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 .
Status of the Claims
Claims 1-7 were previously pending and subject to a non-responsive amendment action mailed 04/07/2026. Claim 1 was amended; no claim was cancelled, or added in a reply filed 06/08/2026. Therefore claims 1-7 are currently pending and subject to the final office action below.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 10/10/2025, 12/11/2025, 02/23/2026 and 07/01/2026 was considered by the examiner.
Response to Arguments
Applicant's arguments filed 01/02/2026 in regard to 101 rejection have been fully considered but they are not persuasive.
Applicant argues “Claim 1, as amended, recites "collecting, by a Real-Time Data Mesh (RTDM) module, heterogeneous data from at least two predefined sources, wherein the heterogeneous data comprises structured and unstructured data representing one or more aspects of system activity, user behavior, or environmental signals, and transforming the collected data into a standardized format for downstream processing, wherein the RTDM applies change data capture (CDC) to detect and propagate data modifications in a data layer, and updates one or more persistent Purposive Data Stores (PDSes) based on the CDC-detected changes…Claim 1 further recites analyzing, by the AAML module, the standardized data to identify predictive patterns using defined neural network models, generating predictive outputs corresponding to the identified patterns, detecting CDC-detected data modifications affecting at least one PDS accessed during execution of the analysis, re-executing the analysis using updated data to maintain consistency between the predictive outputs and the underlying data, monitoring execution results of the re-executed analysis in real time, and refining model weighting parameters based on the monitored execution results for subsequent executions under changing data conditions.” (remarks p. 7-8).
Examiner respectfully disagrees. Applicant’s recitation of the claim does not demonstrate that the claim is directed to an improvement in computer functionality. The lcaim is directed to principally to collecting and standardizing information; storing and updating information; analyzing the information to identify predictive patterns; generating predictive outputs; monitoring changes and execution results; repeating the analysis when the underlying information changes; and modifying model parameters based on feedback. These limitations describe information collection, evaluation, prediction, and iterative refinement. The RTDM, CDC, pDS, and AAML limitations identify the computer tools used to perform those functions, but do not recite a new data mesh protocol, a new CDC detection mechanism, a new database structure, a new neural network architecture, or a new model training technique.
The federal circuit has recently treated materially similar claims involving collecting data, applying machine learning, updating results, and retraining or refining models as directed to an abstract idea when the claims did not specify how the alleged technological improvement was achieved (please see Recentive Analytics, inc v. Fox Corp 134 F.4th 1205).
Applicant argues “he Office Action alleges that "[l]imitation such as 'analyzing the collected data with a machine learning model to identify one or more patterns, wherein the model includes at least one of a convolutional neural network for image data patterns and a recurrent neural network for temporal data patterns' are directed towards mathematical concepts because it describes using a generic machine learning model to perform an analysis." Office Action, 4. Applicant disagrees.” (remarks p. 8).
The argument is not persuasive. Claim 1 recites analyzing standardized data using a CNN or RNN to identify predictive patterns and subsequently “refining model weighting parameters” based on monitored results. Neural network weighting parameters represent numerical relationships between model inputs, intermediate activations, and outputs. Refining those parameters based on model result describes, at the claimed level of generality, mathematical optimization of the model.
The claim does not recite a particular improvement to CNN or RNN operation. It does not specify, for example, a new network topology, activation function, loss function, gradient calculation, weight update rule, memory arrangement, or hardware implementation. The CNN and RNN are instead invoked as tools for carrying out the claimed pattern identification and predictive analysis.
Additionally, the limitations of identifying predictive patterns, generating predictive outputs, monitoring execution results and feedback, and determining how to refine subsequent predictions encompass evaluations and judgements falling within the mental process grouping. The fact that those evaluations are assigned to an AAML module does not remove their abstract character.
This analysis is consistent with USPTO Example 47, which explains that broadly recited neural network training, pattern detection, and analysis may recite mathematical calculations and mental evaluations where the claim does not explain how the analysis is performed (please see UPSTO Eligibility Example 47).
