DETAILED ACTION
The following is a FINAL office action upon examination of the application number 19/065589.
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 7/13/2026 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Response to Amendment
Claims 1, 4, 6, 8, 13, and 15 have been amended.
Claims 5, 12, and 19 have been canceled.
Claims 21-23 are new.
Claims 1-4, 6-11, 13-18, and 20-23 are pending in the application and have been examined on the merits discussed below.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-4, 6-11, 13-18, and 20-23 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
(Step 1) Claims 1-4, 6, 7, and 21-23 are directed to a system comprising one or more processing devices; therefore, these claims are directed to a machine which is a statutory category of invention. Claims 8-11 and 13-14 are directed to a non-transitory computer-readable medium, which is a manufacture, and this a statutory category of invention. Claims 15-18 and 20 are directed to a method; thus these claims are directed to a process, which is one of the statutory categories of invention.
(Step 2A) The claims recite an abstract idea instructing how to orchestrate completion of an adaptive workflow, which is described by claim limitations reciting:
at least one data storage configured to store historical user inputs and historical generated workflows; and
… analyze the historical user inputs and the historical generated workflows … to estimate a context of a current user input;
… generate one or more adaptive workflows in response to the current user input based on the context estimated … such that the one or more adaptive workflows are adjusted using feedback from one or more execution results and are visualized …;
… map one or more computing tasks to the worker …;
… provide one or more status updates of task execution …; and … recommend one or more next steps to one or more users based on one or more results from the one or more adaptive workflows,
wherein the worker … provide one or more real-time status updates of workflow execution … such that the one or more real-time status updates are compared to one or more expected outputs generated … to determine if the one or more adaptive workflows are proceeding without error.
The identified limitations in the claims describing orchestrating completion of an adaptive workflow (i.e., the abstract idea) fall within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas, which covers fundamental economic practices or, alternatively, the “Mental Processes” grouping of abstract ideas since the identified limitations can be performed by a human, mentally or with pen and paper. Dependent claims 2-4, 7, 9-11, 13-14, 16-18, and 20-21 recite limitations that further narrow/describe the abstract idea; therefore, these claims are also found to recite an abstract idea.
This judicial exception is not integrated into a practical application because additional elements such as the one or more processing devices configured to execute a context estimation agent (CEA), a workflow composer agent (WCA), a workflow supervisor agent (WSA), a worker agent, and a recommendation agent (RA); and visualization agent (VA) in claim 1; the non-transitory machine-readable medium including program code; context estimation agent (CEA), a workflow composer agent (WCA), a workflow supervisor agent (WSA), recommendation agent (RA); and visualization agent (VA) in claim 8; and the context estimation agent (CEA), a workflow composer agent (WCA), a workflow supervisor agent (WSA), recommendation agent (RA); and visualization agent (VA) in claim 15, do not add a meaningful limitation to the abstract idea since these elements are only broadly applied to the abstract ideas at a high level of generality; thus, none of recited hardware offers a meaningful limitation beyond generally linking the abstract idea to a particular technological environment, in this case, implementation via a processor/computer.
Additional elements such as using retrieval-augmented generation do not yield an improvement in the functioning of the computer itself, nor do they yield improvements to a technical field or technology; further, these additional elements are recited at a high level of generality and only generally link the abstract idea to a technical environment. Similarly, additional elements in claims 22 and 23 related to using the LLM and a RAG database do not provide an improvement and only generally link the abstract idea to a technological environment. Accordingly, these additional element do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
(Step 2B) The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because as discussed above with respect to integration of the abstract idea into a practical application, the hardware additional elements amount to no more than mere instructions to apply the exception using a generic computer component (see Spec. [0006]). Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. Additional elements such as using retrieval-augmented generation do not yield an improvement and only generally link the abstract idea to a technical environment. Additional elements in claims 22 and 23 related to using the LLM and a RAG database do not provide an improvement and only generally link the abstract idea to a technological environment. In addition, when taken as an ordered combination, the ordered combination adds nothing that is not already present as when the elements are taken individually. There is no indication that the combination of elements improves the functioning of a computer or improves any other technology.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1-4, 6-11, 13-18, and 20 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by US 2025/0371449 (Thompson).
