DETAILED ACTION
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 .
Claim Rejections - 35 USC § 103
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.
Claim(s) 1-9, 13-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Bell et al. (US 20240406166) in view of Ho et al. (US 20250291854) and in further view of STOILOS et al. (US 20200242133).
Regarding claim 1, Bell teaches a system, comprising: one or more processors coupled with memory to:
generate, using one or more machine learning (ML) models ([0237]) and based on a query associated with an account of a processing framework ([0179] “uses information about the patient or subject (e.g., age, gender, medical history, current medications, etc.) as context information for the query … uses information about the user (e.g., a medical professional) as context information for the query”, [0358] “restrict access to each document collection based on user authentication data”, [0449], [0453], [0457] “comparing the user identifier with the permission data”), an embedding corresponding to a vector representation of the query ([0149] “sends the query to an embedding component … to generate an embedding based on one or more portions of the query text received from the client”, [0157]);
perform, using the embedding, a vector semantic ([0153] “semantic comparison of embeddings”, [0205] “identifies semantically similar documents and/or text”) search in a vector space of a plurality of queries and a plurality of question response pairs corresponding to a set of documentation ([0108] “knowledge (e.g., scripted questions and answers on various frequently asked questions”, [0152] “data from the knowledge database (e.g., a question-and-answer dataset)” [0184] “use a knowledge base and/or question and answer pairs”, [0335]) and a validation status ([0144], [0181] “validate the data to ensure it is appropriate”, [0446]) to identify documentation associated with the query and
identify, from a knowledge database more ML models ([0150] “embeddings are generated from data in a knowledge database and the embeddings are stored in a vector database … documents from the knowledge database are split into snippets and the snippets are tokenized and added to the vector database”), using the documentation and metadata associated with the account, one or more entities related to the question response pair and one or more relationships between the one or more entities ([0242], [0244], [0397], [0405]),
wherein the knowledge database
select, from a plurality of agents of the processing framework, based on the one or more entities and the one or more relations, an agent to provide an interaction with a client device to address the one or more entities ([0179] “a routing agent module configured to route subsets of the commands and/or tokens to appropriate agent modules (e.g., based on the query intent, previous interactions, and/or information about the particular user or subject)”; [0181] “receive/parse user queries and commands, a super-agent module to identify the appropriate agent modules and/or tools for responding to respective queries and/or commands, and a routing agent module to transmit information to and from the identified modules and tools”, [0236], [0499]) and the one or more relationships for the question response pair ([0090] “node architecture includes a plurality of paths to traverse from an input to an output node, such as paths of branching trees … each node is interconnected, such by an edge, to at least one other node, the output from one node may be supplied as input to a different node in order to form chains, or orders, or nodes in the node architecture”, [0242], [0528] “distance between reference entities in the same cluster can be significantly less than the distance between the reference entities in different cluster”); and
provide, via the processing framework to the client device, a response associated with the query based on the question response pair, responsive to the interaction ([0181] “generate a complete response to a respective que … transform the complete response to a natural language response and to provide the natural language response to the user”; [0184] “use a knowledge base and/or question and answer pairs”, [0228]).
Although Bell teaches node architecture and traversing nodes in an otology tree ([0090], [0219]), Bell does not explicitly teach, however Ho discloses –
identify, from a knowledge graph comprising entity and relationship data curated from the set of documentation according to the validation status using the one or more ML models ([0014], [0060]), using the documentation and metadata associated with the account ([0053]), one or more entities related to the question response pair ([0071] “generate potential questions and corresponding answers associated with the document using AI-techniques or … stored historical data of previous questions, previous answers”) and one or more relationships between the one or more entities ([0055]), wherein the knowledge graph defines relationships between entities associated with the queries and responses associated with the queries ([0054], [0079], [0081]).
Further, note Ho explicitly teaches a vector space of a plurality of queries ([0014]-[0015], [0035], [0072]), one or more entities and the one or more relations and one or more relationships for the question response pair ([0015], [0054], [0059], [0084]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Bell to include a knowledge graph as disclosed by Ho. Doing so would improve efficiency and efficacy of LLM functionality for a QA system (Ho [0014]).
Bell does not explicitly teach, however STOILOS discloses – identify documentation associated with the query and a matching question response pair ([0037]-[0038], [0110]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Bell to identify documentation associated with the query and a matching question response pair as disclosed by STOILOS. Doing so would help bridge the gap between user queries and a set of pre-defined concepts in order to try and reduce the set of candidates to eventually a single concept that captures the initial user intention (STOILOS [0160]).
