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 .
Claims 1-27 are pending for examination.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 14, 16-18 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim language in the following claims is not clearly understood:
As per claim 14, line 3, it is unclear whether “each AI agent instance” is referring to one of the “plurality of AI agent instances” of claim 1 (i.e. consistent term should be used with “the” or “said” if they are the same)
As per claim 16, line 2, it is unclear whether “recommendations” are referring to “optimization recommendations” in claim 3 (i.e. consistent term should be used with “the” or “said” if they are the same)
Line 2-3, it is unclear whether “AI agent instances” is referring to one of the “plurality of AI agent instances” of claim 1 (i.e. consistent term should be used with “the” or “said” if they are the same)
As per claim 17, it is unclear whether “AI agents” is referring to one of the “AI agents” of claim 1 (i.e. consistent term should be used with “the” or “said” if they are the same)
As per claim 18, it is unclear whether “AI agents” is referring to one of the “AI agents” of claim 1 (i.e. consistent term should be used with “the” or “said” if they are the same)
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.
Claim(s) 24 is/are rejected under 35 U.S.C. 103 as being unpatentable over Duford US Pub 2022/0391729 (hereafter Duford) in view of Sohum et al. US Pub 2021/0117893 (hereafter Sohum).
Reference Duford and Sohum were cited in the previous office action.
As per claim 24, As per claim 1, Duford teaches the invention substantially as claimed including a computer implemented method for centralized management and control of AI agents, comprising: monitoring, by one or more processors, a plurality of AI agent instances deployed across at least a selected portion of the platform (para[0149-0153, 0164-0173, 0196-0197, 0206], FIG. 28 and 32, operating a first AI agent and a second AI agent, and the operation performances of the first AI agent and second AI agent are monitored to generate first indication of operation performances of the AI agents);
presenting, by the one or more processors, a management and control user interface for displaying information about the monitored plurality of Al agents (para[0151-0153, 0164, 0171-0172], FIG. 28 and 32, dashboard displays performance of the first and second AI agents based on the monitored and gathered performance information);
providing, by the one or more processors, control functionality through the management and control user interface for performing actions on the monitored plurality of AI agent instances (para[0153-0157, 0164-0169, 0196-0198, 0207], FIG. 28-30, generate and display indications relating to operation performance of the first and second AI agents, and provide user interface to receive user input to modify/update the AI agents (configuration) related to the workflow).
Duford does not explicitly teach AI agents within a SaaS platform.
However, Sohum teaches AI agents within a SaaS platform (para[0103, 0111], AI agents
(chatbot) are deployed within SaaS).
It would have been obvious to one of ordinary skill in the art before the effective filing
date of the claimed invention to incorporate Sohum’s teaching to Duford’s invention in order to provide an online agent marketplace which enables selection of a suitable intelligent agent from a plurality of agents based on the ranking and performance, and the marketplace categorizes data compliance of each of the one or more intelligent agents along with keeping security precautions in place (para[0065, 0109-0110]).
Claim(s) 1, 4, 8, 11-14, 19-23, 25-27 is/are rejected under 35 U.S.C. 103 as being unpatentable over Duford US Pub 2022/0391729 (hereafter Duford) in view of Sohum et al. US Pub 2021/0117893 (hereafter Sohum) and further in view of Yamane et al. US Pub 10,956,833 (hereafter Yamane).
Reference Duford and Sohum were cited in the previous office action.
As per claim 1, Duford teaches the invention substantially as claimed including a system for centralized management and control of AI agents, comprising: one or more processors configured to: monitor a plurality of AI agent instances deployed across at least a selected portion of the platform (para[0149-0153, 0164-0173, 0196-0197, 0206], FIG. 28 and 32, operating a first AI agent and a second AI agent, and the operation performances of the first AI agent and second AI agent are monitored to generate first indication of operation performances of the AI agents);
wherein each AI agent instance of the plurality of AI agent instances comprises one or more executable skills that define discrete functional capabilities of the AI agent instance (para[0072, 0134-0135, 0150], each IA agent provides different processing functionalities);
store information about the monitored plurality of Al agent instances in a data repository; present a management and control user interface for displaying the stored information about the monitored plurality of Al agent instances (para[0151-0153, 0164, 0171-0172], FIG. 28 and 32, dashboard displays performance of the first and second AI agents based on the monitored and gathered performance information);
provide control functionality through the management and control user interface for performing actions on the monitored plurality of AI agent instances (para[0153-0157, 0164-0169, 0196-0198, 0207], FIG. 28-30, generate and display indications relating to operation performance of the first and second AI agents, and provide user interface to receive user input to modify/update the AI agents (configuration) related to the workflow).
