DETAILED ACTION
This is a Final Office action is in response to communications filed on May 18th, 2026. Claim 1-2, 4, 7-11, 13, and 16-20 is/are amended. Claims 1-20 have been examined in this application.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 112 (First Paragraph)
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
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.
Claims 1-20 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 pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention. The claims were amended to include the following limitations:
Independent Claims 1, 10, and 19:
“identifying one or more featured products based on the user query, wherein the one or more featured products comprises an organic search result and a sponsored search result”
“wherein the query response comprises an order of presentation of the organic search result and the sponsored search result indifferent positions of a scrollable product carousel in a graphical user interface (GUI) of the client device”
“causing the client device to display the GUI by rendering the organic search result and the sponsored search result in respective positions of the scrollable product carousel, the respective positions corresponding to the order of presentation”
“fine-tuning the machine-learned generative language model based on the data on user interactions with the query response to improve a prediction of click-through rate (CTR) by the machine-learned generative language model for the one or more featured products, wherein fine-tuning the machine-learned generative language model comprises: adjusting, by the machine-learned generative language model, the order of presentation of the organic search result and the sponsored search result in the scrollable product carousel”
Applicant's original disclosure suggests that an online system may generate a prompt for a machine-learned generative language model specifying a request related to a user query and a request to suggest one or more featured products, and that the prompt may include a list of featured products or data describing the list of featured products indexed into an index document generated based on LlamaIndex™ or LangChain™. The disclosure further suggests that the one or more featured products may be sponsored products, and that the one or more featured products are selected based on a bidding process in which a bidder indicates a bid amount for a particular featured product. The disclosure also suggests that a response from the model may be parsed to extract one or more consumer packaged good (CPG) products, that the extracted CPG products may be matched with featured products in a database, that relevant sponsored content items may be retrieved from a sponsored content server, that each relevant sponsored content item has a predicted click-through rate, and that the relevant sponsored content items may go through an auction process to identify a highest bidder.
The disclosure additionally suggests that content items may be presented in different views, such as list view, grid view, and carousel view, and that in a carousel view items are presented in one or more carousels horizontally or vertically through which users may scroll. The disclosure suggests that a carousel may be generated for each CPG product, each carousel showing the sponsored content associated with featured products corresponding to that particular CPG product, and that both sponsored and organic search results are mixed within these carousels. The disclosure further suggests that data on user interactions may be collected and used to retrain and/or fine-tune a machine-learned model that predicts the click-through rate (CTR) for each product, or, in the alternative, used to retrain and/or fine-tune the machine-learned language model. Thus, the disclosure may support prompt generation specifying featured products, index documents resolving prompt size limitations, bidding and auction processes for selecting sponsored content items, parsing of model responses to extract CPG products, matching of extracted CPG products to featured products in a database, presentation of content items in list, grid, or carousel views, carousels in which sponsored and organic search results are mixed, and retraining of either a click-through-rate prediction model or the machine-learned language model. However, the disclosure does not reasonably convey possession of one or more featured products that comprise an organic search result, nor does it describe an order of presentation of an organic search result and a sponsored search result in different positions of a scrollable product carousel.
Further, the disclosure does not describe rendering an organic search result and a sponsored search result in respective positions of a scrollable product carousel corresponding to an order of presentation. The disclosure also does not describe fine-tuning the machine-learned generative language model to improve a prediction of click-through rate by the machine-learned generative language model, or fine-tuning that comprises the machine-learned generative language model adjusting an order of presentation of an organic search result and a sponsored search result in a scrollable product carousel. The original disclosure is instead directed to an online system that monetizes responses generated by a machine-learned generative language model by injecting featured or sponsored products into those responses. For example, the disclosure describes generating a prompt requesting the model to suggest featured products or to include CPG products in a response, providing the prompt to a model serving system, parsing the returned response to extract CPG products, matching the extracted CPG products against featured products stored in a data store, retrieving associated sponsored content items, determining a predicted click-through rate for each sponsored content item, conducting an auction to identify a highest bidder among the sponsored content items, conducting a search over the extracted CPG products, and presenting the search results together with the winning sponsored content items. The disclosure states only that both sponsored and organic search results are "mixed within these carousels," and further states that the search results may be presented in grid view at a first portion of the user interface while sponsored content items are presented in carousels above, below, or between the grids. These teachings do not explain how the claimed system determines an order of presentation of an organic search result relative to a sponsored search result, how respective positions within a carousel are assigned, or how the machine-learned generative language model performs the ordering function recited in claims 1, 10, and 19.
Thus the disclosure, however, never suggests how the one or more featured products comprise an organic search result, given that the disclosure describes featured products as products for which bidders submit bid amounts and describes organic search results only as the output of a search conducted over CPG products extracted from a response already generated by the model; how a query response comprises an order of presentation of an organic search result and a sponsored search result in different positions of a scrollable product carousel; how the client device renders those results in respective positions corresponding to that order of presentation; how fine-tuning improves a prediction of click-through rate by the machine-learned generative language model, given that the disclosure attributes click-through-rate prediction to a separate machine-learned model recited in the alternative; and how the machine-learned generative language model adjusts the order of presentation within the scrollable product carousel, given that the disclosure assigns all presentation processing to the online system rather than to the model. If the original disclosure does describe these functions, Applicant is encouraged to identify the portion of the original disclosure describing these functions.
When examining computer-implemented functional claims, examiners should determine whether the specification discloses the computer and the algorithm (e.g., the necessary steps and/or flowcharts) that perform the claimed function in sufficient detail such that one of ordinary skill in the art can reasonably conclude that the inventor possessed the claimed subject matter at the time of filing. It is not enough that one skilled in the art could write a program to achieve the claimed function because the specification must explain how the inventor intends to achieve the claimed function to satisfy the written description requirement. See, e.g., Hyatt v. Dudas, 492 F.3d 1365, 1370, n.4 (Fed. Cir. 2007) (citing MPEP § 2163.04). Also, See, e.g., Vasudevan Software, Inc. v. MicroStrategy, Inc., 782 F.3d 671, 681-683, 114 USPQ2d 1349, 1356, 1357 (Fed. Cir. 2015) (reversing and remanding the district court’s grant of summary judgment of invalidity for lack of adequate written description where there were genuine issues of material fact regarding "whether the specification show[ed] possession by the inventor of how accessing disparate databases is achieved"). If the specification does not provide a disclosure of the computer and algorithm in sufficient detail to demonstrate to one of ordinary skill in the art that the inventor possessed the invention a rejection under 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph, for lack of written description must be made. See MPEP § 2161.01(I).