Applicant argues “The claimed systems are not abstract methods of organizing human activity or mathematical concepts performed in the human mind. Rather, the claimed method is directed to a specific technical solution involving a multi-module computing architecture configured to maintain correctness and consistency of predictive analytics execution in a distributed, real-time data environment…These operations depend on machine-level data synchronization, event-driven execution control, and stateful interaction between persistent data stores and executing analytics processes. Such functions cannot be performed with pen and paper, are not mental processes, and do not recite mathematical concepts in the abstract. In particular, CDC-based detection of insert, update, and delete events, real-time propagation of those events across heterogeneous data domains, and automated re-execution of analytics in response to detected data mutations are inherently computer-implemented operations.” (remarks p. 10).
The argument that the claims as a whole cannot be performed mentally does not establish eligibility. Under step 2A, prong one, the question is whether particular claim limitations recite a judicial exception; the entire claim need not be performable in the human mind. Computer specific limitations may be evaluated as additional elements under step 2A, prong Two (please see MPEP 2106.04(a) and 2106.04(d)).
Moreover, the inability to perform CDC or neural network execution manually does not itself establish an improvement to computer technology. MPEP 2106.05(f) explained that the use of computers to perform data operations more quickly, continuously, accurately, or automatically does not establish eligibility when the claim does not recite a specific improvement in how the computer performs those operations and they are simply performing generic functions of the computer.
Applciant’s argument also characterizes the claim as requiring “controlled re-execution”, “stateful interaction”, “distributed architecture” and “automated re-execution”. However, claim 1 does not recite: pausing or checkpointing an executing analysis; maintaining a snapshot or version identifier for the data used in the analysis; determining which portion of an analysis depends on a modified record; invalidating an intermediate analytical result; rolling back an executing process; selectively recomputing an affected portion of the analysis, etc…
The claim requires only detecting a modification “during execution” and in response “re-executing the analysis” using updated data. The phrase “during execution” establishes the timing of the detected event, but does not provide a technical mechanism for coordinating the modification with the executing analysis. Repeating an analysis because its input data changed remains result oriented data processing.
Applicant argues “Under Step 2A, Prong Two of the 2019 Revised Patent Subject Matter Eligibility Guidance, even if the claims were found to recite an abstract idea, they must still be analyzed to determine whether the claims integrate the alleged exception into a practical application. The amended claims do so by employing a specific, technical architecture that solves a problem rooted in distributed computing systems, namely, maintaining reliable and consistent execution of predictive analytics in the presence of asynchronous, real-time data modifications across heterogeneous data sources.” (remarks p. 11)
Examiner agrees that Step 2A, Prong two must be applied. The disagreement concerns the result of that analysis. The additional elements, RDM, CDC, PDSes, AAML module, feedback analysis module, and real time access, do not integrate the abstract idea into a practical application because they are not recited as generic computing tools performing their ordinary data processing functions: RTDM collects and transforms data; CDC detect and propagates data changes; PDSes store schema defined data; AAML analyses data and generates predictions; the feedback module monitors results; and the AAML module modifies parameters and subsequently repeats the analysis.
The claim does not impose a meaningful technological limitation on how these functions are performed. Instead, it applies the abstract analysis and prediction process in the technological environment of a real time data architecture. Limiting an abstract idea to a particular computer environment or data architecture does not, without a claimed technological improvement, integrate the exception into a practical application.
Applicant argues “As recited, the claims integrate predictive analytics into a real-time data mesh that applies CDC to detect data mutations, persist updated records into structured, schema-defined data stores, and automatically re-execute analytics when data changes affect an executing analysis. This configuration improves the functioning of the computing system itself by preventing analytics outputs from becoming stale or inconsistent due to mid-execution data changes, thereby ensuring that predictive results accurately reflect the current data state. The claimed architecture thus applies any alleged abstract concepts in a concrete and practical manner that is tightly bound to the operation of distributed data processing systems.” (remarks p. 11).
Examiner respectfully disagrees. The identified benefit is an improvement to the informational accuracy or timeliness of the predictive output, not an improvement to the functioning of the computer, database, network, or machine learning model itself. The claim does not recite how it determines that a data modification “affect” the analysis beyond stating the desired result. Nor does it recite how consistency is measured, enforced, or verified. The limitation “to maintain consistency” states the intended result of using updated data, rather than a technical process for achieving a particular consistency model.