Regarding claim 13, Bell teaches a method, comprising: generating, by one or more processors coupled with memory, using one or more machine learning (ML) models and based on a query associated with an account of a processing framework, an embedding corresponding to a vector representation of the query; performing, by the one or more processors, using the embedding, a vector semantic search in a vector space of a plurality of queries and a plurality of question response pairs corresponding to a set of documentation and a validation status to identify documentation associated with the query and a matching question response pair, the vector space generated using the one or more ML models; identifying, by the one or more processors, from a knowledge graph comprising entity and relationship data curated from the set of documentation according to the validation status using the one or more ML models, using the documentation and metadata associated with the account, one or more entities related to the question response pair and one or more relationships between the one or more entities, wherein the knowledge graph defines relationships between entities associated with the queries and responses associated with the queries; selecting, by the one or more processors, from a plurality of agents of the processing framework, based on the one or more entities and the one or more relations, an agent to provide an interaction with a client device to address the one or more entities and the one or more relationships for the question response pair; and providing, by the one or more processors, via the processing framework to the client device, a response of the query based on the question response pair, responsive to the interaction.
Claim 13 recites substantially the same limitations as claim 1, and is rejected for substantially the same reasons.
Regarding claims 2 and 14, Bell as modified teaches the system and the method, wherein the one or more processors further:
determine, based on at least one of the embedding or the one or more entities identified in the knowledge graph, that the query is ambiguous (Bell [0184] “allows for open-ended questions answering”, Ho [0034], [0038], STOILOS [0061], [0084]);
generate, via the selected agent, one or more follow-up questions to the client device in response to the determination that the query is ambiguous (STOILOS [0069]-[0070], [0107], [0116]);
receive, via the selected agent, one or more follow-up responses to the one or more follow-up questions (STOILOS [0116], [0119], [0162] “construct on the fly, a small dialogue that asks the user a few “clarification” questions in an attempt to “activate” the proper entity from the KB”);
identify, based on the one or more follow-up responses and using the knowledge graph (STOILOS [0110]), a refined matching question response pair from the plurality of question response pairs (STOILOS [0087]); and provide, to the client device via the processing framework, the response to the query response pair comprising the refined matching question response pair (STOILOS [0103], [0156]-[0157], Ho [0039], [0056], [0069]).
Regarding claims 3 and 15, Bell as modified teaches the system and the method, wherein the one or more processors are configured to identify the matching question response pair based on a vector semantic search between the embedding and a vector of a question of the matching question response pair satisfying a similarity threshold (Bell [0149], [0153], Ho [0069], [0072], STOILOS [0036], [0094]).
Regarding claims 4 and 16, Bell as modified teaches the system and the method, wherein the one or more processors are configured to:
identify, based on the vector semantic search, a plurality of question response pairs corresponding to the embedding (Bell [0108] “knowledge (e.g., scripted questions and answers on various frequently asked questions”, [0152] “data from the knowledge database (e.g., a question-and-answer dataset)” [0184] “use a knowledge base and/or question and answer pairs”, [0335], STOILOS [0037]-[0038], [0110]);
identify, using the knowledge graph, the one or more entities corresponding to one or more follow up questions to distinguish between the plurality of question response pairs (Ho [0069], [0161], STOILOS [0148]);
receive, via an agent of the plurality of agents, one or more follow up responses to the one or more follow up questions (STOILOS [0116], [0119], [0162]); and
identify, from the plurality of question response pairs based on the one or more follow up responses, the matching question response pair (STOILOS [0116], [0119], [0162]).
Regarding claims 5 and 17, Bell as modified teaches the system and the method, wherein the one or more processors are configured to:
select, from the plurality of agents, a question and answer agent configured to process queries related based on the matched question response pair (Bell [0088], [0109], [0115], [0158]), [0302], [0315]);
determine, by the question and answer agent that the matched question response pair does not satisfy a similarity threshold in the vector space (STOILOS [0094], [0096], [0099], [0145]); and
provide, in response to the determination, using the knowledge graph, a second matching question response pair (STOILOS [0099]-[0100], [0112]).
Regarding claims 6 and 18, Bell as modified teaches the system and the method, wherein the one or more processors are configured to:
select, from the plurality of agents, a guided conversational agent configured to generate one or more follow up questions based on the one or more entities (Bell [0122], [0188], STOILOS [0116], [0119], [0162]-[0163]);
receive, via the guided conversational agent, responsive to the one or more follow up questions, one or more responses define the one or more entities (Bell [0122], [0188], STOILOS [0037], [0081]); and
provide, the response, responsive to the one or more responses (STOILOS [0103], [0156]-[0157], Ho [0039], [0056], [0069]).
