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 Claims
This is a Final Action on the merits in response to the claims filed on 1/22/2026.
Claims 1 – 20 have been amended;
Claims 1 – 20 are currently pending in this application.
Response to Remarks
Examiner’s Response to Remarks:
Rejections under 35 U.S.C. § 101;
Rejections under 35 U.S.C. § 103.
Examiner’s Response to Rejections Under 35 U.S.C. § 101.
Applicant argues the claims are patent-eligible under the two-part test set forth in Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 134 S. Ct. 2347 (2014). The claims are not directed to an abstract idea and, set forth "significantly more" than any abstract idea and thus are eligible.
Examiner respectfully disagrees. Applicant’s claim 1 is directed to a method, which is a statutory category. However, claim 1 recites the abstract idea of mathematical concepts where the claim is computing a feature vector; and computing an estimated revenue: computing a user adoption propensity of the product; computing a usage of the product by the user; computing an overall revenue growth of the user from the one or more products; and computing a retention of the user. However, the limitations are merely performing mathematical calculations. Claim 1 also recites certain methods of organizing human activity, where the claim involves commercial interactions such as sales activities, as the claim involves product adoption by user, user adoption propensity, as well as revenue growth of the user from the one or more products. Thus claim 1 recites an abstract idea.
The additional elements a user, a platform, one or more products, a provider, consumers, a first trained machine learning model, a multi-stage adoption funnel, a second trained machine learning model, a third trained machine learning model, and a fourth trained machine learning model are merely generic computer components performing generic computer functions. Although the claim recites trained machine learning models as additional elements, the trained machine learning model merely constitute a mathematical concept, such as the concept of using known data to set and adjust coefficients; and so the trained machine learning models are generic computer components performing generic computer functions and amount to no more than mere instructions using generic computer components to implement the judicial exception; and thus Applicant’s claim 1 is not an improvement and is merely an abstract idea. Contrastingly, Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, “Enfish”, claims recites an improvement in existing technology and the claims are patent eligible. In Applicant’s matter, claim 1 merely provides instructions to apply an exception using generic computer components, as the claim resolves the business problem of evaluating adoption propensity and retention of users with products on its platform. At each stage of the multi-stage adoption funnel, Applicant is merely running a trained model for a calculation or an estimation, such as for evaluating an estimated revenue. Regarding Ex parte Desjardins, Appeal No. 2024-000567, at 9 (P.T.A.B. September 26, 2025) “Desjardins,” when evaluating Applicant’s Specification ¶¶ 0036 – 0039, and 0044 – 0046, to determine if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement in the functioning of a computer, or an improvement to other technology or a technical field, the evaluation reveals there are no details of an improvement to the functioning of a computer, or an improvement to other technology or a technical field, as these paragraphs merely discuss the stages of the multi-stage adoption funnel, the data collected, the analysis, and results for an estimation, such as revenue, which describes a business problem, and the data analysis to resolve the business problem of calculating and estimating revenue; and this is very different than Desjardins recited improvements.
Further when evaluating claim 1 in Step 2B to determine whether the additional elements in combination (as well as individually) amount to an inventive concept, e.g., because they are more than the non-conventional and non-generic arrangement of known, conventional elements, Examiner determines the additional elements are not more than the non-conventional and non-generic arrangement of known, conventional elements; as running trained machine learning models for estimating, notwithstanding the trained machine learning models are sequentially, in parallel, or concurrently, are conventional methods. See Low, Yucheng, et al. "Graphlab: A new framework for parallel machine learning." arXiv preprint arXiv:1408.2041 (2014); Kim, Jin Kyu, et al. "Strads: A distributed framework for scheduled model parallel machine learning." Proceedings of the Eleventh European Conference on Computer Systems. 2016; and Lee, Jungwon, et al. "A comparison and interpretation of machine learning algorithm for the prediction of online purchase conversion." Journal of Theoretical and Applied Electronic Commerce Research 16.5 (2021): 1472-1491. In contrast In Applicant’s case, when performing statistical analysis of trained machine learning models of claim 1 and the described statistical analysis of the model trainer in view of Applicant’s Specification 0062, an improvement is based on accuracy and iterations of the model.
