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 .
Response to Amendment
The amendment filed on May 28, 2026 cancelled no claims. Claims 1, 9, and 15 were amended and no new claims were added. Thus, the currently pending claims addressed below are claims 1, 4-9, 12-15, and 18-20.
Claim Interpretation
The following claim terms have been interpreted in light of the applicant’s specification:
Data structure: a content campaign (Paragraph 44 of the applicant’s specification states: the content campaign can be understood as a data structure or set of data structures that are coordinated to execute computerized promotional activities over a network related to the generation, manipulation and/or delivery of specified forms of digital content.);
Facet: text-based aspects that make up a subject or an object (The applicant’s specification does not define the term facet. The Merriam-Webster Online dictionary defines a facet as any of the definable aspects that make up a subject or an object. However, the claim requires the facets be processed by an NLP model which requires the facets to be text, which narrows the definition to be text-based aspects that make up a subject or an object. The examples of facets in at least paragraphs 47 and 53-54 of the applicant’s specification are consistent with the above interpretation. While paragraph 53 of the applicant’s specification indicates that facets include plain text from an NLP processing the facet data and creative elements (e.g., media type, media format and the like), the creative element facets are determined by AI/ML and/or LLM processing, not NPL processing. Finally, the claim requires determining both a set of facets and at least one of a media type or media format, wherein the set of facets are input into an NLP model to generate plain text. Thus, the claimed set of facets are different and distinct from the creative elements (e.g., media type, media format and the like);
A set of facets related to the data structure: a set of features related to the content campaign (based on the interpretations of the terms facet and data structure above);
Large language model (LLM) prompt: plain-text facet data (The applicant’s specification provides no definition of an LLM prompt. However, the claim limitation itself requires the prompt generated, using the NLP model and based on the set of facets, to be plain-text facet data.).
Natural language processing (NLP) model: a statistical NLP model or a machine learning NLP model (The applicant’s specification does not provided a definition of an NPL model. Since there are both statistical NLP models and machine learning NLP models the term is broad enough to encompass statistical NLP models as well.)
Claim Rejections - 35 USC § 112
The amendment filed on May 28, 2026 has overcome the 35 U.S.C. 112(a) rejections of claims 1, 4-9, 12-15, and 18-20 which was directed to “generating, by the device, based on the set of facets, a large language model (LLM) prompt, the LLM prompt generation comprising serializing, via a natural language processing (NLP) model, the set of facets into plain-text facet data” because the term serializing was removed from the claims. Thus, these rejections are hereby withdrawn.
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1, 4-9, 12-15, and 18-20 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 applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Independent claims 1, 9 and 15, have been amended to recite “the data structure executing activities over a network related to at least one of generation, manipulation, or delivery of digital content”. The examiner has been unable to find support for this limitation in the applicant’s specification. There is no support in the applicant’s disclosure for a “data structure” that executes activities over a network. The closest support is found in paragraph 44, where a data structure is a “content campaign” and the invention used the data in the data structure to execute promotional activities over a network, wherein the promotional activities include generating specified forms of digital content, manipulating specified forms of specified digital content, and/or delivering, over the network, the specified forms of digital content. Nothing in paragraph 44, indicates that the “data structure” itself which is a content campaign executes any activities. As made clear in paragraph 45, the executing of the content campaign is defined and/or controlled a DSPs, ad formats, ad channels, targeting parameters, target audience, keyword optimizations, tracking tools and analytics, budget and compliance, and the like, or some combination thereof, can be tailored to campaign objectives. Accordingly, the content campaign can correlate to and/or indicate a goal or an intent, which can be associated with a type of audience, an intended audience, and/or a time period, geographic area and/or form or type of content, among other parameters defined by the campaign. As such, the data structure is merely a content campaign comprising data such as a goal or an intent of the campaign, as well as a type of audience, an intended audience, and/or a time period, geographic area and/or form or type of content, among other parameters. It is not computer code and, as such, cannot execute any activity on the computer on which it is stored, much less execute activities that occur on other computers over a network. It is clear that independent claims 1, 9, and 15, as currently amended, require that the “data structure” actually execute activities, and that a data structure performing such execution of activities is not supported by the applicant’s disclosure. Therefore, independent claims 1, 9, and 15, as currently amended, fail to comply with the written description requirement.
. Dependent claims 4-8, 12-14, and 18-20 fail to cure the deficiencies of the claims from which they depend and, as such, are rejected by virtue of dependency.