Regarding claim 7, Bell as modified teaches the system and the method, wherein the one or more processors are configured to:
select, from the plurality of agents, a system and/or a ground truth”, [0016], [0124], [0133]); generate, using the
Bell does not explicitly teach a system of record (SOR) query agent. Instead, Bell teaches a plurality of various records trained on ground truth. However, a System of Record (SOR) query agent is an AI assistant or software bot designed to securely and intelligently search, retrieve, and summarize data directly from an organization's authoritative "master" databases. It queries the definitive "single source of truth" to prevent employees from using outdated or conflicting information. Thus, an agent, disclosed by Bell trained on the authorities ground truth and knowledge database is obviously analogous to the SOR query agent. Wherein one of ordinary skill in art would benefit from using various query agents per design choice to prevent users from using outdated or conflicting information.
Still, the SOR agent is well-known and disclosed by Mudulodu et al. (US 20250005021) in par.[0022] and Singaraju et al. (US 20230206087) in [0048] and further obviate the teachings of Bell.
Regarding claim 8, Bell as modified teaches the system of claim 7, wherein the structured data comprises one or more of confidential information of associated with the account or confidential information associated with an enterprise corresponding to the account (Bell [0120], [0130], [0217]).
Regarding claim 9, Bell as modified teaches the system of claim 1, wherein the one or more processors are configured to: select, from the plurality of agents, a smart actions agent configured to identify a guided workflow of actions corresponding to the one or more entities (Bell [0122], [0188], STOILOS [0037], [0081], [0116], [0119], [0162]-[0163], Ho [0065]-[0066]); implement one or more actions of the guided workflow to identify information to resolve the one or more entities (Ho [0054], [0060], STOILOS [0035], [0083], [0159], [0162]); and provide the response based on the information (STOILOS [0103], [0156]-[0157], Ho [0039], [0056], [0069]).
Regarding claim 12, Bell as modified teaches the system of claim 1, wherein the one or more processors further reference, in the response provided to the client device, a citation to a section of a document from the set of documentation used to generate the matching question response pair (Bell [0150], [0152], STOILOS [0103], [0156]-[0157], Ho [0039], [0056], [0069]).
Regarding claim 19, Bell as modified teaches the method of claim 13, comprising: selecting, by the one or more processors, from the plurality of agents, a system of record (SOR) query agent configured to access a knowledge graph with function calls to metadata to retrieve structured data corresponding to the one or more entities (Bell [0355] “task-specific agents are trained using one or more abstraction sheets and/or a ground truth”, [0016], [0124], [0133]); generating, by the one or more processors, using the SOR query agent, the response of queries based on the structured data (Bell [0122], [0159]), wherein the structured data comprises one or more of confidential information of associated with the account or confidential information associated with an enterprise corresponding to the account (Bell [0120], [0130], [0217]).
Bell does not explicitly teach a system of record (SOR) query agent. Instead, Bell teaches a plurality of various records trained on ground truth. However, a System of Record (SOR) query agent is an AI assistant or software bot designed to securely and intelligently search, retrieve, and summarize data directly from an organization's authoritative "master" databases. It queries the definitive "single source of truth" to prevent employees from using outdated or conflicting information. Thus, an agent, disclosed by Bell trained on the authorities ground truth and knowledge database is obviously analogous to the SOR query agent. Wherein one of ordinary skill in art would benefit from using various query agents per design choice to prevent users from using outdated or conflicting information.
Still, the SOR agent is well-known and disclosed by Mudulodu et al. (US 20250005021) in par.[0022] and Singaraju et al. (US 20230206087) in [0048] and further obviate the teachings of Bell.
Regarding claim 20, Bell teaches a non-transitory computer readable medium storing program instructions for causing at least one processor to: generate, using one or more machine learning (ML) models and based on a query associated with an account of a processing framework, an embedding corresponding to a vector representation of the query; perform, using the embedding, a vector semantic search in a vector space of a plurality of queries to identify documentation associated with the query and a matching question response pair, the vector space generated using the one or more ML models; identify, from a knowledge graph using the one or more ML models, using the documentation and metadata associated with the account, one or more entities related to the question response pair and one or more relationships between the one or more entities, wherein the knowledge graph defines relationships between entities associated with the queries and responses associated with the queries; select, from a plurality of agents of the processing framework, based on the one or more entities and the one or more relations, an agent to provide an interaction with a client device to address the one or more entities and the one or more relationships for the question response pair; and provide, via the processing framework to the client device, a response of the query based on the question response pair, responsive to the interaction.