Claims 8 and 14 are substantially similar and recite the same subject matter as claim 1; and the dependent claims inherit the same deficiencies as the independent claims. Thus, claims 1 – 20, are rejected under 35 U.S.C. § 101.
Examiner’s Response to Rejections Under 35 U.S.C. § 103.
Applicant argues under 35 U.S.C. § 103 the amended claims are patentable over Qu, Huashuai et al. (U.S. Publication No. 2024/0420180) in view of Fotso, Stephane et al. (U.S. Publication No. 2023/0214890).
Examiner respectfully disagrees. Applicant has amended claims 1 – 20. A new search was necessitated due to the amendments and new art has been applied. Thus claims 1 – 20, remain rejected under 35 U.S.C. § 103.
Claim Objection
Claim 10 is objected to because Claim 10 is not written in its proper amended form.
Appropriate correction is required.
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.
Claims 1-20 are rejected under 35 U.S.C. 112(a) 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(s), at the time the
application was filed, had possession of the claimed invention.
Claims 1, 3 – 9, 11 – 14, and 18 are directed to the limitation of “yield” and is new matter. Applicant’s Specification in its entirety is silent to “yield”.
Therefore, the claims and their dependent claims are rejected under 35 U.S.C. 112(a), written description, as being directed to non-statutory subject matter.
Claim Rejections – 35 U.S.C. §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 – 20 are rejected under 35 U.S.C. §101 because the claimed invention is directed towards an abstract idea without significantly more.
Claims 1, 8, and 15 recites:
evaluating a feature vector;
evaluating an estimated yield of the one or more products based on a plurality of usage factors where in the estimated yield is configured to adjust based on an increased promotion of the one or more products and the plurality of usage factors represent intermediate outputs of corresponding stages of the multi-stage adoption funnel , the plurality of usage factors including;
evaluating a user adoption propensity of the one or more products;
evaluating a usage of the one or more products;
evaluating an overall yield growth of the user from the one or more products;
and evaluating a retention of the user;
The limitations of claim 1, under its broadest reasonable interpretation, recites the abstract idea of mathematical concepts; and uses a computer as a tool to perform mathematical concepts. For example, claim 1 is evaluating a feature vector; evaluating an estimated yield of the one or more products and evaluating a user adoption propensity of the one or more products; evaluating a usage of the one or more products; evaluating an overall yield growth of the user; and evaluating a retention of the user; and all involve evaluation of data and particularly recites mathematical calculations. Claims 8 and 14 are substantially similar and recite the same subject matter as claim 1. Accordingly, claims 1, 8, and 14 recite mathematical concepts.
Additionally, the claim limitations above further recite user product distribution growth through product adoption, user adoption propensity, and computing revenue through overall revenue growth. This falls within the abstract idea grouping of certain methods of organizing human activity, specifically sales activities and commercial interactions. It also falls under the subgrouping of activity of where the activity is between a person such as in the instant claim, a user, and a platform. See MPEP 2106.04(a)(2)(II).