Duford does not explicitly teach AI agents within a SaaS platform; each executable skill of the one or more executable skills comprising a set of instructions defining a task and being associated with a manifest comprising a structured metadata record that declares one or more permissions of the executable skill; gate access by each executable skill of the one or more executable skills to at least one of agent configuration data of the AI agent instance, credential data, or another executable skill of the one or more executable skills, such that the access is permitted only when the access is declared in the manifest associated with that executable skill and approved by the one or more processors according to a policy.
However, Sohum teaches AI agents within a SaaS platform (para[0103, 0111], AI agents (chatbot) are deployed within SaaS).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Sohum’s teaching to Duford’s invention in order to provide an online agent marketplace which enables selection of a suitable intelligent agent from a plurality of agents based on the ranking and performance, and the marketplace categorizes data compliance of each of the one or more intelligent agents along with keeping security precautions in place (para[0065, 0109-0110]).
Sohum and Duford do not explicitly teach each executable skill of the one or more executable skills comprising a set of instructions defining a task and being associated with a manifest comprising a structured metadata record that declares one or more permissions of the executable skill; gate access by each executable skill of the one or more executable skills to at least one of agent configuration data of the AI agent instance, credential data, or another executable skill of the one or more executable skills, such that the access is permitted only when the access is declared in the manifest associated with that executable skill and approved by the one or more processors according to a policy.
However, Yamane teaches each executable skill of the one or more executable skills comprising a set of instructions defining a task and being associated with a manifest comprising a structured metadata record that declares one or more permissions of the executable skill (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 15, line 55-67, col 16, line 1-21, a first task with ‘instructions to perform specific task’ input represent the “executable skills” comprising a set of instructions defining how a task should be executed using each AI agent and the first version of the online software product, where the login credentials are provided to grant the AI agent to access the online software product, thus the login credentials associated with the task represents the metadata record that declares permissions of the accessing the online software product by the executing the task);
gate access by each executable skill of the one or more executable skills to at least one of agent configuration data of the AI agent instance, credential data, or another executable skill of the one or more executable skills, such that the access is permitted only when the access is declared in the manifest associated with that executable skill and approved by the one or more processors according to a policy (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 15, line 55-67, col 16, line 1-21, based on the received login credentials, the first task is permitted to login using the credential and use the online software product to complete the performing the first task by the AI agent, where the credential includes username and password, an access code, biometrics, and the instructions indicates to the AI agent to logging into an application using a certain method credentials).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Yamane’s teaching to Sohum and Duford’s invention to provide a method of using AI agent to mimic the behaviors and responses of users to new products, design and processes, and to application of machine learning to train simulated users for evaluation and development of more effective, usable and adoptable software application, using demographic characteristics and its performance of the application (col 1, line 15-21).
As per claim 4, Sohum, Duford and Yamane teaches wherein the control functionality enables activation and deactivation of specific agents (para[0021, 0205], activate the selected AI agents);
modification of agent operational parameters and permissions, and resource allocation adjustments and usage limit configurations (para[0023, 0196-0198, 0207], configure/modify the parameters of the AI agents, where the agents can be updated and/or replaced).
As per claim 8, Duford teaches wherein the control functionality further enables bulk management operations for multiple agents simultaneously (para[0197-0207], FIG. 36, dashboard receive inputs from the user to manage the plurality of agents).