Therefore, the claims are rejected under 112 1st paragraph as failing to comply with the written description requirement as applicant is only entitled to claim the invention the Applicant possessed and disclosed at the time of invention. Hence, independent claims 1, 10, and 19, its dependent claims 2-9, 11-18, and 20 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.
Claim Rejections - 35 USC § 112 (Second Paragraph)
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 1-20 is/are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
Claims 1, 10, and 19 recite the phrase “wherein fine-tuning the machine-learned generative language model comprises: adjusting, by the machine-learned generative language model, the order of presentation of the organic search result and the sponsored search result in the scrollable product carousel.” Here a training operation is recited as comprising a display operation. Applicant's published specification [0081, 0182, 0083, 0085], defines training as the modification of model parameters, stating that the machine learning training module “generates the set of parameters for a machine learning model by 'training' the machine learning model” by “updating weights associated for the machine learning model through a back-propagation process,” and may “apply gradient descent to update the set of parameters,” and further that fine-tuning consists of obtaining “a pre-trained transformer language model and further fine tune the parameters of the transformer model.” Adjusting an order of presentation within a graphical user interface element is not a modification of model parameters. It therefore cannot be determined what act constitutes performance of the recited fine-tuning step, whether that step occurs during training or during rendering of a query response, or when infringement of the step would occur. The claim is amenable to at least two plausible constructions and is indefinite. See MPEP § 2173.02.
Further, the independent Claims 1, 10, and 19 recite “wherein the one or more featured products comprises an organic search result and a sponsored search result.” Here a product is recited as comprising a search result. Applicant's published specification [0027], defines an item as “a good or product that can be provided to the customer through the online system 140,” whereas a search result is a record returned in response to a query. It cannot be determined whether the recited featured products are goods, records identifying goods, or both. Further, the specification [0126], describes featured products as products for which bidders submit bid amounts, stating that “a bid amount is a value indicating a priority for having a corresponding featured product to be suggested to the user,” while an organic search result is not the subject of a bid. Whether a product subject to a bid may simultaneously constitute an organic search result cannot be determined.
As such, the claims is/are indefinite under 35 U.S.C. § 112, second paragraph, as it fails to distinctly point out and particularly claim the subject matter which the inventor regards as the invention.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e. an abstract idea) without significantly more.
Step 1: Claims 1-9 is/are drawn to method (i.e., a process), and claims 10-18 is/are drawn to computer readable medium (i.e., a manufacture), and claims 19-20 is/are drawn to system (i.e., a manufacture). (Step 1: YES).
Step 2A - Prong One: In prong one of step 2A, the claim(s) is/are analyzed to evaluate whether it/they recite(s) a judicial exception.
Claim 1: A method comprising:
receiving, from a client device and via an interface, a user query;
identifying one or more featured products based on the user query, wherein the one or more featured products comprises an organic search result and a sponsored search result;
generating a prompt for input to a machine-learned generative language model, the prompt specifying at least a request related to the user query and a request to suggest the one or more featured products in association with a response to the prompt;
providing the prompt to a model serving system for execution by the machine-learned generative language model;
receiving, from the model serving system, a response generated by executing the machine-learned generative language model on the prompt, the response including at least one of the one or more featured products;
generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products;
transmitting instructions, to the client device, to cause presentation of the query response to the user query, wherein the query response comprises an order of presentation of the organic search result and the sponsored search result indifferent positions of a scrollable product carousel in a graphical user interface (GUI) of the client device;
causing the client device to display the GUI by rendering the organic search result and the sponsored search result in respective positions of the scrollable product carousel, the respective positions corresponding to the order of presentation;
receiving and collecting data on user interactions with the query response;
and fine-tuning the machine-learned generative language model based on the data on user interactions with the query response to improve a prediction of click-through rate (CTR) by the machine-learned generative language model for the one or more featured products, wherein fine-tuning the machine-learned generative language model comprises:
adjusting, by the machine-learned generative language model, the order of presentation of the organic search result and the sponsored search result in the scrollable product carousel.
(Examiner notes: The underlined claim terms above are interpreted as additional elements beyond the abstract idea and are further analyzed under Step 2A - Prong Two)
Under their broadest reasonable interpretation, independent claims 1, 10, and 19 recite a commercial product-recommendation and advertising practice comprising receiving a user request, identifying products responsive to the request, including organic and sponsored product results, requesting and obtaining generated content suggesting one or more products, presenting the organic and sponsored product recommendations to the user in selected positions, collecting information regarding the user’s interactions with the recommendations, and using the interaction information and predicted click-through rate to modify the order in which the products are presented. The recited limitations therefore describe selecting, presenting, and optimizing product promotions based on a user’s request and subsequent engagement. This constitutes a commercial interaction involving advertising, marketing, and sales activities or behaviors and therefore falls within the “certain methods of organizing human activity” grouping of abstract ideas. The recited limitations therefore describe selecting, presenting, and optimizing product promotions based on a user’s request and subsequent engagement. This constitutes a commercial interaction involving advertising, marketing, and sales activities or behaviors. Advertising, marketing, and sales activities fall within the “certain methods of organizing human activity” grouping of abstract ideas identified in MPEP § 2106.04(a)(2)(II). More particularly, the independent claims recite identifying organic and sponsored product results, suggesting the products in connection with a response to a user query, presenting those results in an ordered product carousel, monitoring user engagement with the displayed products, and adjusting product placement to improve predicted CTR. These limitations concern which products are promoted to a potential customer, where the promoted products are placed, and how later product placement is optimized based on customer behavior. The amended claim language expressly recites the organic and sponsored results, carousel ordering, interaction-data collection, CTR prediction, and subsequent adjustment of the presentation order. The commercial and advertising character of the claimed process is also consistent with the applicant’s description. Paragraph [0090] explains that an LLM-generated response may include a response to the user’s request together with a suggestion for one or more featured products, that the featured products may be sponsored products, and that the product suggestions may be intertwined with the response to make the advertising opportunity less intrusive or disruptive. Thus, the specification identifies the integration of sponsored product suggestions into generated responses as an advertising opportunity. The Examiner notes that although the claim limitations are summarized, the analysis regarding subject matter eligibility considers the entirety of the claim and all of the claim elements individually, as a whole, and in ordered combination.