The dependent claims encompass the same abstract ideas as well. For instance, claims 2, 10, and 16 are directed towards observing the first trained machine learning model is trained based on training data from historical interactions between a plurality of existing users and a plurality of existing platforms providing access to corresponding selections of the plurality of available products; claims 3, 11, and 17 are directed towards observing the user is a prospective user of the platform, and the second trained machine learning model is a model that evaluates expected yield for the prospective user; claims 4, 12, and 18 are directed towards observing the feature vector is evaluated based on one or more of: visits to a product page on a website of the provider; total yield of the user; time on platform of the user; and industry of the user; claim 5 is directed towards observing the platform is associated with a first industry, and the method further comprises evaluating a second estimated yield based on a second platform adopting the one or more products, and the second platform is associated with a second industry different from the first industry; claim 6 is directed towards evaluating a plurality of feature vectors representing a plurality of users of the platform; and evaluating an estimated overall yield for the plurality of users based on the plurality of feature vectors; claim 7 is directed towards evaluating an estimated overall yield based on the plurality of machine learning models; ranking the plurality of available products based on the estimated overall yield corresponding; and displaying the ranking of the plurality of available products on a user interface; claim 9 is directed towards evaluating a ranking of the plurality of available products based on an overall estimated yield for each of the plurality of available products, the overall estimated yield based on the user adoption propensity, the usage, the overall yield growth, and the retention of the user; claim 13 is directed towards observing the instructions to evaluate the estimated yield… adoption propensity for the one or more products, representing a likelihood that the platform will adopt the one or more products; claim 15 is directed towards observing the feature vector is evaluated based on text information and non-text information; claim 19 is directed towards evaluating a report of the ranking of the one or more products; and claim 20 is directed towards transmit a promotion of the one or more products to the platform, the one or more products being selected based on their ranking. Thus, the dependent claims further limit the abstract ideas.
These judicial exceptions are not integrated into a practical application. Claim 1 recites the additional elements of a user platform providing access to one or more products, a first trained machine learning model, plurality of machine learning models, a machine learning model comprised in a multi-stage user adoption funnel, a second trained machine learning model, a third trained machine learning model, and a fourth trained machine learning model, supplying the feature vector to a first, second, third, and fourth trained machine learning model; Claim 8 recites the additional elements of a user platform providing access to one or more products, a first trained machine learning model, plurality of machine learning models, a machine learning model comprised in a multi-stage user adoption funnel, a second trained machine learning model, a third trained machine learning model, and a fourth trained machine learning model, a processor, memory, and a system, supplying the feature vector to a first, second, third, and fourth trained machine learning model; claim 14 recites the additional elements of a non-transitory computer-readable storage medium, processor, a machine learning model comprised in a multi-stage user adoption funnel, a first trained machine learning model, plurality of machine learning models, a multi-stage user adoption funnel, a second trained machine learning model, a third trained machine learning model, and a fourth trained machine learning model, supplying the feature vector to a first, second, third, and fourth trained machine learning model which are generic computer components as per Applicant’s Specifications shown below:
“[00107] FIG. 8 is a block diagram illustrating components of a machine 800, according to some example embodiments, able to read instructions from a non-transitory machine-readable medium (e.g., a computer-readable storage medium) and perform any one or more of the methodologies discussed herein. Specifically, FIG. 8 shows a diagrammatic representation of the machine 800 in the example form of a computer system, within which instructions 810 (e.g., software, a program, an application, an applet, an app, or other executable code) for causing the machine 800 to perform any one or more of the methodologies discussed herein may be executed. As such, the instructions 810 may be used to implement modules or components described herein. The instructions 810 transform the general, non-programmed machine 800 into a particular machine 800 programmed to carry out the described and illustrated functions in the manner described. In alternative embodiments, the machine 800 operates as a standalone device or may be coupled (e.g., networked) to other machines. In a networked deployment, the machine 800 may operate in the capacity of a server machine or a client machine in a server-client network environment, or as a peer machine in a peer-to-peer (or distributed) network environment. The machine 800 may include, but not be limited to, a server computer, a client computer, a personal computer (PC), a tablet computer, a laptop computer, a netbook, a set-top box (STB), a personal digital assistant (PDA), an entertainment media system, a cellular telephone, a smart phone, a mobile device, a wearable device (e.g., a smart watch), a smart home device (e.g., a smart appliance), other smart devices, a web appliance, a network router, a network switch, a network bridge, or any machine capable of executing the instructions 810, sequentially or in parallel or concurrently, that specify actions to be taken by the machine 800. Further, while only a single machine 800 is illustrated, the term “machine” or “processing circuit” shall also be taken to include a collection of machines that individually or jointly execute the instructions 810 to perform any one or more of the methodologies discussed herein.”
and thus are not practically integrated nor significantly more.