As per claim 1, Thompson teaches: a system comprising: at least one data storage configured to store historical user inputs and historical generated workflows; and ([0040] … The context data can include user-, entity- and/or environment-specific preferences pertaining to the agents' roles, capabilities, and/or control (e.g., the agents' level or degree of autonomy). The context data may be derived from data logged or otherwise obtained as a result of users' historical use of agents and/or other software applications such as search engines, social networks, and/or domain applications. Different versions of context data may be stored in different memory levels, and different memory layers can be integrated with different portions of different workflows, such that, for example, examples may customize the selections of context data from different memory levels for different tasks or even for subsequent iterations of the same task. [0105] The data resources can include, alternatively or additionally, entity profile data (e.g., user profiles data, company profiles, job postings, etc.), activity data (e.g., historical interaction data such as search histories, chat histories, and/or interaction histories associated with the user's use [0159] … Workflows include data stored in the agent's memory system [0245] … the first context model can include user preferences extracted from an online profile and/or interaction log of the user at or prior to the operation 710. The memory can be a multi-layer memory that includes one or more memory layers, such as working memory, episodic memory, semantic memory, procedural memory, or collective memory. [0252] … The memory can store the second workflow, second micro-prompts, and second context model in one or more memory layers, such as working memory, episodic memory, semantic memory, procedural memory, or collective memory).
one or more processing devices configured to execute a context estimation agent (CEA), a workflow composer agent (WCA), a workflow supervisor agent (WSA), a worker agent, and a recommendation agent (RA), wherein: ([0072] … the distributed multi-agent system 105 includes a plurality of sub-agents 106A, 106B, . . . , 106N, a communication service 108, an adaptive machine learning service 110, and a multi-layer memory structure 111. Any reference to N herein can refer to an Nth element of a device, component, system, or process, where N is a positive integer and the value of N can vary depending on the context. For example, in FIG. 1, the computing system 100 can include N applications 104, N sub-agents 106, N memory layers, N context models, N artificial intelligence services 114, N data resources 116, and N tools 118, where the value of N can be the same or different in each case. [0073] The sub-agents 106A, 106B, . . . , 106N cooperate and coordinate with each other to perform tasks on behalf of the user. Each sub-agent 106A, 106B, . . . , 106N can have a specific role or function; Examiner notes that the ‘names’ of the claimed agents do not have provide any functional limitations to the claims. The ‘agents’ disclosed by Thompson are found to satisfy this limitation)
the CEA is configured to analyze the historical user inputs and the historical generated workflows using retrieval-augmented generation to estimate a context of a current user input; ([0081] … obtain the inputs required to invoke an agent, e.g., to obtain relevant context, interaction history, learned preferences, etc., to parameterize an action (e.g., a specific task of a workflow performed by an agent). The parameterization of actions enables the actions to be configured and customized dynamically using the most current relevant information. For example, when an action is invoked, a workflow is executed that obtains the relevant context, historical data, and learned preferences from memory and parameterizes the action with that information. Other examples of generalized workflows include workflows for performing adaptive machine learning processes to build context models (e.g., models of users and environments), updating semantic memory, and translating interaction experiences into procedural memory. [0130] … can obtain a context model via one or more memory layers of the multi-layer memory 218 and determine an objective for the automated agent based on the event and the context model. For example, at block 304, the orchestrator 216 can load the context model from a working memory layer of the multi-layer memory 218. The context model can represent the current situation, environment, and/or knowledge of the automated agent and can include various types of information, such as user preferences, user history, user goals, data sources, AI services, workflows, plans, actions, results, feedback, or any other information relevant to the task or the goal of the automated agent. [0245] … The first context model can be a representation of the relevant information, data, or knowledge that can be used by the automated agent to perform tasks or actions. For example, the first context model can include user preferences extracted from an online profile and/or interaction log of the user at or prior to the operation 710. [0296] … model 906 can be created or trained using historical data, e.g., interaction data logged during use of the automated agent, pertaining to system actions and associated user responses)
the WCA is configured to generate one or more adaptive workflows in response to the current user input based on the context estimated by the CEA such that the one or more adaptive workflows are adjusted using feedback from one or more execution results and are visualized using a visualization agent (VA); ([0131] … The workflow can be predefined, customized, or generated by the adaptive machine learning-based orchestrator 216 based on the event, the context model [0132] At block 308, the adaptive machine learning-based orchestrator 216 can invoke a planner to use the workflow to generate or update a plan to accomplish the objective. The planner can be an AI service 230, such as an LLM, which can create, modify, or optimize a plan based on the workflow, the event, the context model, and/or feedback from the user [0248] …measuring an actual user response to the output of the first execution of the automated agent, comparing the actual user response to the predicted user response, and updating one or more of the first workflow, the first micro-prompts, and the first context model based on the comparison [0250] The second workflow can be a modified version of the first workflow [0374] …dialog element 1312 enables workflows to be dynamically altered or customized for individual users in an intuitive and efficient way. [0384] … the automated agent presents the updated workflow or plan 1338 including the tasks added or modified by the user and incorporating the user's feedback [0398] … the automated agent presents a summary of the plan created [0400] … indicates the status of the sub-tasks, e.g., check marks indicate that sub-tasks have been completed.)