The dependent claims further elaborate on aspects of the same abstract idea by adding features relating to ranking, bidding, auction-based selection, hyperlink presentation, recipe generation, multiple-user handling, and catalog mapping. Claims 2, 11, and 20 recite generating relevance scores for candidate products and selecting products having scores above a threshold. These limitations further specify evaluating and prioritizing products for presentation to a potential customer and therefore refine the underlying marketing and product-recommendation activity. Claims 3 and 12 recite receiving bid values from product providers and selecting products based on bid values above a threshold. Claims 8 and 17 further recite performing an auction to select a product. These limitations describe competitive bidding and paid product placement, which are commercial interactions and fundamental economic practices associated with advertising and sales. Claims 4 and 13 recite presenting recommended products in connection with a recipe page. Claims 5 and 14 recite incorporating the product suggestion into textual content. These limitations merely specify the commercial context and informational format in which the product recommendations are presented. Claims 6 and 15 recite receiving another user query and generating another response that includes CPG products. Claims 7 and 16 recite mapping the CPG products to products in an online catalog and presenting the mapped products to the user. These limitations further specify matching a potential customer’s expressed interest to products available for sale through an online commercial system. Claims 9 and 18 recite hyperlinks that allow the user to interact directly with the recommended products from within the response. These limitations facilitate customer access to, and interaction with, promoted products and therefore further implement the same advertising and sales activity. Accordingly, the dependent claims do not introduce a new technological concept but instead add further refinements to the underlying abstract idea of product recommendation, ranking, and advertising optimization based on user input and interaction data. Accordingly, claims 1-20 are directed to an abstract idea under 35 U.S.C. §101, using conventional communication and recordkeeping techniques. As such, the claims are directed to an abstract idea involving certain methods of organizing human activity and mental processes, which falls within a judicial exception under 35 U.S.C. §101.
Independent claim(s) 10 and 19 recite/describe nearly identical steps (and therefore also recite limitations that fall within this subject matter grouping of abstract ideas), and this/these claim(s) is/are therefore determined to recite an abstract idea under the same analysis.
As such, the Examiner concludes that claims 1 recites an abstract idea (Step 2A – Prong One: YES).
Step 2A - Prong Two: In prong two of step 2A, an evaluation is made whether a claim recites any additional element, or combination of additional elements, that integrate the exception into a practical application of that exception. An “addition element” is an element that is recited in the claim in addition to (beyond) the judicial exception (i.e., an element/limitation that sets forth an abstract idea is not an additional element). The phrase “integration into a practical application” is defined as requiring an additional element or a combination of additional elements in the claim to apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that it is more than a drafting effort designed to monopolize the exception.
The requirement to execute the claimed steps/functions using a client device and via an interface, machine-learned generative language model, model serving system, processors, graphical user interface (GUI), etc. (Claims 1, 10, and 19) is/are equivalent to adding the words “apply it” on a generic computer and/or mere instructions to implement the abstract idea on a generic computer.
Similarly, the limitations of using a client device and via an interface, machine-learned generative language model, model serving system, processors, graphical user interface (GUI), etc. (Claims 1, 10, and 19, and dependent claims 2-9, 11-18, and 20) are recited at a high level of generality and amount to no more than mere instructions to apply the exception using generic computer components. This/these limitation(s) do/does not impose any meaningful limits on practicing the abstract idea, and therefore do/does not integrate the abstract idea into a practical application (see MPEP 2106.05(f)).
Further, the additional limitations beyond the abstract idea identified above, serves merely to generally link the use of the judicial exception to a particular technological environment or field of use. Specifically, it/they serve(s) to limit the application of the abstract idea to computerized environments (e.g., receive, identify, generate, provide, transmit, collect, fine-tune, etc. steps performed by a client device and via an interface, machine-learned generative language model, model serving system, processors, graphical user interface (GUI), etc.). This reasoning was demonstrated in Intellectual Ventures I LLC v. Capital One Bank (Fed. Cir. 2015), where the court determined "an abstract idea does not become nonabstract by limiting the invention to a particular field of use or technological environment, such as the Internet [or] a computer"). This/these limitation(s) do/does not impose any meaningful limits on practicing the abstract idea, and therefore do/does not integrate the abstract idea into a practical application (see MPEP 2106.05(h)).
The recited additional element(s) of receiving the user query, providing a generated prompt to a model-serving system, receiving the generated response, transmitting instructions for displaying the response, rendering the organic and sponsored product results in respective positions of a scrollable product carousel, and collecting data regarding user interactions with the displayed recommendations merely gather, prepare, communicate, present, and collect the information used to perform the abstract commercial product-recommendation and advertising process. The user query supplies the information used to identify and recommend products, while the prompt and generated response merely place that information into forms suitable for producing and communicating the product recommendations. The transmission and carousel-rendering limitations merely present the results of the abstract recommendation and advertising process to the user in a particular informational format. The collection of clicks, interaction data, and predicted click-through-rate information merely gathers statistics concerning how users respond to the presented organic and sponsored product recommendations. To the extent these receiving, prompting, model-serving, transmitting, rendering, and interaction-collection limitations are considered additional elements rather than part of the judicial exception, they merely append insignificant pre-solution and post-solution activity to the abstract idea, including collecting the information necessary to select products, presenting the resulting recommendations, and gathering customer response information for subsequent evaluation. Similarly, fine-tuning the machine-learned generative language model based on the collected interaction data merely uses the results of the claimed commercial evaluation to refine the same product-recommendation, advertising, and product ordering process for future iterations and does not apply those results to produce a separate technological result or improve the operation of the computer or model itself. The received query, generated prompt and response, displayed carousel, interaction data, and CTR information merely supply the inputs, outputs, and evaluation criteria used to select, arrange, and optimize product advertisements (Independent Claims 1, 10, and 19, additionally and/or alternatively simply append insignificant extra-solution activity to the judicial exception, (e.g., mere pre-solution activity, such as data gathering, in conjunction with an abstract idea)). This/these limitation(s) do/does not impose any meaningful limits on practicing the abstract idea, and therefore do/does not integrate the abstract idea into a practical application. (See MPEP 2106.05(g)).
Dependent claims 2-9, 11-18, and 20 fail to include any additional elements. In other words, each of the limitations/elements recited in respective dependent claims is/are further part of the abstract idea as identified by the Examiner for each respective dependent claim (i.e., they are part of the abstract idea recited in each respective claim).
The Examiner has therefore determined that the additional elements, or combination of additional elements, do not integrate the abstract idea into a practical application. Accordingly, the claim(s) is/are directed to an abstract idea (Step 2A – Prong two: NO).
Step 2B: In step 2B, the claims are analyzed to determine whether any additional element, or combination of additional elements, is/are sufficient to ensure that the claims amount to significantly more than the judicial exception. This analysis is also termed a search for an "inventive concept." An "inventive concept" is furnished by an element or combination of elements that is recited in the claim in addition to (beyond) the judicial exception, and is sufficient to ensure that the claim as a whole amounts to significantly more than the judicial exception itself. Alice Corp., 134 S. Ct. at 2355, 110 USPQ2d at 1981 (citing Mayo, 566 U.S. at 72-73, 101 USPQ2d at 1966).