At most, the claim ensures that an analysis is run again with more current information. More accurate, current, or complete information resulting from reprocessing updated data does not establish an improvement in computer functionality.
Applicant argues “This structural and functional improvement is analogous to the claims found patent- eligible in DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1257-59 (Fed. Cir. 2014). There, the Federal Circuit held that claims directed to a web page framework for retaining visitors at a host website by dynamically generating a hybrid webpage were patent-eligible, even though they involved known web technologies. What made the claims eligible, the Court explained, was that the ordered combination of elements solved a specific, non-abstract problem "rooted in computer technology," a characterization that equally applies to the present invention's use of CDC-updated, machine learning-integrated PDS structures and defined machine-learning operations, within a real-time system.” (remarks p. 11).
Examiner respectfully disagrees. DDR Holdings is distinguishable. The claims there altered the conventional operation of internet hyperlink navigation by generating a particular hybrid webpage that combined visual elements of a host website with content associated with a third party merchant. The claimed solution did not merely use the internet to perform a preexisting information processing practice. Here, claim 1 does not alter the operation of CDC, databases, data meshes, or neural networks. Each component performs its ordinary function: CDC detects data changes, PDSes store data, and the AAML module analyses data. Repeating analysis when its input changes does not create a new mode of computer or network operation analogous to the hybrid webpage mechanism in DDR holdings.
The statement that a problem is “rooted in computer technology” is not independently sufficient. The claim must recite the particular technological solution, rather than only the desired result of maintaining current or consistent information.
Applicant argues “Moreover, the invention reflects the kind of non-conventional arrangement of known components recognized in Amdocs (Israel) Ltd. v. Openet Telecom, Inc., 841 F.3d 1288, 1300- 01 (Fed. Cir. 2016). In Amdocs, the Court upheld claims that used generic components (e.g., network monitors, gatherers) in a distributed fashion to yield a non-abstract improvement in functionality. The Court emphasized that "the claim's enhancing limitation necessarily requires that these generic components operate in an unconventional manner to achieve an improvement in computer functionality." Id. at 1301. Here, similarly, the claim does not simply apply predictive analytics in a vacuum. Rather, it uses CDC mechanisms to continuously update PDSes, each structured and purposed for a specific data type, and feeds this categorized and synchronized data into tailored neural networks (CNN/RNN) within the AAML module. The result is a dynamic, feedback-driven architecture that refines its own weighting parameters and publishes updates into live systems. This is a machine-specific, real-time infrastructure that could not be performed by humans and does not simply automate known abstract practices. As Amdocs concluded: "the claim is tied to a specific structure of various components operating in a distributed manner to achieve a technological solution to a technological problem." Id. at 1301. That logic applies squarely here.” (remarks p. 12).
Examiner respectfully disagrees. Amdocs is distinguishable. The eligible claims in Amdocs required a particular distributed arrangement in which network accounting information was enhanced close to its source, thereby reducing congestion and resource usage associated with transporting and processing massive records at a centralized location.
Claim 1 does not similarly define where the RTDM, PDSes, and AAML module are physically or logically located relative to the data sources. It does not require that processing be distributed among particular network components, reduce network traffic, reduce memory consumption, or relocate processing to solve a computer resource problem. Calling the collection of modules a “distributed architecture” does not supply the concrete distributed arrangement required by Amdocs. The modules are functionally identified but are not claimed as operating in a specific unconventional placement or communication structure.
Applicant additionally argues that the architecture “publishes updates into live systems” that limitation does not appear in the presently argued claim and therefore cannot establish eligibility. Arguments based on unclaimed features are not persuasive.
Applicant argues “Furthermore, the Federal Circuit has clarified that even if claim limitations, considered individually, could be seen as conventional or routine, the ordered combination can still amount to a patent-eligible inventive concept. See Bascom Glob. Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341, 1350 (Fed. Cir. 2016) ("an inventive concept can be found in the non- conventional and non-generic arrangement of known, conventional pieces"). Applicant's amended claim presents such a non-conventional arrangement by combining real-time ingestion via CDC, intelligent structuring via PDSes, predictive analysis via dual- modal ML models, and iterative refinement using real-world feedback loops. Accordingly, the claims as a whole integrate any arguable abstract ideas into a practical application. Thus, the claims are eligible for a patent under 35 U.S.C. § 101 because it is not directed to a judicial exception, and it is unnecessary for us to proceed to Step 2B.” (remarks p. 12-13).