the WSA is configured to map one or more computing tasks to the worker agent; ([0073] The sub-agents 106A, 106B, . . . , 106N cooperate and coordinate with each other to perform tasks on behalf of the user [0108] To generate a response to the input, the automated agent 102 can invoke one or more sub-agents of the distributed multi-agent system 105. For example, a planner sub-agent of the distributed multi-agent system 105 can be invoked to generate a plan for responding to the input using as input a workflow, a profile, and a context model obtained from one or more of the memory layers. The plan can include a plurality of actions that need to be performed (e.g., in sequence or in parallel), where each action has an associated action sub-agent. [0133] …execution of one or more actions from the plan via one or more agents [0234] The one or more action agents 510, 512, 514 can be agents that can perform one or more actions as part of the plans generated).
the worker agent is configured to provide one or more status updates of task execution to the CEA; and ([0133] …can monitor the status of the plan generated or updated at block 308 [0138] … The plan status can indicate the progress, completion, failure, or interruption of the plan or any of its steps, tasks, actions, or functions. [0386] …automated agent asks the user whether the user would like the agent to send progress updates to the user as the plan is executed. [0400] … indicates the status of the sub-tasks, e.g., check marks indicate that sub-tasks have been completed)
the RA is configured to recommend one or more next steps to one or more users based on one or more results from the one or more adaptive workflows ([0218] …Decide Next Step: Once a plan or course of action is established, the agent can determine its subsequent action [0220] …the agent identifies the need for information or action from other skills or AI agents, the agent can formulate a corresponding request. [0222] Publish Message: Depending on the decision in the previous step, the agent can push out a message into the system. The message can be a feedback request to humans [0287] …which incorporates feedback and makes improvements. The result is then sent back to the originator (e.g., human or agent) for review. …The reviewer can instruct the agent to proceed (approve), cancel (terminate), or implement feedback recommendations. [Fig 12A]; 1210 shows Agent recommending inclusion of additional skills)
wherein the worker agent is configured to provide one or more real-time status updates of workflow execution to the CEA such that the one or more real-time status updates are compared to one or more expected outputs generated by the WCA to determine if the one or more adaptive workflows are proceeding without error ([0098] … ground-truth examples of desired model output are paired with respective inputs, and these input-example output pairs are used to train or fine tune one or more models. [0132] … create, modify, or optimize a plan based on … feedback from the user or another agent…. The plan can also include various information, such as inputs, outputs, resources, agents, tools, or results associated with each step, task, action [0133] … monitor the status of the plan generated or updated at block 308, and initiate or cause execution of one or more actions [0134] … perform or delegate one or more actions and/or to monitor the performance of actions and/or the plan as a whole. [0138] … determine whether the action has been achieved based on the event, the context model, the plan, and/or the result of the action. In some embodiments, at block 316, an adaptive machine learning-based process is used to compare the result of executing the action to a predicted expected result and analyze the difference between the predicted expected result and the actual result… . The plan status can indicate the progress, completion, failure, or interruption [0160] … the agent can formulate a plan with well-defined goals and concrete actions required to achieve them. These plans consider: [0161] Preconditions: Conditions that must be met before an action is taken. [0162] Effects: The expected state after an action is performed. [0386]).