As discussed above in “Step 2A – Prong 2”, the identified additional elements in independent Claims 1, 10, and 19, and dependent claims 2-9, 11-18, and 20 are equivalent to adding the words “apply it” on a generic computer, and/or generally link the use of the judicial exception to a particular technological environment or field of use. Therefore, the claims as a whole do not amount to significantly more than the judicial exception itself.
The recited additional element(s) of a client device, an interface, a model-serving system, a machine-learned generative language model, a graphical user interface, and a scrollable product carousel, considered individually and in combination, do not amount to significantly more than the judicial exception. The recited steps of receiving a user query, generating and transmitting a prompt, providing the prompt to the model-serving system, receiving model-generated information, transmitting display instructions, rendering organic and sponsored product results, and collecting user-interaction data merely receive, transmit, process, present, and collect the information used to implement the abstract commercial product-recommendation and advertising practice. The model-serving system and machine-learned generative language model are recited at a high level of generality and perform their ordinary functions of receiving a prompt, executing a model on the prompt, and returning generated content (Independent Claims 1, 10, and 19), additionally and/or alternatively simply append insignificant extra-solution activity to the judicial exception, (e.g., mere pre-solution activity, such as data gathering, in conjunction with an abstract idea), i.e. receiving the user query, providing a generated prompt to a model-serving system, receiving the generated response, transmitting instructions for displaying the response, rendering the organic and sponsored product results in respective positions of a scrollable product carousel, and collecting data regarding user interactions with the displayed recommendations merely gather, prepare, communicate, present, and collect the information, which is similar to “Receiving or transmitting data over a network, e.g., using the Internet to gather data”, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information), “Storing and retrieving information in memory”, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93; “Presenting offers to potential customers and gathering statistics generated based on the testing about how potential customers responded to the offers; the statistics are then used to calculate an optimized price”, OIP Technologies, 788 F.3d at 1363, 115 USPQ2d at 1092-93, Determining an estimated outcome and setting a price, OIP Techs., 788 F.3d at 1362-63, 115 USPQ2d at 1092-93, is a well-understood, routine, and conventional function when it is claimed in a merely generic manner (as it is here) (See MPEP 2106.05(d) (II)).
This conclusion is based on a factual determination. Applicant’s own disclosure at paragraph [0071-0072] acknowledges that “the recommendations are in the form of a list of potential recipes the ingredients can fulfill, and a list of additional ingredients to fulfill each recipe. The content presentation module 210 may present each suggested recipe and the list of additional ingredients for fulfilling the recipe to the customer. The content presentation module 210 may allow the customer to automatically place one or more additional ingredients in the basket of the customer …The order management module 220 that manages orders for items from customers. The order management module 220 receives orders from a customer client device 100 and offers the orders to pickers for service based on picker data. For example, the order management module 220 offers an order to a picker based on the picker’s location and the location of the retailer from which the ordered items are to be collected” This additional element therefore do not ensure the claim amounts to significantly more than the abstract idea.
Viewing the additional limitations in combination also shows that they fail to ensure the claims amount to significantly more than the abstract idea. When considered as an ordered combination, the additional components of the claims add nothing that is not already present when considered separately, and thus simply append the abstract idea with words equivalent to “apply it” on a generic computer and/or mere instructions to implement the abstract idea on a generic computer or/and append the abstract idea with insignificant extra solution activity associated with the implementation of the judicial exception, (e.g., mere data gathering, post-solution activity) and/or simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception.
The dependent claims 2-9, 11-18, and 20 fail to include any additional elements. In other words, each of the limitations/elements recited in respective independent claims is/are further part of the abstract idea as identified by the Examiner for each respective dependent claim (i.e., they are part of the abstract idea recited in each respective claim).
Specifically, claims 2, 11, and 20 recite generating relevance scores and selecting products based on a threshold. Assigning scores to items, ranking items according to those scores, and applying thresholds to filter results were conventional techniques in information retrieval systems, search engines, and recommendation platforms. Such ranking and filtering mechanisms merely apply conventional mathematical evaluation and comparison techniques to implement the abstract idea of recommending products. Claims 3, 8, 12, and 17 recite receiving bid values from product providers and performing an auction process to select products. Competitive bidding, sponsored search ranking, and auction-based advertisement placement were widely used commercial practices implemented on generic computer systems prior to the effective filing date. Selecting items for display based on bid amounts and competitive ranking constitutes a conventional monetization mechanism in online advertising and does not add a technological improvement to computer functionality. Claims 4, 5, 6, 13, 14, and 18 recite generating recipe pages, producing textual content incorporating product suggestions, creating hyperlinks, and formatting output for display. Presenting information in webpage form, embedding hyperlinks, and generating textual recommendation content are routine output presentation techniques performed by generic web servers and client devices. These limitations merely specify the format in which the abstract idea is communicated to the user and do not provide a technical improvement. Claims 7, 15, and 16 recite handling multiple user queries and mapping products to a catalog of an online system. Managing multiple users, associating items with entries in a product catalog, and filtering results based on stored user preferences were conventional database and e-commerce operations performed using generic computing infrastructure. When considered as an ordered combination, the dependent claims simply apply conventional scoring, bidding, ranking, catalog matching, and presentation techniques to implement the abstract idea of generating and refining product recommendations. The claims do not recite any specialized hardware, unconventional data structure, specific machine-learning architecture, or technical improvement to computer performance. Instead, they rely on generic computing components performing their expected functions. Accordingly, the additional limitations of the dependent claims do not amount to significantly more than the abstract idea and therefore fail to provide an inventive concept under Step 2B. Because these elements do not solve a specific technical problem or offer a technical improvement over existing systems, they are viewed as merely "applying" the abstract idea on a generic computer, thus failing to provide a practical application that would render the claims patent-eligible, and therefore do not add an inventive concept sufficient to transform the abstract idea into patent-eligible subject matter.
When viewed as an ordered combination, the additional elements of claims 2-9, 11-18, and 20 merely instruct to implement the abstract idea using generic computer components to collect, store, represent, and display information. The claims do not recite any unconventional arrangement of elements, nor do they effect an improvement to computer functionality or another technical field and therefore fail to integrate the abstract concept into a practical application and it is recited at a high level of generality and does not integrate the judicial exception into a practical application.
The Examiner has therefore determined that no additional element, or combination of additional claims elements is/are sufficient to ensure the claim(s) amount to significantly more than the abstract idea identified above (Step 2B: NO).
Therefore, claims 1-20 are not eligible subject matter under 35 USC 101.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis 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 of this title, 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.