As per claim 11, Duford teaches wherein the control functionality comprises skill deployment management enabling simultaneous deployment of one or more of the executable skills to multiple compatible AI agent instances of the plurality of Al agent instances (para[0150, 0205], the selection of the AI solution provides a single point of entry to centrally view, monitor and manage all AI solutions deployments, and the first and second AI agents are activated simultaneously).
As per claim 12, Duford teaches further comprising version control management functionality, wherein the one or more processors store skill version history for each Al agent (para[0192], execution of the command by AI agents are tracked and monitored).
As per claim 13, Duford teaches wherein the management and control user interface enables selective rollback one or more of the executive skills to previous versions for an individual AI agent instance of the plurality of the AI agent instances independently of other AI agent instances of the plurality of AI agent instances (para[0207], an updated version of the first AI agent has the same configuration as the previous version of the first AI agent).
As per claim 14, Sohum teaches wherein the control functionality comprises deployment isolation management that monitors skill deployments in secure sandbox environments for each AI agent instance, a secure sandbox environment of the secure sandbox environments isolating an executable skill of the one or more executable skills from core logic of the AI agent instance (para[0073, 0104], automated testing determines performance of the agents with different inputs within a sandbox environment).
In addition, Yamane teaches preventing the executable skill from accessing the agent configuration data absent the access being declared in the manifest associated with that executable skill and approved by the one or more processors according to the policy (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 15, line 55-67, col 16, line 1-21, based on the received login credentials, the first task is permitted to login using the credential and use the online software product to complete the performing the first task by the AI agent).
As per claim 19, Duford teaches further comprising enterprise orchestration and compliance capabilities, wherein the management and control user interface implements global orchestration layers for enterprise-wide coordination and policy enforcement (para[0141, 0184, 0191], workload optimizer allows configuration and orchestration of the runtime environment for the setup, performance and monitoring of applications, and the command orchestrator manage the execution of the commands by the AI agents).
As per claim 20, Duford teaches wherein the enterprise orchestration capabilities present orchestration and integration data as quantifiable business metrics including one or more of workflow efficiency measurements, cost reduction analytics, and operational performance indicators (para[0141, 0184, 0191, 0193], the command orchestrator manage the execution of the commands by the AI agents, and the dashboard displays performance indicators of the AI agents).
As per claim 21, Duford teaches further comprising knowledge base module management, wherein the management and control user interface displays organizational knowledge base modules and user-specific knowledge base modules with semantic linkage mapping between related concepts (para[0149-0152, 0201-0202, 0206], monitoring interface provides a user with an overview of AI model performance metrics alongside business level KPIs, and dashboard receives inputs from the user, and when data source fails to match input type of the first AI agent, a warning is displayed to the user).
As per claim 22, Duford teaches wherein the management and control user interface enables configuration of hierarchical access priorities between conflicting information sources, monitoring of data lineage tracking, and management of cross-module access rules (para[0149-0152, 0201-0202, 0206], monitoring interface provides a user with an overview of AI model performance metrics alongside business level KPIs, and dashboard receives inputs from the user to modify configurations of the AI agents, and when data source fails to match input type of the first AI agent, a warning is displayed to the user).
As per claim 23, Sohum teaches further comprising dedicated database instance management, wherein the management and control user interface enables instantiation of dedicated database instances for each user account with both organizational and user-specific knowledge base content while maintaining logical separation between data domains (para[0063, 0086, 0088], FIG. 4, profile, behavioral, roles, authentication information regarding each user is stored as a first set of data dedicated for the user, and the logical separation is maintained for the different user’s information).
As per claim 24, it is a computer implemented method claim of claim 1 above, thus it is rejected for the same rationale.
As per claim 25, Duford, Sohum and Yamane teach the system of claim 1, Yamane teaches wherein the one or more processors maintain credential data of the AI agent instance separately from the one or more executable skills (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 15, line 55-67, col 16, line 1-21, login credentials associated with the AI agent is inputted, and the first task is permitted to login using the credential and use the online software product to complete the performing the first task);
and mediate access by an executable skill of the one or more executable skills to a resource requiring the credential data in accordance with the manifest associated with that executable skill and the policy (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 13, line 58-67, col 14, line 1-10, col 15, line 55-67, col 16, line 1-21, perform specific task is given to the AI agent using the given “instruction to perform specific task”, which provides instructions whether to use the credential data or some other method to login as needed for to perform the task).