As per claim 2, Thompson teaches: wherein the at least one data storage is configured to categorize the historical user inputs and the historical generated workflows based on domain classification for efficient retrieval by the CEA ([0101] … “classify the user input [input1] into action_a, action_b, or action_c,” where [input1] is a placeholder for the user input and/or associated context data and action_a, action_b, and action_c are possible intents into which the large language model may classify input1. [0107] … The input can be associated with context data, such as a timestamp, a geographic location, a device, network, or session identifier, a topic, a goal, or any other information that can affect the interpretation of the input by the automated agent 102 or the response generated by the automated agent 102 to the input. [0467] … uses the at least one input to determine an objective. For example, the processing device can use one or more machine learning models to extract an objective from the at least one input or classify the at least one input as corresponding to a particular objective. The objective can be a goal, intent, or task to be achieved via the use of an automated agent. For instance, a binary classification machine learning model can be used to classify the at least one input as an objective, goal, intent, or task. [0468] … a first plan comprising one or more tasks to achieve the objective [0624] … historical interactions between the entity and the automated agent [0086] … historical user activity data [0087] … historical data logged during the user's prior uses [0105] … activity data (e.g., historical interaction data such as search histories, chat histories, and/or interaction histories associated with the user's use of applications 104) [0130] … user history, user goals, data sources, AI services, workflows, plans, actions, results [0296] … historical data, e.g., interaction data logged during use of the automated agent, pertaining to system actions and associated user responses [0441] … historical data relating to the creation and execution of workflow)
As per claim 3, Thompson teaches: wherein the CEA is configured to identify one or more patterns in the historical user inputs and the historical generated workflows to estimate the context of the current user input ([0040] … The context data can include user-, entity- and/or environment-specific preferences pertaining to the agents' roles, capabilities, and/or control (e.g., the agents' level or degree of autonomy). The context data may be derived from data logged or otherwise obtained as a result of users' historical use of agents [0085] … store context models 113A, 113B, . . . , 113N that are related to the current dialog or task, the episodic memory 112B can store context models 113A, 113B, . . . , 113N that are related to previous dialogs or tasks, and the collective memory 112C can store context models 113A, 113B, . . . , 113N that are related to general or domain knowledge. [0245] … context model can include user preferences extracted from an online profile and/or interaction log of the user at or prior to the operation 710.)
As per claim 4, Thompson teaches: wherein the CEA is configured to use the RA and the VA to update the one or more adaptive workflows dynamically based on user feedback and one or more execution outcomes ([0046] … The computer system may obtain feedback f1 related to the response r1 at a time t2 which is greater than or equal to the time t1. At a time t3, the first workflow w1 is modified or updated by the computer system based on the feedback f1. [0251] Any one or more of the workflow, micro-prompts and context models may be dynamically updated in response to the adaptive machine learning process [0385] … the user has provided feedback 1344 in a conversational natural language format. In response to the feedback 1344, the automated agent can use the feedback 1344 to update one or more workflows, plans, context models, etc.)
As per claim 6, Thompson teaches: wherein the worker agent is configured to communicate a presence of any error to the CEA and the RA ([0138] …The plan status can indicate the progress, completion, failure, or interruption [0200] error: An optional field that contains error information if the message is an error message. [0315] … The self-correction function can evaluate the output of agent executions and, if the output of the agent executions does not exceed threshold criteria, return to the context preparation function. [0316] … an error handling function, and a response formulation function. The validation function can validate, evaluate, or assess the results or the outputs of the task or the sub-tasks, e.g., the output of the execution process 1008 [0317] … The error handling function can detect, identify, or classify the errors, the exceptions, or the failures and resolve, recover, or mitigate the errors, the exceptions, or the failures).
As per claim 7, Thompson teaches: wherein the CEA is configured to receive one or more instructions from one or more users to modify the one or more adaptive workflows in real-time and provide one or more instructions for a workflow modification to the WCA for continuous adaptation ([0046] … The computer system may obtain feedback f1 related to the response r1 at a time t2 which is greater than or equal to the time t1. At a time t3, the first workflow w1 is modified or updated by the computer system based on the feedback f1 [0378] … the automated agent presents the updated workflow or plan 1326 including the tasks added or modified by the user and incorporating the user's feedback regarding the order of performance of the steps, etc. [0384] …FIG. 13A3, the automated agent presents the updated workflow or plan 1338 including the tasks added or modified by the user and incorporating the user's feedback regarding the order of performance of the steps, etc. The automated agent prompts the user to add steps to and/or remove steps from the plan or workflow 1338).
As per claims 8 and 15, these claims recite limitations substantially similar as those addressed by the rejection of claim 1, above; therefore, the same rejection applies.
As per claims 9 and 16, these claims recite limitations substantially similar as those addressed by the rejection of claim 2, above; therefore, the same rejection applies.
As per claims 10 and 17, these claims recite limitations substantially similar as those addressed by the rejection of claim 3, above; therefore, the same rejection applies.
As per claims 11 and 18, these claims recite limitations substantially similar as those addressed by the rejection of claim 4, above; therefore, the same rejection applies.
As per claim 13, this claim recites limitations substantially similar as those addressed by the rejection of claim 6, above; therefore, the same rejection applies.
As per claims 14 and 20, these claims recite limitations substantially similar as those addressed by the rejection of claim 7, above; therefore, the same rejection applies.