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 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 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 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, 5-7, 9-10, 14-16, and 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pub. 20240331004 (“Shitrit”) in view of U.S. Pub. 20220277345 (“Nag”) in further view of U.S. Pub. 20240249318 (“Spiegel”).
As per claims 1, 10, and 19, Shitrit discloses, receiving, from a client device and via an interface, a user query (Examiner interprets that Shitrit discloses, a user enters a question or search query into input text bar 112 of UI 100, and backend server computers receive question 314 from client UI 302) (0013-0016, 0032-0033);
identifying one or more featured products based on the user query (Examiner interprets that the Shitrit identifies candidate product-related noun phrases from an answer generated in response to the user’s question, scores and ranks the noun phrases based on the question-answer context, and uses the ranked noun phrases to identify related products in products database 362) (0012, 0016-0018, 0032-0033),
generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products (Examiner interprets Shitrit enhances the generated answer with associated product recommendations, links product-related phrases in the answer to corresponding products, and generates the enhanced answer for presentation to the user) (0013-0016, 0032-0033);
transmitting instructions, to the client device, to cause presentation of the query response to the user query (Examiner interprets UI generator 372 generates an answer containing linked product results, and the enhanced answer is transmitted through network 310 for display by client UI 302) (0016, 0032-0033),
Shitrit discloses, receiving a natural-language question, generating a responsive answer, extracting and ranking product-related noun phrases from the answer, finding corresponding products in a product database, and returning the product-enhanced answer to the client interface, but specifically doesn’t disclose, wherein the one or more featured products comprises an organic search result and a sponsored search result, wherein the query response comprises an order of presentation of the organic search result and the sponsored search result indifferent positions of a scrollable product carousel in a graphical user interface (GUI) of the client device, causing the client device to display the GUI by rendering the organic search result and the sponsored search result in respective positions of the scrollable product carousel, the respective positions corresponding to the order of presentation, receiving and collecting data on user interactions with the query response, receiving and collecting data on user interactions with the query response, wherein fine-tuning the machine-learned generative language model comprises: adjusting, by the machine-learned generative language model, the order of presentation of the organic search result and the sponsored search result in the scrollable product carousel, however Nag discloses,
wherein the one or more featured products comprises an organic search result and a sponsored search result (Examiner interprets that Nag generates recommendation results containing relevant non-sponsored items and sponsored or promotional items and injects the sponsored items at positions among the relevant non-sponsored items i.e. Nag’s relevant non-sponsored search or recommendation result corresponds to the claimed organic search result) (0029-0030, 0035-0040, 0077-0084);
wherein the query response comprises an order of presentation of the organic search result and the sponsored search result indifferent positions of a scrollable product carousel in a graphical user interface (GUI) of the client device (Examiner interprets that Nag ranks sponsored and non-sponsored items, evaluates permutations of available presentation positions, injects sponsored items into positions among relevant items, and expressly identifies positions within a carousel. Nag also explains that an item displayed at a position to which the user is unlikely to scroll may be ineffective) (0030, 0036-0040, 0073-0075);
causing the client device to display the GUI by rendering the organic search result and the sponsored search result in respective positions of the scrollable product carousel, the respective positions corresponding to the order of presentation (Examiner interprets that Nag selects a combination and associated item positions, generates an ordered list identifying those positions, transmits the recommendations to the web server, and presents the selected items at the selected positions on the user device) (0045, 0074-0076, 0084-0085);
receiving and collecting data on user interactions with the query response (Nag collects item impressions, item clicks, recommendations viewed or clicked, advertisements viewed or clicked, cart additions, conversions, and click-through rates from the displayed recommendation interface) (0024, 0030, 0060-0061);
wherein fine-tuning the machine-learned generative language model comprises: adjusting, by the machine-learned generative language model, the order of presentation of the organic search result and the sponsored search result in the scrollable product carousel (Nag trains a Deep-Q network and position model to determine combinations and positions of relevant non-sponsored and sponsored items and presents the items in the optimized positions) (0048-0049, 0067-0076).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, wherein the one or more featured products comprises an organic search result and a sponsored search result, wherein the query response comprises an order of presentation of the organic search result and the sponsored search result indifferent positions of a scrollable product carousel in a graphical user interface (GUI) of the client device, causing the client device to display the GUI by rendering the organic search result and the sponsored search result in respective positions of the scrollable product carousel, the respective positions corresponding to the order of presentation, receiving and collecting data on user interactions with the query response, receiving and collecting data on user interactions with the query response, wherein fine-tuning the machine-learned generative language model comprises: adjusting, by the machine-learned generative language model, the order of presentation of the organic search result and the sponsored search result in the scrollable product carousel, as taught by Nag for the purpose for using mixed sponsored/non-sponsored carousel to identify product recommendations to produce a product enhanced response containing relevant organic and sponsored results displayed at optimized carousel positions.
Shitrit specifically doesn’t disclose, generating a prompt for input to a machine-learned generative language model, providing the prompt to a model serving system for execution by the machine-learned generative language model, receiving, from the model serving system, a response generated by executing the machine-learned generative language model on the prompt, receiving, from the model serving system, a response generated by executing the machine-learned generative language model on the prompt, and fine-tuning the machine-learned generative language model based on the data on user interactions with the query response, however Spiegel discloses,
generating a prompt for input to a machine-learned generative language model (Examiner notes that Spiegel discloses, chatbot system 300 receives user prompt 328, may generate adjusted input prompt 368, and communicates the prompt to LLM 338 for execution) (0083-0087), the prompt specifying at least a request related to the user query and a request to suggest the one or more featured products in association with a response to the prompt (Examiner interprets the prompt includes the user’s request, Spiegel trains the LLM on products currently offered by advertisers, the LLM returns product information responsive to requests for particular product categories, and advertising knowledge or textual advertising components may be incorporated into the LLM response) (0083, 0107, 0111-0112);
providing the prompt to a model serving system for execution by the machine-learned generative language model (Examiner notes that Spiegel states that LLM 338 may be hosted by a server system separate from the chatbot system and that the chatbot system communicates the user prompt to the remotely hosted LLM over a network) (0085-0087);
receiving, from the model serving system, a response generated by executing the machine-learned generative language model on the prompt (Examiner interprets that the LLM 338 receives prompt 328, generates raw response 362, and communicates raw response 362 to chatbot system 300 for subsequent processing) (0085-0087), the response including at least one of the one or more featured products (Spiegel trains the LLM on products offered by advertisers and states that the LLM provides product information responsive to product requests. Spiegel also describes a chatbot response including a list of hotels and a hotel promotion) (0107-0112);
and fine-tuning the machine-learned generative language model based on the data on user interactions with the query response (Spiegel expressly states that LLM 338 is continuously retrained or fine-tuned based on user interactions with advertising content and that collected advertising-engagement metrics provide reinforcement to the LLM) (0108) to improve a prediction of click-through rate (CTR) by the machine-learned generative language model for the one or more featured products (Spiegel further discloses collecting advertising-engagement information, including CTR-related information, generating expected-engagement scores, ranking advertising content, determining advertising priority or order, selecting top-ranked advertisements, and refining the ranking model using subsequent engagement) (0121, 0124-0132).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, generating a prompt for input to a machine-learned generative language model, providing the prompt to a model serving system for execution by the machine-learned generative language model, receiving, from the model serving system, a response generated by executing the machine-learned generative language model on the prompt, receiving, from the model serving system, a response generated by executing the machine-learned generative language model on the prompt, and fine-tuning the machine-learned generative language model based on the data on user interactions with the query response, as taught by Spiegel for the purpose to use the engagement information produced or refined through LLM based advertising system as an input, thereby adjusting the presentation order of sponsored and non-sponsored products in subsequent product responses according to improved engagement or CTR estimates.