Claim 20 recites substantially the same limitations as claim 1, and is rejected for substantially the same reasons.
Claim(s) 10, 12 is/are rejected under 35 U.S.C. 103 as being unpatentable over Bell as modified and in further view of Ganesh et al. (US 12282504).
Regarding claim 10, Bell as modified teaches the system of claim 1, wherein the one or more processors further: determine a plurality of similarity scores between the embedding and a plurality of candidate question embeddings (Bell [0149], [0153], Ho [0069], [0072], STOILOS [0036], [0094]);
Bell does not explicitly teach, however Ganesh discloses calculate a relative similarity threshold based on a distribution of the similarity scores (C19L47-55). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Bell to calculate a relative similarity threshold based on a distribution of the similarity scores as disclosed by Ganesh. Doing so would effectively filter out anomalies and outliers that may be contextually distant, despite a high semantic similarity score (Ganesh C3L13-15).
Regarding claim 12, if Bell as modified does not explicitly teach, however Ganesh discloses the system of claim 1, wherein the one or more processors further reference, in the response provided to the client device, a citation to a section of a document from the set of documentation used to generate the matching question response pair (C15L1-30).
Bell does not explicitly teach, however Ganesh discloses calculate a relative similarity threshold based on a distribution of the similarity scores (C19L47-55). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Bell to provide a citation to a section of a document as disclosed by Ganesh. Doing so would provide accurate, relevant, and contextually appropriate responses to user queries (Ganesh C1L14-15).
Claim(s) 11 is/are additionally or alternatively rejected under 35 U.S.C. 103 as being unpatentable over Bell as modified and in further view of Vishnoi et al. (US 20200342850).
Regarding claim 11, Bell as modified does not explicitly teach however, Vishnoi discloses wherein the one or more processors further:
identify a workflow of one or more queries comprising the query ([0031], [0040], [0050]);
detect a change in query context based on a second query ([0058], [0067]);
identify, in response to detecting the change in the query context, a second agent of the plurality of agents for the second query ([0056], [0089], [0155]-[0159]); and resume, upon completion of processing of the second query by the second agent, the workflow ([0057], [0165]-[0172], [0207]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Bell as modified to identify a second agent for the second query in response to detecting a change in the query context, as disclosed by Vishnoi. Doing so would effectively able to route user input to chatbots in an intelligent manner, so that the user input is sent to the chatbot that the user intends to interact with (Vishnoi [0003]).
Claim(s) 5 and 17 is/are additionally or alternatively rejected under 35 U.S.C. 103 as being unpatentable over Bell as modified and in further view of Du et al. (US 20250337701) and Galitsky (US 20180329880).
Regarding claims 5 and 17, Bell as modified teaches the system and the method as disclosed above However, if Bell does not explicitly teach, Du further discloses select, from the plurality of agents, a question and answer agent configured to process queries related based on the matched question response pair ([0071], [0083], [0086] [0090], [0095], [0156] [0176]); determine, by the question and answer agent that the matched question response pair does not satisfy a similarity threshold in the vector space ([0073], [0088], [0118], [0161]).
Further, if Bell as modified by STOILOS does not explicitly teach, Galitsky further discloses
provide, in response to the determination, using the knowledge graph, a second matching question response pair ([0061] “determines whether the question-answer pair is above a threshold level of matching, e.g., indicating whether the answer addresses the question. If not, the rhetoric classification application continues to analyze additional pairs that include the question and a different answer until a suitable answer is found”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Bell to include question and answer agent configured to process queries related based on the matched question response pair as disclosed by Du and determine second matching question response pair as disclosed by Galitsky. Doing so improve responsiveness and efficiency (Du [0104]) and enables improved automated agents and improved search engine performance over traditional statistical-based approaches that able to determine whether an answer is fully responsive to a question (Galitsky [0043]-[0044]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure is indicated on PTO-892.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to POLINA G PEACH whose telephone number is (571)270-7646. The examiner can normally be reached Monday-Friday, 9:30 - 5:30.
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, Aleksandr Kerzhner can be reached at 571-270-1760. 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.
/POLINA G PEACH/ Primary Examiner, Art Unit 2165 August 9, 2026
Prosecution Insights