Examiner respectfully disagrees. Examiner agrees with the general proposition that an inventive concept may reside in an ordered combination. Applicant, however, has not identified a nonconventional arrangement comparable to BASCOM.
BASCOM involved installation of a filtering tool at a specific location remote from end users while retaining individualized filtering controls. That particular placement provided benefits associated with both local and remote filtering. In contrast, claim 1 recites the ordinary sequence of: collecting and standardizing data; updating stored data when changes are detected; analyzing the stored data; repeating the analysis when the data changes; monitoring results and feedback; and modifying parameters for a later execution. The claimed sequence does not require an unconventional location, interconnection, allocation of functionality, or processing mechanism. The recited components merely perform the functions ordinarily associated with data ingestion, database updating, predictive analytics, monitoring, and model refinement.
Furthermore, “dual modal ML models” overstates the claim. The claim requires “at least one of” a CNN or an RNN; it does not require that both models be used or that their outputs be combined. Thus, the claim does not require a dual modal architecture.
Applicant argues “The claimed invention is not directed to a judicial exception, or in the alternative, integrates any alleged exception into a practical application using a non-conventional system architecture. Applicant respectfully submits that the claims as amended should be found patent- eligible under 35 U.S.C. § 101, consistent with DDR Holdings, Amdocs, and Bascom.
Since the claims are not directed to an abstract idea for at least the reasons outlined above, Applicant respectfully requests that the Examiner withdraw the rejections under § 101 and pass all claims to allowance” (remarks p. 13).
Examiner respectfully disagrees. Even when considered individually and as an ordered combination, the additional elements do not amount to significantly more than the abstract idea. The claim invokes RTDM, CDC, PDSes, CNN/RNN processing, real time monitoring, and model parameter refinement at a high level of generality. The claim does not recite a technological modification to any of those components. Rather, the components are used according to their established functions to obtain updated data, repeat predictive analysis, and improve future predictive outputs.
This conclusion is not based merely on finding each element separately in the prior art. Instead, the eligibility deficiency is that the claimed combination itself remains functionally and result oriented. The combination does not recite a specific implementation that changes how the computer, database, network, CDC mechanism, or machine learning model operates. Accordingly, the claims are directed towards an abstract idea that is not integrated into a practical application nor does it provide significantly more.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim(s) 1-7 is/are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Claim 1 recites “d) detecting, during execution of the analysis, a CDC-detected data modification affecting at least one PDS accessed by the AAML module; (e) in response to detecting the CDC-detected data modification, re-executing the analysis by the AAML module using updated data from the affected PDS to maintain consistency between the predictive insights associated with vendor product roadmaps and the underlying data;”
the limitation above is new matter because Examiner is unable to find support for it in the specification. The closest support in the specification discloses CDC generally. Paragraph 91 states “ The RTDM module 310 can be configured to capture changes in data across multiple transactional systems in real-time. It employs a sophisticated Change Data Capture (CDC) mechanism that constantly monitors the transactional systems, detecting any updates or modifications.” Paragraph 144 further states “ [0144] To ensure real-time data synchronization, data layer 620 can be configured to employ one or more change data capture (CDC) mechanisms. These CDC mechanisms can be integrated with the transactional systems, such as legacy ERPs like SAP, Impulse, META, and I-SCALA, as well as other enterprise-wide systems. CDC constantly monitors these systems for any updates, modifications, or new transactions and captures them in real-time. By capturing these changes, data layer 620 ensures that the data within the data lake 622 and PDSes 624 remains current, providing users with real-time insights into the distribution ecosystem.”
In regards to PDSes, the specification discloses in paragraph 139 ,” Within the data mesh, multiple Purposive Datastores (PDS) can be deployed to store specific types of data, such as customer data, product data, or inventory data. Each PDS can be optimized for efficient data retrieval based on specific use cases and requirements. The PDSes can be configured to store specific types of data, such as customer data, product data, finance data, and more. These PDS serve as repositories for canonized and/or standardized data, ensuring data consistency and integrity across the system.”