Each of the additional limitations are no more than mere instructions to apply the exception using generic computer components (e.g., processor). The combination of these additional elements are no more than mere instructions to apply the exception using generic computer components (e.g., processor). The combination of these additional elements are no more than mere instructions to apply the exception using generic computer components (e.g., processor). The claims do not include additional elements that are sufficient to amount significantly more than the judicial exception because the additional elements do not impose meaningful limits on practicing the idea, and amount to no more than mere instructions using generic computer components to implement the judicial exception. Mere instructions to apply an exception using generic computer components cannot provide an inventive concept.. Thus, the claims are directed to an abstract idea.
Dependent claims 2 – 7, 9 – 13, and 15 – 20, when analyzed both individually and in combination are also held to be ineligible for the same reason above and the additional recited limitations fail to establish that the claims are not directed to an abstract idea. The additional limitations of the dependent claims when considered individually and as an ordered combination do not amount to significantly more than the abstract idea.
Looking at these limitations as an ordered combination and individually add nothing additional that is sufficient to amount to significantly more than the recited abstract idea because they simply provide instructions using generic computer components, to “apply” the recited abstract idea. Thus, the elements of the claims, considered both individually and as an ordered combination, are not sufficient to ensure that the claim as a whole amount to significantly more than the abstract idea itself. Therefore, claims 1 – 20 are not patent eligible.
Claim Rejections – 35 U.S.C. § 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.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103(a) are summarized as follows:
Determining the scope and contents of the prior art.
Ascertaining the differences between the prior art and the claims at issue.
Resolving the level of ordinary skill in the pertinent art.
Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1 – 20, are rejected under 35 U.S.C. § 103 as being unpatentable over Qu, Huashuai et al. (U.S. Publication No. 2024/0420180) hereinafter “Qu” in view of Fotso, Stephane et al. (U.S. Publication No. 2023/0214890) hereinafter “Fotso” in view of Userpilot Team “Understanding the Adoption Funnel: Stages, Examples & More” https://userpilot.com/blog/adoption-funnel/ (January 16, 2023) hereinafter “Userpilot Team”.
Claims 1, 8, and 14:
determining, via a plurality of machine learning models, an estimated yield of the one or more products based on a plurality of usage factors, wherein the estimated yield is configured to adjust based on an increased promotion of the one or more products, and the plurality of usage factors represent intermediate outputs, the plurality of usage factors including: there is no support for “yield” in Applicant’s Specification; yield is taken to mean increase, produce, revenue, output, or the like; Qu teaches in ¶ 0020, components of a processing circuit or a processor, according to some example embodiments, configured to read instructions from a non-transitory computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methods discussed herein; Qu teaches in ¶ 0025, calculating the conditional revenue to be earned if a customer were to adopt a given product, which may be referred to as a conditional product revenue. Qu teaches in ¶ 0034, a first model predicting the propensity of a prospective customer; Qu teaches in ¶ 0046, where the model trainer is implemented using one or more processing circuits executing instructions stored in one or more memory circuits, where the instructions configure the processing circuits to perform as special purpose devices to perform operations; Qu teaches in ¶ 0131, The customer may be a live customer of the service platform, and the memory may further store instructions that, when executed by the processor, cause the processor to collect platform usage data from the customer based on interactions between the customer and the service platform.