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) 21 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 2025/0371449 (Thompson); in view of US 2023/0229534 (Mohanty).
As per claim 21, although not explicitly taught by Thompson, Mohanty teaches: wherein, to estimate the context of the user input, the CEA is further configured to perform cosine matching ([0036] … response to the user input (e.g., the user clicking/tapping the control), API marketplace 108 can send the new API specification along with a request [0060] … Data may be entered using an input device of GUI 414 [0025] … the matching of API specifications is achieved by predicting the context and intent of a new API that is being designed from its API specification. … natural language processing (NLP) and topic modeling algorithms can be utilized to predict the context and intent of the new API from its API specification. The API specifications corresponding to existing APIs can then searched to identify API specifications corresponding to existing APIs whose context and intent are similar to the predicted context and intent of the new API. [0032] … the keywords correspond to or are indicative of the context and intent of the existing API specification rather than simply a list of words found the contents [0033] …identify one or more existing API specifications whose context and intent are similar [0034] … similarity prediction module 106 may use a distance measure, such as cosine similarity).
It would have been obvious, before the effective filing date of the claimed invention, for one of ordinary skill in the art to have modified the teachings of Thompson with the aforementioned teachings of Mohanty with the motivation of predicting context for a request. Further, one of ordinary skill in the art would have recognized that applying the teachings of Mohanty to the system of Thompson would have yielded predictable results and doing so would have been recognized by those of ordinary skill in the art as resulting in an improved system that would allow for use of cosine matching.
Claim(s) 22 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 2025/0371449 (Thompson); in view of AGENTIC RETRIEVAL-AUGMENTED GENERATION: A SURVEY ONAGENTIC RAG (Singh, Jan 2025).
As per claim 22, Thompson teaches: wherein a large language model (LLM) is coupled to one or more of the CEA, the WCA, the WSA, or the worker agent ([0077] … agent 102 or any sub-agent 106A, 106B, . . . , 106N can be implemented as a stateful, LM- and/or LLM-based multi-actor application with built-in persistent memory)
…one or more adaptive workflows accepted by the one or more users ([0378] … The automated agent prompts the user to confirm that the workflow or plan 1326 is acceptable to the user (e.g., meets the user's goals or objectives)).
Although not explicitly taught by Thompson, Singh teaches: filter, using the LLM, the context of the current user input by keeping only content that is material to one of the one or more adaptive workflows (Page 2 …agents, including LLM powered and mobile agents [13] … multi-agent collaboration [16], enabling them to manage dynamic workflows …Agentic RAG employs autonomous agents to orchestrate retrieval, filter relevant information, and refine responses … Processes retrieved data, extracting and summarizing the most relevant information to align with the query context Page 4 prioritize the most contextually relevant information. Page 9 agents can adapt to complex tasks and provide more accurate and contextually relevant outputs. Page 13 Data Integration and LLM Synthesis: Once retrieval is complete, the data from all agents is passed to a Large Language Model (LLM). The LLM synthesizes the retrieved information into a coherent and contextually relevant response Page 25 RAG systems dynamically retrieve the most relevant information, adapt to the user’s context, and generate personalized responses).
It would have been obvious, before the effective filing date of the claimed invention, for one of ordinary skill in the art to have modified the teachings of Thompson with the aforementioned teachings of Singh with the motivation of improving relevance of output provided (Singh Page 2). Further, one of ordinary skill in the art would have recognized that applying the teachings of Singh to the system of Thompson would have yielded predictable results and doing so would have been recognized by those of ordinary skill in the art as resulting in an improved system that would allow for providing the most relevant information.
Claim(s) 23 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 2025/0371449 (Thompson); in view of AGENTIC RETRIEVAL-AUGMENTED GENERATION: A SURVEY ONAGENTIC RAG (Singh, Jan 2025); in view of US 2024/0289863 (Smith).
As per claim 23, although not explicitly taught by Thompson, Singh teaches: … the filtered context of the current user input… (Page 2 …Agentic RAG employs autonomous agents to orchestrate retrieval, filter relevant information, and refine responses … Processes retrieved data, extracting and summarizing the most relevant information to align with the query context).
It would have been obvious, before the effective filing date of the claimed invention, for one of ordinary skill in the art to have modified the teachings of Thompson with the aforementioned teachings of Singh with the motivation of improving relevance of output provided (Singh Page 2). Further, one of ordinary skill in the art would have recognized that applying the teachings of Singh to the system of Thompson would have yielded predictable results and doing so would have been recognized by those of ordinary skill in the art as resulting in an improved system that would allow for providing the most relevant information.