These paragraphs support CDC monitoring of transactional systems and updating the data maintained in the data lake and PDSes. They do not describe CDC monitoring of an AAML execution. In particular, they do not disclose: an AAML analysis that is presently executing against a particular PDS; detecting a CDC event while that analysis remains in execution; determining that the CDC event affects data accessed by that execution; invalidating the executing analysis or its intermate results; and/or initiating another execution of the analysis because the relevant PDS changed during the first execution.
Paragraph 144 states that CDC “constantly monitors” transactional systems. That disclosure establishes continuing CDC operation, but it does not establish that the CDC mechanism monitors the execution state of an AAML analysis or correlates a detected modification with data being used by a particular ongoing AAML execution.
Likewise, paragraph 144 explains that CDC keeps PDS data “current”. It does not explain what happens when the PDS becomes current after an AAML analysis has already begun using a prior data state.
Paragraph 150 discloses “ RTDM module 600 can include an AI module 630 configured to implement one or more algorithms and machine learning models to analyze the stored data in data layer 620 and derive meaningful insights. In some non-limiting examples, AI module 630 can apply predictive analytics, anomaly detection, and optimization algorithms to identify patterns, trends, and potential risks within the supply chain. AI module 630 can continuously learns from new data inputs and adapts its models to provide accurate and current insights. AI module 630 can generate predictions, recommendations, and alerts and publish such insights to dedicated data feeds.” This paragraph supports analyzing stored data, learning from new data inputs, and adapting models to provide current insights. Nevertheless, “continuously learns from new data inputs” does not describe re-executing an ongoing analysis in response to a CDC event detected during that analysis.
Continous learning may encompass successive training, periodic refinement, incremental model updating, or subsequent analyses based on newly received data. The paragraph does not disclose that a particular execution is monitored for CDC events, that the data used by that execution is determined to have changed, or that the execution is restarted or re-executed as a consequence.
In regards to AAML, paragraph 156 discloses “[0156] In an embodiment, FIG. 7 depicts System 700 configured for generating predictive insights for vendor product roadmaps. System 700 includes Real-Time Data Mesh (RTDM) 710, Single Pane of Glass User Interface (SPoG UI) 705, Advanced Analytics and Machine-Learning (AAML) Module 715, and Product Roadmap Insights (PRI) Module 720. System 700 can integrate streaming data processing capabilities within RTDM 710, enabling the handling of high-velocity, real-time data essential for immediate market responsiveness. RTDM 710 aggregates and standardizes data from diverse sources and processes this information in real-time, providing a foundation for dynamic, data-driven decision-making. RTDM 710 also integrates advanced data security measures, including encryption and fine-grained access controls, safeguarding sensitive information and ensuring compliance with stringent data protection regulations.”.
Paragraph 159 discloses “ [0159] Advanced Analytics and Machine-Learning (AAML) Module 715 serves as the analytical engine of System 700, leveraging specialized algorithms to analyze data aggregated by RTDM 710. AAML Module 715 utilizes big data analytics tools and deep learning capabilities to conduct sentiment analysis, trend forecasting, and customer behavior analytics. It integrates machine learning algorithms trained on historical and real-time data to identify patterns and generate predictive insights for product roadmap adjustments. The module's architecture supports continuous learning, enabling the refinement of algorithms based on feedback loops to improve the precision and accuracy of insights.” This disclosure establishes that RTDM 710 handles real time data and that AAML module 715 analyzes historical and real time data. They also establish continuous learning through feedback loops. They do not disclose the claimed interaction between CDC and the execution state of AAML module 715.
More specifically, the paragraphs above, do not state that the AAML module beings an analysis using data from a PDS; a CDC modification is detected while that analysis is still executing; the system determine that the modified data is contained in or affects the PDS being accessed by that execution; and the AAML analysis is re-executed in response to that determination.
The statement that machine learning algorithms are trained on “historical and real time data” describes the type of freshness of the input data. It does not disclose monitoring for a change occurring after an analysis begins or restarting the analysis because of that mid-execution change.