As per claim 26, Yamane teaches wherein the structured metadata record of the manifest comprises a skill identifier, a skill type, an input and output schema, the one or more permissions, and one or more execution constraints (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 13, line 58-67, col 14, line 1-10, col 15, line 55-67, col 16, line 1-21, first task is related to a function that is performed using online software application, such as depositing checks, making payment, logging on etc., where the inputs include instruction to perform the specific task, user account data, AI training algorithm, login credential and the outputs include detected obstacles, AI session report).
As per claim 27, Yamane teaches wherein the set of instructions of each executable skill of the one or more executable skills comprises one or more of a prompt-based instruction set, executable code, or an integration interface to an external service (col 1, line 48-67, col 8, line 3-29, col 9, line 10-45, col 13, line 58-67, col 14, line 1-10, col 15, line 55-67, col 16, line 1-21, first task is related to a function that is performed using online software application, which AI agents is executing the executable code according to the instruction to perform specific task).
Claim(s) 2-3, 5-7, 9-10, 15-18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Duford in view of Sohum and Yamane as applied to claim 1 above, and further in view of Jain et al. US Patent 12,106,205 (hereafter Jain).
As per claim 2, Duford, Sohum and Yamane teach the system of claim 1, but they do not explicitly teach wherein the one or more processors are configured to track computational resource data associated with the plurality of AI agent instances including computational tokens consumed by each AI agent instance of the plurality of AI agent instances over specified time periods and expected token usage based on AI agent instance workload and historical patterns.
However, Jain teaches the one or more processors are configured to track computational resource data associated with the plurality of AI agent instances including computational tokens consumed by each AI agent instance of the plurality of AI agent instances over specified time periods and expected token usage based on AI agent instance workload and historical patterns (col 33, line 49-67, col 34, line 1-19, predicts a number of output tokens (expected token usage) and estimate of resource usage based on the historical data).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Jain’s teaching to Duford, Sohum and Yamane’s invention in order to provide a platform which enables selection of particular machine learning models on the basis of a predicted resource allocation requirement and estimated performance metric values, thereby optimizing the performance metric values and improving the efficiency of the system (abstract).
As per claim 3, Duford, Sohum and Yamane teach the system of claim 1, and Duford teaches wherein the management and control user interface displays agent real-time agent status and activity summaries (para[0075, 0151-0153, 0164, 0171-0172], FIG. 28 and 32, dashboard displays performance of the first and second AI agents based on the monitored and gathered performance information).
Duford, Sohum and Yamane do not explicitly teach utilization metrics and resource consumption data, and token usage analytics with cost projections and optimization recommendations.
However, Jain teaches utilization metrics and resource consumption data, and token usage analytics with cost projections and optimization recommendations (col 31, line 9-50, col 33, line 49-67, col 34, line 1-19, col 36, line 25-49, performance metric including CPU utilization, cost, memory usage, and predicts a number of output tokens (expected token usage)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Jain’s teaching to Duford, Sohum and Yamane’s invention in order to provide a platform which enables selection of particular machine learning models on the basis of a predicted resource allocation requirement and estimated performance metric values, thereby optimizing the performance metric values and improving the efficiency of the system (abstract).
As per claim 5, Duford, Sohum, Yamane and Jain teaches the system of claim 3, and Duford teaches wherein the agent utilization metrics comprise task completion rates (para[0206], dashboard displays rate at which the AI agents are processing input).
In addition, Sohum teaches error frequency indicators, and user satisfaction scores (para[0073, 0077, 0106], determines level of trust in the trust score of intelligent agents, which is determined based on vendor reputation, likelihood of satisfaction, detecting agents with unexpected results, and confidence score that indicates likeliness of switch between agents).
In addition, Jain teaches response time measurements (col 17, line 60-67, col 18, line 1-13, time associated with the processing the request).