As per claims 5 and 14, Shitrit discloses, wherein generating the query response comprises generating textual content incorporating the suggestion for the one or more featured products in the textual content (Examiner interprets the Shitrit automatically generates a textual answer responsive to a user’s natural-language question and Shitrit identifies product-related noun phrases within the textual answer, highlights or links those phrases to corresponding product results, and returns the enhanced textual answer to the client UI) (0012-0016, 0032-0033).
As per claims 6 and 15, Shitrit specifically doesn’t disclose, receiving a second user query from a second client device and a second request to include one or more consumer packaged good (CPG) products in a response, however Nag discloses, receiving a second user query from a second client device (Nag supports any number of separate customer computing devices. Each customer device may communicate with the web server, submit a search or recommendation query, and interact with displayed recommendations) (0022-0026);
generating a second prompt for input to the machine-learned generative language model, the second prompt specifying at least a second request related to the second user query and a second request to include one or more consumer packaged good (CPG) products in a response (Examiner notes that the underlined limitation is disclosed by another prior art. Nag’s catalog expressly includes grocery products, such as milk, with product IDs, brands, types, descriptions, options, and prices. Limiting Spiegel’s product request to CPG or grocery products is a predictable selection of product category.) (0064).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, receiving a second user query from a second client device and a second request to include one or more consumer packaged good (CPG) products in a response, as taught by Nag for the purpose for using mixed sponsored/non-sponsored carousel to identify product recommendations to produce a product enhanced response containing relevant organic and sponsored results displayed at optimized carousel positions.
Shitrit specifically doesn’t disclose, generating a second prompt for input to the machine-learned generative language model, the second prompt specifying at least a second request related to the second user query and receiving a second response generated by executing the machine-learned generative language model on the second prompt, the second response including the one or more CPG products, however Spiegel discloses,
generating a second prompt for input to the machine-learned generative language model, the second prompt specifying at least a second request related to the second user query (Spiegel receives client prompts, provides the prompts to an LLM, and conducts repeated or additional interactive sessions. Repeating the known prompt-generation process for another client is a predictable use of the multiuser server; The user prompt specifies the user’s conversational request, and the LLM generates a responsive answer based on the prompt) (0083-0087, 0102),
and receiving a second response generated by executing the machine-learned generative language model on the second prompt, the second response including the one or more CPG products (Spiegel’s product-aware LLM provides product information responsive to product-category requests and may include a product or service promotion in the generated response. Applying this operation to Nag’s grocery/CPG category produces a response containing CPG products.) (0107, 0111-0112).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, generating a second prompt for input to the machine-learned generative language model, the second prompt specifying at least a second request related to the second user query and receiving a second response generated by executing the machine-learned generative language model on the second prompt, the second response including the one or more CPG products, as taught by Spiegel for the purpose to allow prompt and response process for another user operating another customer device, and selecting a particular product category that does not alter the underlying LLM prompt architecture to maintain grocery products in its retailer catalog for second request to identify CPG or grocery products.
As per claims 7 and 16, Shitrit discloses, generating a second query response to the second user query by mapping the one or more CPG products to one or more products in a catalog of an online system (Examiner interprets that Shitrit parses a generated answer to identify product-related noun phrases, supplies the phrases to product search engine 360, searches products database 362, and identifies corresponding catalog products. The product phrase from the generated answer corresponds to the claimed generic CPG product; the returned database product corresponds to the mapped catalog product) (0016, 0032-0033);
and transmitting instructions to the second client device to cause presentation of the one or more products to a second user (UI generator 372 associates the identified product with the answer phrase, generates the enhanced response, and transmits the response through network 310 for display by client UI 302) (0016, 0032-0033).
As per claims 9 and 18, Allen discloses, wherein generating the suggestion for the one or more featured products includes creating hyperlinks associated with the one or more featured products, allowing direct interaction with the one or more featured products within the response (Examiner interprets that Shitrit highlights a product-related noun phrase in the generated answer and links or associates the phrase with corresponding product search results. The association may be implemented through selectable text or another UI association. Selecting the highlighted noun phrase launches a product-recommendation widget, product search, or shopping flow. Shitrit also displays an add-to-cart control associated with the product result.) (0012-0016, 0032-0033).
Claims 2, 11, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pub. 20240331004 (“Shitrit”) in view of U.S. Pub. 20220277345 (“Nag”) in further view of U.S. Pub. 20240249318 (“Spiegel”) in further view U.S. Pub. 20180293644 (“Allen”).
As per claims 2, 11, and 20, Shitrit specifically doesn’t disclose, generating a relevance score for each of a set of candidate featured products, the relevance score indicating a level of relevance of a respective featured product to the user query and selecting the one or more featured products, however Nag discloses, generating a relevance score for each of a set of candidate featured products, the relevance score indicating a level of relevance of a respective featured product to the user query (Nag generates semantic-similarity representations for sponsored and non-sponsored items and uses the representations to rank and score the items based on their relevance to the user for a current user session. Nag further discloses that each relevant item may have a corresponding relevancy score determined by relevance model 392 based on user-session and user-transaction data) (0031-0032, 0066),
and selecting the one or more featured products having relevance scores above a threshold for inclusion in the prompt (Examiner notes that the underlined limitation is disclosed by another prior art. Nag further discloses selecting sponsored items based on their relevance scores) (0039).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, generating a relevance score for each of a set of candidate featured products, the relevance score indicating a level of relevance of a respective featured product to the user query and selecting the one or more featured products, as taught by Nag for the purpose for using mixed sponsored/non-sponsored carousel to identify product recommendations to produce a product enhanced response containing relevant organic and sponsored results displayed at optimized carousel positions.