Paragraph 169 discloses “ [0169] At operation 810, the system initiates the predictive analytics process. AAML Module 715 begins by aggregating data collected by RTDM 710, which includes an extensive range of inputs such as customer feedback, market trends, competitive intelligence, and product performance metrics. This aggregation process is foundational, ensuring that the analysis is based on a holistic view of the data landscape relevant to product development and market positioning.” And paragraph 171 discloses “ [0171] At operation 830, AAML Module 715 can perform trend forecasting and customer behavior analytics. Using statistical models, machine learning algorithms and/or a combination thereof, AAML Module 715 analyzes historical and real-time data to predict future market trends and customer behaviors. This predictive modeling considers factors such as seasonal variations, market dynamics, and emerging trends to forecast future states and identify potential opportunities or challenges for product strategies.”
Method 800 describes initiating the predictive analytics process, aggregating data collected by RTDM, and then analyzing historical and real time data. It does not disclose detecting a CDC modification during operation or returning to operation 810 or operation 830 in response to such a modification. Indeed, the disclosed sequence indicates that RTDM collected data is aggregated as a foundation for the analysis. Nothing in paragraphs 169/171 describes what occurs if the underlying PDS changes after the analysis begins.
Paragraph 181 discloses “ [0181] At operation 950, the system establishes feedback loops for continuous learning and adaptation. This involves continuously refining predictive models and strategies based on real-time market response and product performance data. By incorporating feedback mechanisms, the system ensures that the predictive analytics and roadmap optimization processes remain dynamic and responsive to changing market conditions and customer feedback.”
This paragraph supports refinement of predictive models and strategies based on real time market response and product performance data. However, operation 950 follows the generation and implementation of roadmap adjustments. It therefore describes a feedback process based on the results or market effects of prior predictive and roadmap processes, not detection of a PDS modification while an AAML analysis is executing. The specification’s feedback loop may support subsequent model refinement and later analysis using updated information. It doesn’t support aborting, invalidating, restarting, or re-executing an analysis because a CDC modification occurred during the analysis.
Paragraph 189 discloses “ [0189] At operation 1050, the standardized and secured data is subjected to AAML Module 715 analysis. Operation 1050 can include performing analytics to process data in real-time, allowing for the immediate identification of trends, insights, and potential opportunities about product roadmaps. Through this dynamic analysis, the system can generate actionable intelligence that informs decision-making and strategy formulation.” And paragraph 191 discloses “ [0191] At operation 1070, the PIPR system establishes a continuous feedback loop. This operation involves monitoring the outcomes of implemented strategies and the performance of products in the market. The feedback collected through this loop is fed back into the system, enabling continuous learning and adaptation of the predictive models and strategies based on real-world performance data and market response.” Paragraph 189 supports processing data in real time. But “processing data in real time” does not disclose detecting a CDC event during that processing and re-executing the processing because its underlying data changed. Paragraph 191 expressly describes monitoring “the outcomes of implemented strategies” and “the performance of products in the market”. That feedback necessarily concerns results occurring after predictive insights or strategies have been generated and implemented. It supports later learning and adaptation, not re-execution triggered by a CDC modification during an ongoing AAML analysis.
The above disclosures may reasonably support repeated or subsequent analysis using updated real-time data. They do not reasonably convey possession of the narrower claimed combination requiring an AAML analysis presently executing against at least one PDS; detection of a CDC modification during that execution; a determination that the modification affects the PDS accessed by the ongoing execution; and re-execution of the analysis in response to the mid execution modification to maintain consistency between the predictive insights and the underlying data.
Accordingly, the limitation is new matter.
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 1-7 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.
Claim 1 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claim recites “(a) collecting roadmap heterogeneous data from at least two predefined sources, wherein the heterogeneous roadmap data comprises structured and unstructured data representing one or more aspects of product planning, user behavior, or environmental signals, and transforming the collected data into a standardized format for downstream processing; (b) analyzing the standardized roadmap data to identify one or more predictive patterns, (c) generating predictive insights associated with vendor product roadmaps, wherein the predictive insights identify one or more proposed modifications to product development attributes (d) detecting data modification (e ) re-executing the analysis; (f) monitoring execution results of the re-executed analysis in real time; and (g) refining a set of weighting parameters used in pattern analysis based on the monitored execution results and market feedback, wherein the refined parameters are applied in a subsequent execution of the method to iteratively improve predictive accuracy under changing data conditions for future roadmap predictions.”