a user adoption propensity of the one or more products determined by supplying the feature vector to a first trained machine learning model comprised in the plurality of machine learning models; Qu teaches in ¶ 0029, in some embodiments, a model trainer 160 is configured to retrieve the usage data in the customer data store 150 and to train or update the trained models 110 based on the usage data in the customer data store 150. In some embodiments, the model trainer periodically retrains the trained models 110 based on updates to the usage data in the customer data store 150 and/or due to changes in the product offerings (e.g., addition or removal of product offerings); Qu teaches in ¶ 0072, depict some embodiments of methods for training and implementing statistical models to compute the propensity of a prospective customer or a newly onboarded customer to be upsold on various products offered by a service provider based on training data from the product adoption behavior of other customers of the service provider and based on platform usage data;
a usage of the one or more products by the user determined by supplying the feature vector to a second trained machine learning model comprised in the plurality of machine learning models; Qu teaches in ¶ 0023, An additional neural network(s) may be trained to receive as input various information pertaining to merchant services that are available for purchase and use by the merchants and provides as output a numerical representation for each merchant service. The inputs to the trained neural network may be transformed versions (may be referred to as vectorized versions) of any type of available data descriptive of the merchant services. For example, there may be a text-based description of a given merchant service available on the World Wide Web; Qu teaches in ¶ 0043, The conditional expected product revenue values may be computed by a second statistical model, the second statistical model trained based on historical data associating customer feature embeddings to customer-level revenue for the corresponding products offered by the service platform; Qu teaches in ¶ 0124, a second statistical model;
overall yield growth of the user determined by supplying the feature vector; Qu teaches in ¶ 0034, compute the conditional product revenue associated with a customer. These models 110B may then be used to generate predictions 131 based on the available information 121 for a given customer (e.g., prospective customer 104, newly onboarded customer 105, or live customer 106), where these predictions may include product propensity predictions 132, product revenue predictions 133, and product and bundle recommendations 134. Qu teaches in 0041, configured to generate a customer feature embedding 216 (or feature vector) of the pre-processed text (e.g., a representation of the text as a vector of numbers) in an embedding space (e.g., or latent space or latent feature space, where similar customers have similar customer feature embeddings);
to a third trained machine learning model comprised in the plurality of machine learning models; Qu teaches in ¶ 0053, merchant profile data of type (3) above may be referred to as converted merchant services that can be separated into two different categories of conversions. The first category may be merchant services developed by payment processing system 108 and may be referred to as service provider's merchant services. Information on previous conversions of service provider's merchant services may be collected by element 218 in Fig. 2. The second category of converted merchant services may be third-party merchant services converted by a given merchant in the past and may be referred to as previous third-party merchant services conversion. Information on previous third-party merchant services conversion may be collected by element 220 in Fig. 2. Similar to categorical features 212-1 through 212-N, instances of transformer logics 222-1 and 222-2 may be applied to information at elements 218 and 220. The output of transformer logics 222-1 and 222-2 may then be fed as input into second trained neural network 224 and third trained neural network 226 (e.g., convolutional networks), respectively. Alternatively, outputs of transformer logics 222-1 and 222-2 may be fed into a single instance of a trained neural network as opposed to two separate neural networks 224 and 226;
Qu teaches production usage data, product adoption, expected revenue, products actively used by customers, a first model, usage activity of the platform, second statistical model, and feature vector, and Qu and Fotso are similar where Qu and Fotso teach feature vectors for user and customer data and adoption of merchant and customer products and services and Fotso further teaches the following:
A method comprising: determining a feature vector representing a user of a platform, the platform providing access to one or more products from a provider among a plurality of available products; Fotso teaches in ¶ 0037, merchant specific profile data may be collected by payment processing system 108 through monitoring and retrieving various types of information about products and services offered for sale by each merchant; Fotso teaches in ¶ 0039, additionally, AI-based merchant service recommendation module 112 may include one or more additional neural networks trained to receive, as input, merchant service data, and provide, as output, a numerical representation of each merchant service. The inputs to the trained neural network may be transformed versions (vectorized versions) of any type of available information on the merchant services. For example, there may be a text-based description of a given merchant service available on the World Wide Web 114. Natural language processing techniques may be utilized to transform the text-based description into a numerical representation to be provided as input to the trained neural network. Similarly, there may be various other forms of description for a given merchant service such as images, video and/or audio based descriptions (e.g., promotional videos and audio descriptions) for a given merchant service. Complex transformational techniques may be utilized to transform these forms of description of a merchant service into a numerical representation to be provided as input into the trained neural network. Using the vectorized version of the merchant service data, the trained neural network(s) can provide a numerical representation of each merchant service as output; Fotso teaches in ¶ 0014, merchant platforms;
and computing a retention of the user determined by based on supplying the feature vector to a fourth trained machine learning model comprised in the plurality of machine learning models; Fotso teaches in ¶ 0054, The output of feed-forward neural network 210, vectorized categorical features 216-1 to 216-N, and outputs of convolutional neural networks 224 and 226 may each be an array of numbers, and concatenation logic 228 may concatenate these outputs; Fotso teaches in ¶ 0055, data ran through a fourth trained neural network; Fotso teaches in ¶ 0088, the output of trained neural network is specific merchant services for merchant that includes loyalty (under sales and marketing category), where loyalty is likened to retention;
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to combine models for predicting customer behavior, including the use or adoption of products by current customers and prospective customers of a service platform offering multiple service products of Qu with a method that includes receiving merchant profile data describing attributes of merchants of Fotso to assist businesses in building trained neural networks with vectorized categorical features (Fotso Spec. ¶ 0054).
Qu teaches production usage data, product adoption, expected revenue, products actively used by customers, a first model, neural networks, usage activity of the platform, second statistical model, and feature vector, and Fotso teaches machine trained models, platforms for merchant services, merchant profile data; and Userpilot Team teaches adoption funnel, and Qu, Fotso, and Userpilot Team are similar where Qu, Fotso, and Userpilot teach user adoption; and Userpilot Team further teaches the following:
comprised in a multi-stage user adoption funnel; of corresponding stages of the multi-stage adoption funnel; Userpilot Team “Understanding the Adoption Funnel: Stages, Examples & More” https://userpilot.com/blog/adoption-funnel/ (January 16, 2023) hereinafter “Userpilot Team” teaches in Pg. 1, ¶¶ 3 – 8, user adoption funnel in stages;
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to combine models for predicting customer behavior, including the use or adoption of products by current customers and prospective customers of a service platform offering multiple service products of Qu and a method that includes receiving merchant profile data describing attributes of merchants of Fotso with using metrics to track adoption in stages of Userpilot Team to assist businesses with implementing a feature adoption funnel to navigate stages (Userpilot Team, Pg. 4, ¶ 1).
Claims 2, 10, and 16:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
wherein the first trained machine learning model is trained based on training data from historical interactions between a plurality of existing users and a plurality of existing platforms providing access to corresponding selections of the plurality of available products; Qu teaches in ¶ 0056, predicting the propensity of a customer to adopt services offered by a platform based on historical interactions with the platform (platform usage data); Qu teaches above in claim 1, a first model predicting the customer propensity; Qu teaches in ¶ 0125, compute a plurality of product propensities for the customer by supplying the customer feature embedding to a trained statistical model, the product propensities representing likelihoods that the customer will adopt corresponding products offered by a service platform.
Claims 3, 11, and 17:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
wherein the user is a prospective user of the platform, and the second trained machine learning model is a model that determines expected yield for the prospective user; Qu teaches in ¶ 0034, a second model predicting the propensity for a current customer; Qu further teaches in ¶ 0034, the predictions include product revenue predictions.