As per claim 6. Duford, Sohum and Yamane teach the system of claim 4, and Duford teaches wherein the agent operational parameters comprise behavioral constraints, access permissions, and capability activation settings (para[0192-0198, 0205-0207], configure parameters of the agents to be updated and/or replaced, activating AI agents, re-executing commands).
Duford, Sohum and Yamane do not explicitly teach memory allocation limits.
However, Jain teaches memory allocation limits (col 17 , line 43-59, determine a maximum memory size based on the system state).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Jain’s teaching to Duford, Sohum and Yamane’s invention in order to provide a platform which enables selection of particular machine learning models on the basis of a predicted resource allocation requirement and estimated performance metric values, thereby optimizing the performance metric values and improving the efficiency of the system (abstract).
As per claim 7, Duford, Sohum, Yamane and Jain teach the system of claim 3, and Duford teaches wherein the management and control user interface further displays (para[0206], dashboard displays information related to the AI agents).
In addition, Jain teaches predictive analytics for future resource requirements based on usage trends (col 31, line 9-50, col 33, line 49-67, col 34, line 1-19, col 36, line 25-49, estimated resource requirements and predicts a number of output tokens (expected token usage)).
As per claim 9, Duford, Sohum, Yamane and Jain teach the system of claim 3, Duford teaches further comprising agent marketplace display functionality, wherein the management and control user interface displays Al agent instances of the plurality of AI agent instances with visual representations showing the functional capabilities and specialization areas of the displayed AI agent instances (para[0197-0207], FIG. 36, dashboard receive inputs from the user to manage the plurality of agents, where the information related to the agents are displayed).
As per claim 10, Duford teaches wherein performance analytics including task completion rates and user satisfaction scores are integrated into the visual representations of the Al agents (para[0148, 0197-0207], FIG. 36, dashboard receive inputs from the user to manage the plurality of agents, and the AI agent outputs confidence score, where a rate at which the AI agents are processing and the information related to the agents are displayed).
As per claim 15, Sohum teaches further comprising marketplace interaction functionality, wherein the management and control user interface presents the executable skills and the functional capabilities of the Al agent instances as quantifiable workflow improvements including task automation capabilities, process efficiency metrics, and resource optimization measurements (para[0026, 0065, 0067], AI agent (chatbot) marketplace includes different chatbots with different functionalities, capabilities, metrics, and optimal chatbot is selected).
As per claim 16, Sohum teaches wherein the marketplace interaction functionality comprises automatically generated recommendations of AI agent instances based on user role and workflow patterns (para[0089-0090], recommendation module includes role based, interest based, behavioral based, profiled based).
As per claim 17, Duford, Sohum and Yamane teach the system of claim 1, but they do not explicitly teach further comprising architecture-agnostic credential management, wherein the management and control user interface provides centralized credential management accessible by Al agents regardless of deployment architecture.
However, Jain teaches architecture-agnostic credential management, wherein the management and control user interface provides centralized credential management accessible by Al agents regardless of deployment architecture (col 8, line 30-53, col 12, line 62-67, the platform includes access control engine controlling credential management)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Jain’s teaching to Duford. Sohum and Yamane’s invention in order to provide a platform which enables selection of particular machine learning models on the basis of a predicted resource allocation requirement and estimated performance metric values, thereby optimizing the performance metric values and improving the efficiency of the system (abstract).
As per claim 18, Duford, Sohum, Yamane and Jain teach the system of claim 17, and Jain teaches wherein the architecture-agnostic credential management supports Al agents deployed as microservices, embedded components, and API-based integrations (col 8, line 30-53, col 12, line 62-67, col 24, line 47-61, the platform includes access control engine controlling credential management, and providing the prompt request to LLM through an API call).
Response to Arguments
Applicant’s arguments with respect to claim(s) 1-27 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Conclusion
THIS ACTION IS MADE FINAL. 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 TAMMY EUNHYE LEE whose telephone number is (571)270-7773. The examiner can normally be reached Mon, Tues, Thur 9PM-4PM.
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, Aimee Li can be reached at (571)272-4169. 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.
/TAMMY E LEE/Primary Examiner, Art Unit 2195
Prosecution Insights