The limitations above, as drafted, is a process that, under its broadest reasonable interpretation, covers collecting, standardizing, and analyzing enterprise data to generate predictive insights and adjust product development roadmaps which is a method of organizing a human activity (i.e. planning and adjusting vendor product strategies, including marketing, pricing, feature selection and timeline management) and mathematical concepts (analyzing data with CNN/RNN models, generating and refining predictive weights, detecting and propagating changes). That is, the method allows for commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations), managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions) and mathematical concepts and relationships.
This judicial exception is not integrated into a practical application. In particular, the claim recites “a Real-Time Data Mesh (RTDM) module”, “an Analytics and Machine-Learning (AAML) module”, “roadmap optimization module”, “wherein the model includes at least one of a convolutional neural network for image data patterns and a recurrent neural network for temporal data patterns”, “feedback analysis module” and “wherein the RTDM applies change data capture (CDC) to detect and propagate data modifications in a data layer, and updates one or more persistent Purposive Data Stores (PDSes) based on the CDC-detected changes, wherein each PDS comprises a structured, dynamically updateable data repository that persists records corresponding to a defined schema for a specific supply chain data domain, and configured for real-time access by an Analytics and Machine Learning (AAML) module for predictive insight generation and roadmap optimization”. The claim further recites “(d) detecting, during execution of the analysis, a CDC-detected data modification affecting at least one PDS accessed by the AAML module; (e) in response to detecting the CDC-detected data modification, re-executing the analysis by the AAML module using updated data from the affected PDS to maintain consistency between the predictive insights associated with vendor product roadmaps and the underlying data.”
Each of the additional limitations is recited at a high level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (i.e. ingesting data streams, synchronizing updates, storing in database, applying ML architectures, and outputting results to business systems). The specification describes these using conventional big data/ML paradigms, and the claim does not require any specific improvement to computer functionality (e.g. no new network protocol, data structure, or measurable performance gain). Simply executing the analysis “in real time” on a computer does not impose a meaningful limit on the abstract idea.
Furthermore, in regards to the detecting the CDC modification and re-executing limitation, the claim does not require improving the CDC mechanism, the AAML module or the PDSes in any way. Instead, the claim functionally requires the result: when data changes during analysis, run the analysis again with updated data. That is an instruction to apply the abstract analysis again under specified data conditions, rather than a particular improvement to CDC, PDS operation, concurrency control, or machine learning execution.
Accordingly, these additional elements, alone or in combination, do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
The claim does not include additional elements, alone or in combination, that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements, alone or in combination, are nothing more than mere instructions to apply the exception on a general computer.
Dependent claim 2 is also directed to an abstract idea without significantly more because it further narrows the abstract idea described in relation to claim 1 without successfully integrating the exception into a practical application (historical data repository of the RTDM is recited at a high level of recitation which amounts to mere instructions to apply the exception in a computer environment) or providing significantly more limitations.
Dependent claim 3 is also directed to an abstract idea without significantly more because it further narrows the abstract idea described in relation to claim 1 without successfully integrating the exception into a practical application (new data sources and types is recited at a high level of recitation which amounts to mere instructions to apply the exception in a computer environment) or providing significantly more limitations.
Dependent claim 4 is also directed to an abstract idea without significantly more because it further narrows the abstract idea described in relation to claim 1 without successfully integrating the exception into a practical application (a user interface is recited at a high level of recitation which amounts to mere instructions to apply the exception in a computer environment) or providing significantly more limitations.
Dependent claim 5 is also directed to an abstract idea without significantly more because it further narrows the abstract idea described in relation to claim 1 without successfully integrating the exception into a practical application (automated communication channels are recited at a high level of recitation which amounts to mere instructions to apply the exception in a computer environment) or providing significantly more limitations.
Dependent claim 6 is also directed to an abstract idea without significantly more because it further narrows the abstract idea described in relation to claim 1 without successfully integrating the exception into a practical application or providing significantly more limitations.
