. Based on the topic comparison, a similarity score may be calculated..."); and
providing the first user with a summarized missed message for the task-related content of which priority has been checked (Leiba, Fig.3, par [044], "…After the top messages are picked at 312, the summary generation sub-system 302 presents the top messages as a summary of most relevant messages (i.e., top messages) that the user missed while away from the group chat...").
Claims 10 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over Robins in view of Leiba further in view of Gardner further in view of Qin (US Pub 2024/0346256).
Regarding Claim 10,
Leiba in view of Robins further in view of Gardner discloses the method of claim 9 but does not explicitly disclose receiving of user feedback on the generated summary or using such feedback to generate multi-turn AI prompt/answer.
Qin, in the analogous field of retrieval-augmented-generation (RAG) response system, discloses receiving a feedback related to the summarized missed message from the first user (Qin, Fig.3, par [049], "…In Step 302, a query is received. For instance, pre-processor 202 of response generator 110 may receive query 216...");
inquiring information related to the feedback in an internal database that includes a second task-related content of the first user (Fig.3, paras [051-052], "…in step 306, the first feature vector is compared to a plurality of second feature vectors, each of which corresponding to a piece of augmentation information, to determine second feature vectors that satisfy a predetermined condition with respect to the first feature vector...In step 308, pieces of augmentation information corresponding to the determined second feature vectors are retrieved...retriever 210 may retrieve augmentation information 232 from dataset(s) 112...");
automatically generating a fifth prompt for generating an answer to the feedback by using the inquired second task-related content and transmitting the fifth prompt to the generative artificial intelligence-based query service (Fig.3, par [053], "…In step 310, an augmented prompt is provided to a large language model...prompt generator 212may generate and provide augmented prompt 236 to LLM 214...requests LLM 214 to respond to query 216 based on contextual information 215 using augmentation information 232…"); and
generating a response message to the feedback by using a fifth answer received from the generative artificial intelligence-based query service in response to the transmission of the fifth prompt (par [054], "…In step 312, a response generated by the large language model is received. For instance, GUI manager 108
may receive from response generator 110 a response 238 generated by LLM 214...").
Therefore, it would have been obvious to one of ordinary skill in the art, before effective filing date of the claimed invention, to have applied a known RAG technique of Qin to a known presence/absence-detection, content clustering, and summarization system of Leiba in view of Robins further in view of Gardner with a reasonable expectation of success that the combination would have yielded the predictable results of a system that receives feedback, inquires related content, and generates a further AI response, because Qin's technique performs the identical general function in both context.
Regarding Claim 11,
Leiba in view of Robins further in view of Gardner discloses the method of claim 9.
Qin, in the analogous field of retrieval-augmented-generation (RAG) response system, further discloses receiving a feedback related to the summarized missed message from the first user (Qin, par [049], "…In Step 302, a query is received. For instance, pre-processor 202 of response generator 110 may receive query 216...");
inquiring information related to the feedback in an internal database that includes a second task-related content of the first user (Fig.3, paras [051-052], "…in step 306, the first feature vector is compared to a plurality of second feature vectors, each of which corresponding to a piece of augmentation information, to determine second feature vectors that satisfy a predetermined condition with respect to the first feature vector...In step 308, pieces of augmentation information corresponding to the determined second feature vectors are retrieved...retriever 210 may retrieve augmentation information 232 from dataset(s) 112..."); and
providing the first user with a link connected to the inquired second task-related content (Qin, Fig.2, par [046], "…augmented prompt 236 may include, identify and/or link to augmentation information 232...").
Rationale for combination is similar to that provided for Claim 11.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Rey et al. (US Pat 11,418,463) discloses a method and system for responding to a message directed to a recipient includes receiving the message including a query from a sender, receiving an indication that the recipient is unavailable to respond to the query, and providing the query to as an input to a machine-learning (ML) model to identify information requested in the query. (Rey, Abstract).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JANGWOEN LEE whose telephone number is (703)756-5597. The examiner can normally be reached Monday-Friday 8:00 am - 5:00 pm ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, BHAVESH MEHTA can be reached at (571)272-7453. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/JANGWOEN LEE/ Examiner, Art Unit 2656
/BHAVESH M MEHTA/ Supervisory Patent Examiner, Art Unit 2656
Prosecution Insights