Shitrit specifically doesn’t disclose, having relevance scores above a threshold, however Allen discloses, having relevance scores above a threshold (Allen discloses determining a respective confidence score for each candidate category, comparing the score with a predetermined threshold, retaining candidates whose scores exceed the threshold, and removing candidates whose scores fail to exceed the threshold from further consideration) (0032).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, having relevance scores above a threshold, as taught by Allen for the purpose to apply threshold filtering technique to product level relevance scores by selecting candidate featured products whose respective relevance scores exceed a predetermined threshold to exclude insufficiently relevant candidate products and retain products having a sufficient relationship to the user query.
Shitrit specifically doesn’t disclose, for inclusion in the prompt, however Spiegel discloses, for inclusion in the prompt (Spiegel discloses generating an input prompt for an LLM, communicating the prompt to the LLM, and training the LLM using information concerning products currently offered by advertisers; Spiegel further teaches that the LLM provides product information in response to user requests for particular product categories) (0083-0087, 0107, 0112).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, for inclusion in the prompt, as taught by Spiegel for the purpose to use LLM prompt to constrain the requested product suggestions to candidates already determined to have a sufficient level of relevance to the user query, thereby reducing irrelevant product suggestions and improving the relevance of the generated response.
Claims 3, 8, 12, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pub. 20240331004 (“Shitrit”) in view of U.S. Pub. 20220277345 (“Nag”) in further view of U.S. Pub. 20240249318 (“Spiegel”) in further view U.S. Pub. 20140006170 (“Collette”).
As per claims 3 and 12, Shitrit specifically doesn’t disclose, receiving a bid value from each of a plurality of product providers that offer a set of candidate featured products; and selecting the one or more featured products having bid values above a threshold for inclusion in the prompt, however Collette discloses, receiving a bid value from each of a plurality of product providers that offer a set of candidate featured products (Examiner interprets AppNexus receives bids from multiple impression buyers. Each bid has an associated bidder attribute and bid-value attribute. In the combined product-advertising system, the impression buyers correspond to advertisers or product providers, and their advertising creatives correspond to candidate featured-product promotions) (Abstract, Fig. 12, steps 1204–1244, §5.6.1);
and selecting the one or more featured products having bid values above a threshold for inclusion in the prompt (Examiner notes that the underlined limitation is disclosed by another prior art. Examiner interprets that Collette compares each bid value with a minimum bid threshold. A bid failing the minimum is excluded from the relevant auction tier; qualifying bids are compared, and the highest qualifying bid is selected as the winner) (Abstract, Fig. 12, steps 1212–1244, §5.6.1-Exemplary Tiered Auctions).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, receiving a bid value from each of a plurality of product providers that offer a set of candidate featured products; and selecting the one or more featured products having bid values above a threshold, as taught by Collette for the purpose to apply minimum-bid auction technique to the candidate sponsored products to exclude nonqualifying sponsored products and select an eligible product for limited promotional placement so that the generated response suggests the sponsored product selected for presentation.
Shitrit specifically doesn’t disclose, for inclusion in the prompt, however Spiegel discloses, for inclusion in the prompt (Spiegel discloses generating an input prompt for an LLM, communicating the prompt to the LLM, and training the LLM using information concerning products currently offered by advertisers; Spiegel further teaches that the LLM provides product information in response to user requests for particular product categories; Spiegel teaches generating an LLM prompt and conditioning product-related LLM responses using advertiser-product and advertising information. Spiegel does not expressly state that the bid-qualified product is inserted into the prompt. Inclusion of the auction-selected product as prompt context is the proposed modification) (0083-0087, 0107, 0112).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, for inclusion in the prompt, as taught by Spiegel for the purpose to use LLM prompt to constrain the requested product suggestions to candidates already determined to have a sufficient level of relevance to the user query, thereby reducing irrelevant product suggestions and improving the relevance of the generated response.
As per claims 8 and 17, Shitrit discloses, wherein generating the second query response further comprises: mapping a CPG product to a set of candidate products (Examiner interprets that Shitrit extracts a generic product-related noun phrase from the generated answer and provides it to product search engine 360. The search engine searches products database 362 and returns related products. A generic product phrase can therefore map to multiple candidate catalog products) (0016, 0032-0033).
Shitrit specifically doesn’t disclose, receiving a bid value for each candidate product in the set of candidate products and performing an auction process to select a product from the set of candidate products having a bid value above a threshold, however Collette discloses, receiving a bid value for each candidate product in the set of candidate products (Examiner interprets that Collette receives bids from multiple impression buyers. Each bid contains a bidder attribute and a bid-value attribute for a candidate advertisement. In the product-advertising combination, the candidate advertisements correspond to sponsored candidate products) (Abstract, Fig. 12, steps 1204-1220, §5.6.1-Exemplary Tiered Auctions);
and performing an auction process to select a product from the set of candidate products having a bid value above a threshold (Examiner interprets that Collette conducts a real-time advertising auction, assigns qualifying bids to auction tiers, selects the active tier, and compares the assigned bids to determine the winning bid and compares each bid with a minimum bid threshold. Bids below the threshold are excluded; the highest bidder meeting the minimum price is selected as the winner.) (Abstract, Fig. 12, steps 1204, 1212-1244, §5.6.1-Exemplary Tiered Auctions).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, receiving a bid value for each candidate product in the set of candidate products and performing an auction process to select a product from the set of candidate products having a bid value above a threshold, as taught by Collette for the purpose apply real-time advertising auction to the candidate products identified through catalog search to select an eligible sponsored candidate product for the limited sponsored-product position while excluding candidates whose bids fail the threshold.
Claims 4 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pub. 20240331004 (“Shitrit”) in view of U.S. Pub. 20220277345 (“Nag”) in further view of U.S. Pub. 20240249318 (“Spiegel”) in further view U.S. Pat. 9,165,320 (“Belvin”).
As per claims 4 and 13, Shitrit specifically doesn’t disclose, wherein generating the query response comprises generating a recipe page including a list of products for fulfilling the recipe, however Belvin discloses, wherein generating the query response comprises generating a recipe page including a list of products for fulfilling recipes, wherein the list of products includes the one or more featured products (Examiner notes that the underlined limitation is disclosed by another prior art. Examiner interprets the Belvin generates interface 300 as a webpage displaying a recipe, including recipe instructions, an ingredient list, nutritional information, and directions. Belvin further identifies items needed to complete a recipe, maps recipe ingredients to electronic-marketplace products, and generates a shopping-cart list containing the products needed to complete the recipe. Fig. 6 additionally lists ingredients needed to complete each recipe) (col. 9, line 65- col. 10, line 28, col. 9, line 25- col. 10, line 28, col. 11, line 65- Co. 12, line 15, Figs. 2–3 and 5-6).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, wherein generating the query response comprises generating a recipe page including a list of products for fulfilling the recipe, as taught by Belvin for the purpose to use recipe-oriented interface when the user query concerns preparing a recipe when presenting the catalog products needed to complete the recipe facilitates identification and purchase of the necessary products thus to generate a recipe response containing the featured products corresponding to required recipe items.