Dependent claim 7 is also directed to an abstract idea without significantly more because it further narrows the abstract idea described in relation to claim 1 without successfully integrating the exception into a practical application or providing significantly more limitations.
Non-obviousness and Novelty
No prior art was applied to claims 1-7 because Examiner is unaware of any prior art alone or in combination which discloses the limitation of claim 1.
The closest prior art are:
Meharwade (US 20190050771)
Confluent, “Data products, Data contracts, and change data capture”, published by Confluent.com on February 21, 2024, hereinafter “Confluent”.
Devore (US 20230282350)
Negi (US 11568305)
Shaya (US 7809601)
The proposed combination is a piecemeal and hindsight driven because each reference is being used to supply a separate claim limitation that is absent from the primary reference, rather than because the primary reference itself suggests the claimed architecture and execution sequence.
Meharwade teaches, at most, a computer implemented product development planning system that uses artificial intelligence and machine learning for agile planning. In claim language, Meharwade teaches “generating predictive insights associated with vendor product roadmaps” because meharwade discloses AI/ML based product development, release planning, iteration planning and planning aligned with a product roadmap
“predictive insights identify one or more proposed modifications to product development attributes” because Meharwade analyzes backlog, requirements, story rank, priority, size, dependencies, timelines, team velocity, task types, and task efforts
“analyzing…data to identify one or more predictive patterns” because it uses historical/project data and ML to predict tasks, efforts, and planning outcomes.
Confluent is brought in to supply the missing RTM/PDS/CDC architecture. Confluent teaches:
“heterogenous…data from at least two predefineds sources” because it discusses building data products from multiple source tables and disparate sources.
“transforming the collected data into a standardized format for downstream processing” because it teaches data products with schemas, metadata, contracts, and modeled external data formats.
“CDC to detect and propagate data modifications in a data layer” because it teaches CDC sourced data products and near real time streams.
“persistent…data repositories corresponding to a defined schema for a specific data domain”…although generally, through data products, schemas, metadata, and domain oriented data products.
Devore teaches:
Detecting…a data modification” because it receives new or updated data events.
“in response to detecting…the data modifications, re-executing the analysis..using updated data” because it re-executes ML models when updated data is received.
“maintain consistency…with the underlying data” only generally because it updates injury probability outputs based on newly received accident data.
Negi teaches:
“recurrent neural network configured for temporal customer feedback data” because it sues RNN/LSTM sequence modeling for temporal customer journey/customer behavior events.
“Heterogeneous…user behavior…data” because it processes heterogeneous customer journey events, interactions, timestamps, ratings, transcripts, outcomes, and user behavior data.
Shaya disclsoes
“monitoring…market response signals” through consumer/product performance response feedback
“refining…weighting parameters used in pattern analysis based on…market feedback” because it uses consumer feedback to update training data and adjust neural network connection weights.
“iteratively improve predictive accuracy…for future…predictions” because it teaches improving future recommendation accuracy based on feedback.
Thus, Meharwade provides only the broad product roadmap/product development planning environment. Confluent is separately needed for the claimed RTM/PDS/CDC architecture. Devore is separately needed for the claimed updated data triggered ML re-execution. Negi is separately needed for the claimed RNN temporal customer data analysis. Shaya is separately needed for the claimed feedback based weighting parameter refinement.
Nothing in Meharwade teaches or suggests modifying its AI/ML product development planning assistants to use CDC updated PDSes, detect CDC detected data modifications during execution of roadmap analysis, re-execute the AAML analysis using updated PDS data to maintain consistency, and then refine roadmap analysis weighting parameters based on execution results and market feedback. The rejection, therefore, reconstructs the claim only by selecting isolated teachings from five unrelated references after Applicant’s claimed arrangement.
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 OMAR ZEROUAL whose telephone number is (571)272-7255. The examiner can normally be reached Flex schedule.
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, Lynda Jasmin can be reached at (571) 272-6782. 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.
OMAR . ZEROUAL
Examiner
Art Unit 3628
/OMAR ZEROUAL/Primary Examiner, Art Unit 3629
Prosecution Insights