Claims 4, 12, and 18:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu teaches production usage data, product adoption, expected revenue, products actively used by customers, a first model, neural networks, usage activity of the platform, second statistical model, and feature vector, and Fotso teaches machine trained models, platforms for merchant services, merchant profile data; and Userpilot Team teaches adoption funnel, and Qu, Fotso, and Userpilot Team are similar where Qu, Fotso, and Userpilot teach user adoption; and Fotso further teaches the following:
wherein the feature vector is determined based on one or more of: visits to a product page on a website of the provider; Fotso teaches in ¶ 0049, merchant profile data can include information on the merchant's customers such as number of visits; Fotso teaches in ¶ 0087, the list of merchant service recommendations at block 358 and categorical flags vectorized at block 356 may be provided to a separate trained neural network. Such trained neural network may adjust the merchant services recommended based on the categorical flags. For example, the neural network may be trained over time and using merchant profile data to learn merchant behavior, trends, and needs to identify categories of merchant services that may be more or less relevant to a particular merchant;
total yield of the user; Fotso teaches in ¶ 0051, data related to the merchant's business-related activities including information related on the merchant’s revenue
time on platform of the user; Fotso teaches in ¶ 0049, amount of time spent researching and reading about specific merchant services;
and industry of the user; Fotso teaches in ¶ 0049, the merchant's categorical information such as MCC.
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to combine models for predicting customer behavior, including the use or adoption of products by current customers and prospective customers of a service platform offering multiple service products of Qu and using metrics to track adoption in stages of Userpilot Team with a method that includes receiving merchant profile data describing attributes of merchants of Fotso to assist businesses in building trained neural networks with vectorized categorical features (Fotso Spec. ¶ 0054).
Claim 5:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
wherein the platform is associated with a first industry, and the method further comprises determining a second estimated yield
Claim 6:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
determining a plurality of feature vectors representing a plurality of users of the platform; Qu teaches in ¶ 0132, compute a plurality of expected revenue values for the corresponding products offered by the service platform;
and determining an estimated overall yield for the plurality of users based on the plurality of feature vectors; Qu teaches in ¶ 0058, categorical data and numerical data representing various characteristics of the customer (e.g., current revenue, transaction volume, and the like); Qu teaches in ¶ 0059, user interactions with the service provider may be collected as features representing the current state of a customer. In various embodiments of the present disclosure, these collected features are processed (e.g., mathematically transformed, such as to a log scale and/or normalized) before being added to the customer feature embedding or feature vector representing the customer.
Claim 7:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
determining an estimated overall yield based on the plurality of machine learning models; Qu teaches in Fig. 1B is a block diagram depicting the use of available information at different stages of a customer or prospective customer relationship with an organization to predict product propensity and conditional revenue and to generate marketing and product promotion messages to customers; Qu teaches in ¶ 0051, In a case where the model training is determined to be complete, then the trained model (e.g., the trained parameters) are output by the model trainer 160, and the trained model may be included in the trained models 110 shown in Fig. 1A and Fig. 1B (e.g., after validating the model using validation data taken from the customer data store 150).
ranking the plurality of available products based on the estimated overall yield; Qu teaches in ¶ 0091, product filter 513 filters the available products (e.g., the products offered by the service provider) to identify targeted products based on the prospective customer product propensities. The product filter sorts the products based on their corresponding propensities and selects some number of highest ranking products; Qu teaches in ¶ 0094, Fig. 5, the marketing and promotion system may receive live customer product propensities and/or expected revenue (e.g., the marketing and promotion system may also, or instead, receive expected revenue calculations for the customer for a plurality of products). The live customer product propensities and/or the expected revenue calculations may be supplied to a product filter to identify targeted products (e.g., products having high propensity and/or high expected revenue).
and displaying the ranking of the plurality of available products on a user interface; Qu teaches in ¶ 0026, displayed on a user interface of a computing system.