Shitrit specifically doesn’t disclose, wherein the list of products includes the one or more featured products, however Nag discloses, wherein the list of products includes the one or more featured products (Examiner notes that Shitrit identifies catalog products associated with generated answer content, while Nag identifies relevant and sponsored products. It would have been obvious to use those identified featured products as some of Belvin’s products needed to fulfill the recipe) (0030, 0035-0040).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the applicant’s invention for receiving, from a client device and via an interface, a user query, identifying one or more featured products based on the user query, generating a query response to the user query based on the response generated by executing the machine-learned generative language model on the prompt, the query response including at least a suggestion for the at least one of the one or more featured products, transmitting instructions, to the client device, to cause presentation of the query response to the user query, as taught by Shitrit, wherein the list of products includes the one or more featured products, as taught by Nag for the purpose for using mixed sponsored/non-sponsored carousel to identify product recommendations to produce a product enhanced response containing relevant organic and sponsored results displayed at optimized carousel positions.
Response to Arguments
With regards to § 101 rejections:
The arguments filed on May 18th, 2026, with respect to the rejection(s) of claims 1-20 under 35 U.S.C 101 have been fully considered but are unpersuasive/moot. he rejection is maintained and updated to address the amended claims.
Applicant states that, that amended independent claims 1, 10, and 19 provide a technical solution to “static LLM-based chatbots” by adjusting products displayed in a scrollable carousel according to user interactions and by fine-tuning a machine-learned model based on improved interaction metrics, such as click-through rate. Applicant further contends that the fine-tuning improves the machine-learning model and adjusts product positioning in the carousel. Remarks 12-14.
Applicant's arguments have been fully considered but they are not persuasive. Considered as a whole, the amended claims remain focused on selecting and presenting organic and sponsored products, monitoring customer interactions with those products, predicting customer engagement, and modifying the relative placement of the products to improve CTR. These activities constitute product recommendation, advertising placement, marketing, and sales optimization. Advertising, marketing, and sales activities or behaviors fall within the commercial-interaction subgroup of “certain methods of organizing human activity.” The recitation of a machine-learned generative language model, model serving system, GUI, scrollable carousel, and fine-tuning does not, by itself, establish an improvement to computer or machine-learning technology. The claims do not recite a particular model architecture, training-data structure, loss function, parameter-update procedure, objective function, or other specific technical mechanism by which the LLM is improved. Instead, the claims recite the desired commercial results of improving a CTR prediction and adjusting the order in which organic and sponsored products are presented. The asserted improvement therefore concerns the effectiveness of product promotion and advertisement placement, rather than an improvement in how the computer or machine learning model itself operates.
Applicant’s reliance on Ex Parte Desjardins is also unpersuasive. In Desjardins, the claims recited a specific machine learning training procedure that determined the importance of model parameters and adjusted those parameters using an objective function and penalty term so that the model could learn a second task while protecting performance on a first task. The disclosed and claimed improvement reduced storage requirements and system complexity and addressed the technical problem of catastrophic forgetting. The Appeals Review Panel found that the claims themselves reflected an improvement in the operation of the machine learning model. Here, by contrast, the amended claims do not recite a comparable technical training procedure or improvement in model operation. The claims use interaction and CTR information to improve the commercial ordering of products, without reciting how the generative model is technically modified to achieve that result.
Applicant’s reliance on PEG Example 37 is likewise unpersuasive. Example 37 recited a specific manner of improving a GUI by automatically moving the most used icons to a defined position closest to the computer’s start icon. That specific arrangement improved the operation and usability of the interface itself. The present claims merely require organic and sponsored product results to occupy different positions in a scrollable carousel and require their order to be adjusted based on commercial engagement information. The claims do not recite a new carousel structure, rendering technique, navigation operation, memory-management process, or other improvement to GUI technology. The carousel merely presents the results of the abstract advertising and recommendation process.
Applicant also states that the claims must be considered as a whole. The Examiner agrees and has considered all limitations individually and as an ordered combination. Nevertheless, the combination merely uses a client device, model serving system, LLM, GUI, and carousel as tools to carry out the abstract commercial practice. Receiving the query supplies information for the recommendation; transmitting the prompt and response communicates that information; rendering the carousel presents the recommendations; and collecting user interactions gathers customer-response statistics used to optimize later product placement. Presenting offers, gathering customer response statistics, and using those statistics to optimize a commercial result constitute insignificant data-gathering and extra-solution activity that does not meaningfully limit the exception. Accordingly, the amended claims continue to recite certain methods of organizing human activity involving commercial product recommendation, advertising, marketing, and sales optimization. The additional elements, individually and as an ordered combination, do not integrate the abstract idea into a practical application and do not amount to significantly more than the judicial exception. The asserted improvement is an improvement to the commercial effectiveness of recommendations and advertisements, rather than an improvement to computer functionality, machine learning technology, or another technical field. See MPEP §§ 2106.05(a)–(c), (e)– (h). Therefore, the rejection of claims 1-20 under 35 U.S.C. §101 is maintained.
With regards to § 103 rejections:
Applicant's arguments, see pages 15-20, filed October 11th, 2024, with respect to the rejection(s) of claims 1-5, 10-14, and 19 under 35 U.S.C 102/103 have been fully considered but are unpersuasive/moots on new ground of rejection.
Thus, the dependent claims 2-5 and 11-14 that depend from claims 1 and 10 respectively are also moots.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. US. Pat. 8306966 (“Liu”).
Liu outlines methods are disclosed for optimizing the relative positions of sponsored and organic search results on a search results page displayed to a user, including calculating, by a search engine, a degree of commerciality of a search query; receiving the search query from a user by the search engine; and delivering, by the search engine to a browser of a user, a plurality of sponsored search results in at least two columns and a plurality of organic search results in at least a third column of one or more search results pages, wherein a layout of the at least three columns depends on the degree of commerciality of the search query.
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 GAUTAM UBALE whose telephone number is (571)272-9861. The examiner can normally be reached Mon-Fri. 7:00 AM- 6:30 PM PST.
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, Marissa Thein can be reached at (571) 272-6764. 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.
/GAUTAM UBALE/
Primary Examiner, Art Unit 3689
Prosecution Insights