Although not explicitly taught by Thompson, Smith teaches: store the filtered context of the current user input in the at least one data storage as a RAG database ([0045] … the conversation history between the system and the user, or a subset of this history identified as important either by the user or by a machine learning model, or a set of higher level summaries or other abstractions of the user's conversation history, or similar text, audio or visual representations of the user's history with the system, may be embedded as indexed vectors in a vector database or similar storage structure for later reference by the system [0049] … vectors, their corresponding chunk of text, and other metadata such as the originating document, tags, upload date, user data, etc., may be stored in a dedicated vector database [0052] … Retrieval Augmented Generation (RAG) whereby a targeted search across the embedded vector database is performed in response to a user query, e.g., in order to produce context for a conversational agent to then generate a response).
One of ordinary skill in the art would have recognized that applying the teachings of Smith to the system of Thompson would have yielded predictable results and doing so would have been recognized by those of ordinary skill in the art as resulting in an improved system that would allow for the storage of context data for future retrieval.
Response to Arguments
Applicant's arguments filed 6/12/2026 have been fully considered but they are not persuasive.
With respect to the rejection under 35 USC 101, Applicant argues that the claims are directed to an improvement to a technology.
Examiner respectfully disagrees. The use of prior context, historical data and execution results to generate adaptive workflows does not amount to an improvement to the technology. Examiner notes that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology. For example, in Trading Technologies Int’l v. IBG, 921 F.3d 1084, 1093-94, 2019 USPQ2d 138290 (Fed. Cir. 2019), the court determined that the claimed user interface simply provided a trader with more information to facilitate market trades, which improved the business process of market trading but did not improve computers or technology.
With respect to the rejection under 35 USC 101, Applicant argues that the claims do not recite an economic practice or mental process.
Examiner respectfully disagrees. Examiner maintains that the identified limitations in the claims describing orchestrating completion of an adaptive workflow (i.e., the abstract idea) fall within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas, which covers fundamental economic practices or, alternatively, the “Mental Processes” grouping of abstract ideas since the identified limitations can be performed by a human, mentally or with pen and paper.
Additional elements related to different agents only generally link the abstract idea to a technological environment and do not provide an improvement. Further, allowing for correction of a workflow during execution does not improve the performance of the computer or technology; an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology.
With respect to the rejection under 35 USC 101, Applicant argues that the claims provide significantly more.
Examiner respectfully disagrees. The use of a feedback loop in execution of a workflow does not yield a technical improvement, while additional elements related to different agents only generally link the abstract idea to a technological environment and do not provide an improvement.
With respect to the rejection under 35 USC 102, Applicant argues that the art of record does not disclose the claimed features.
Examiner respectfully disagrees. Thus Examiner maintains that Thompson discloses wherein the worker agent is configured to provide one or more real-time status updates of workflow execution to the CEA such that the one or more real-time status updates are compared to one or more expected outputs generated by the WCA to determine if the one or more adaptive workflows are proceeding without error ([0098] … ground-truth examples of desired model output are paired with respective inputs, and these input-example output pairs are used to train or fine tune one or more models. [0132] … create, modify, or optimize a plan based on … feedback from the user or another agent…. The plan can also include various information, such as inputs, outputs, resources, agents, tools, or results associated with each step, task, action [0133] … monitor the status of the plan generated or updated at block 308, and initiate or cause execution of one or more actions [0134] … perform or delegate one or more actions and/or to monitor the performance of actions and/or the plan as a whole. [0138] … determine whether the action has been achieved based on the event, the context model, the plan, and/or the result of the action. In some embodiments, at block 316, an adaptive machine learning-based process is used to compare the result of executing the action to a predicted expected result and analyze the difference between the predicted expected result and the actual result… . The plan status can indicate the progress, completion, failure, or interruption [0160] … the agent can formulate a plan with well-defined goals and concrete actions required to achieve them. These plans consider: [0161] Preconditions: Conditions that must be met before an action is taken. [0162] Effects: The expected state after an action is performed. [0386]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 2025/0378092 – discloses the use of RAG databases storing context information ([0040][0046]).
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALAN TORRICO-LOPEZ whose telephone number is (571)272-3247. The examiner can normally be reached M-F 10AM-5PM.
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, Beth Boswell can be reached at (571)272-6737. 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.
/ALAN TORRICO-LOPEZ/Primary Examiner, Art Unit 3625
Prosecution Insights