Claim 9:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
wherein the memory further stores instructions that, when executed by the processor, cause the processor to: generate a ranking of the plurality of available products based on and overall estimated yield for each of the plurality of available products, the overall estimated yield based on the user adoption propensity, the usage, the overall yield growth, and the retention of the user; Qu teaches in ¶ 0035, The processing circuits are configured to perform operations according to various embodiments of the present disclosure using program instructions that may be stored in one or more memory circuits (e.g., the same memory circuits that store the input data, output data, and intermediate results, or different memory circuits); Qu teaches in ¶ 0037, the system 210 is configured to compute, for a given customer, a plurality of propensities 212 corresponding to each of the products 211. Each of the propensities 212 represents a degree of product fit between the product and the given customer (e.g., a likelihood, probability, or other numerical metric). These propensities 212 may therefore be used to evaluate the likelihood that the given customer will adopt the corresponding product (e.g., become a subscriber or user of that corresponding product); Qu teaches in ¶ 0052, training and implementing statistical models to compute the propensity of a prospective customer or a newly onboarded customer to adopt various products offered by a service provider based on training data from the product adoption behavior of other customers of the service provider; Qu teaches in ¶ 0073, some aspects of embodiments of the present disclosure relate to computing estimates of product value or product revenue conditional on the user going live. Multiplying the conditional estimates by the product propensity scores produce the desired expected product value from users going live. (For example, a customer going live on an expensive product may produce a large amount of revenue, but if the product is a poor fit for the customer needs, the customer is unlikely to subscribe to the product and therefore marketing efforts related to that product may be wasted if the customer propensity score for that product is low); Qu teaches above from ¶ 0091, the product filter sorts the products based on their corresponding propensities and selects some number of highest ranking products;
Claim 13:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
wherein the instructions to compute the estimated yield a further comprise instructions that, when executed by the processor, cause the processor to compute a platform adoption propensity for the one or more products 1B to compute the propensity of live customers (current users) to be upsold on products (e.g., begin to use products that they had not previously been using); Qu teaches in ¶ 0128, the memory may further store instructions that, when executed by the processor, cause the processor to retrieve the one or more text descriptions from one or more of: a website associated with the customer; a third-party data source of company information; and a publication regarding the customer.
Claim 15:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
wherein the feature vector is determined based on text information and non-text information; Qu teaches in ¶ 0058, collected information may also include non-textual data; Qu teaches in ¶ 0066, customer data includes information corresponding to the inputs that are to be supplied to the model (e.g., textual data collected from scraping customer websites, from third parties, from published information, and from customer responses to questions during the sign-up process) and features extracted from customer usage data as discussed above (e.g., reaching milestones, application usage information, product activity, interactions with the service provider, and the like).
Claim 19:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
The non-transitory computer-readable medium of claim 14, further storing instructions that, when executed by the processor, cause the processor to generate a report of a ranking of the one or more products; Qu teaches in ¶ 0074, calculation of per-user rankings of not-yet-adopted products by expected value.
Claim 20:
Qu, Fotso, and Userpilot Team teach claims 1, 8, and 14. Qu further teaches the following:
The non-transitory computer-readable medium of claim 19, further storing instructions that, when executed by the processor, cause the processor to transmit a promotion of the one or more products to the platform, the one or more products product being selected from the available products based on their ranking of the products; Qu teaches in ¶ 0092, the marketing and promotion system generates targeted product marketing messages for the prospective customer based on their accessibility via the corresponding prospective customer contact channels. Qu further teaches in ¶ 0094, the marketing and promotion system may receive live customer product propensities and/or expected revenue (e.g., the marketing and promotion system may also, or instead, receive expected revenue calculations for the customer for a plurality of products). The live customer product propensities and/or the expected revenue calculations may be supplied to a product filter to identify targeted products (e.g., products having high propensity and/or high expected revenue).
Conclusion
The prior art made of record and not relied upon is considered relevant but not applied:
Note: these are additional references found but not used.
- Reference Tilton, Scott et al. (U.S. Publication No. 2019/0311268) discloses promotion value model uses deep neural networks to learn to calculate the promotion value of a commercial brand.
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 Frank Alston whose telephone number is 703-756-4510. The Examiner can normally be reached 9:00 AM – 5:00 PM Monday - Friday. Examiner can be reached via Fax at 571-483-7338. 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.
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.
/FRANK MAURICE ALSTON/
Examiner, Art Unit 3625
6/02/2026
/BETH V BOSWELL/Supervisory Patent Examiner, Art Unit 3625
Prosecution Insights