Claims 1, 4-9, 12-15, and 18-20 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 applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Based on independent claims 1, 9 and 15, as currently amended, the LLM prompt comprises only plain-text facet data based on historical facet data. The LLM prompt (comprising only plain-text face data based on historical facet data) is input into the LLM model. The LLM model generates LLM output comprising information related to attributes indicating the intended activity corresponding to the set of facets determined based solely on the historical facet data. The claim then curates the data structure (e.g., either not changing the data structure or changing the format of content within the data structure or changing the quantity of content items within the data structure) by analyzing the attributes from the LLM output, which is based solely on historical facet data, using an ML model, and determining an effectiveness of the data structure. The examiner can find no support in the applicant’s specification for curating the data structure based solely on information derived from historical facet data as currently claimed. The closest support in the applicant’s specification is found in paragraphs 53-59. According to paragraph 53, facets that capture the intent of the current campaign comprise creative elements (e.g., media type, media format and the like), digital platforms, user actions, timeline and scheduling, consistency, planning and buying, campaign objectives, and the like, or some combination thereof. According the paragraph 54, facets can further be based on past/historical facets from previous campaigns. Thus, the facets disclosed in paragraphs 53-54 comprise at least two facets of the current campaign (i.e., some combination of creative elements, digital platforms, user actions, timeline and scheduling, consistency, planning and buying, campaign objectives, and the like) and may additionally include past/historical facets from previous campaigns. Thus, the LLM prompt generated based on the facets, as disclosed in paragraph 55, must include at least two current campaign facets and might include additional historical facets of previous campaigns. The LLM output, attributes, and curation disclosed in paragraph 56-59 are also performed based on at least two current campaign facets and might include additional historical facets of previous campaigns. Therefore, it is clear that paragraphs 56-59 do not support performing the claimed steps using only past/historical faces from previous campaigns. As such, independent claims 1, 9, and 15, as currently amended, fail to comply with the written description requirement.
Dependent claims 4-8, 12-14, and 18-20 fail to cure the deficiencies of the claims from which they depend and, as such, are rejected by virtue of dependency.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1, 4-9, 12-15, and 18-20 are directed to a method, a system, and a computer program product which would be classified under one of the listed statutory classifications (i.e., 2019 Revised Patent Subject Matter Eligibility Guidance (hereinafter “PEG”) “PEG” Step 1=Yes).
However, claims 1, 4-9, 12-15, and 18-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Independent claim(s) 1, 9, and 15 recite(s) the following abstract idea: (Examiner note: the network has been included as part of the abstract idea because it is outside the scope of the claimed invention and, as such, cannot be considered an “additional element” of the claimed invention)
identifying a data structure (content campaign) on a platform that comprises at least one intended user action (e.g., goal or intent with regards to clicks, views, conversions, likes, shares etc.: applicant’s specification paragraphs 45 and 58) and a plurality of implementation parameters, the data structure executing activities over a network related to at least one of generation, manipulation, or delivery of digital content;
determining, based on historic facet data, a set of facets related to the data structure (i.e., a set of features related to the content campaign) and at least one of a media type or a media format, the set of facets corresponding to intended activity related to the data structure on a network;
generating. based on the set of facets, a large language model (LLM) prompt (i.e., plain text facet data), the LLM prompt generation comprising processing, via a natural language processing (NLP) model, the set of facets into plain-text facet data
generating via a language model using the LLM prompt (i.e., plain text facet data), a language model output comprising information related to attributes indicating the activity related to the data structure;
curating, based on the LLM output comprising information related to attributes indicating the activity related to the data structure, the data structure. the curation comprising:
analyzing, by an algorithmic model, the attributes from the language model output;
determining an effectiveness of the data structure, wherein the curation of the data structure is based on the determined effectiveness;
determining whether the effectiveness of the data structure satisfies a performance threshold, wherein: when the performance threshold is satisfied, the curation comprises maintaining the data structure in its current form, and when the performance threshold is not satisfied, the curation comprises modifying the data structure by changing a format of content withing the data structure and changing a quantity of content items within the data structure; and
communicating, over the network, the curated data structure to a set of network resources.
The limitations as detailed above, as drafted, falls within the “Certain Method of Organizing Human Activity” grouping of abstract ideas namely advertising, marketing, or sales related activities or behaviors. Accordingly, the claim recites an abstract idea (i.e., “PEG” Revised Step 2A Prong One=Yes).
This judicial exception is not integrated into a practical application because the claim only recites the additional elements of:
a device with a processor executing instructions (e.g., a general-purpose computer),
a large language model (LLM) (e.g., a generic computer component), and
a machine learning (ML) model (e.g., a generic computer component).
The following limitations, if removed from the abstract idea and considered additional elements, merely perform generic computer function of processing, communicating (e.g., transmitting and receiving), and displaying:
communicating, over the network, the curated data structure to a set of network resources (transmitting data).
The additional technical elements above are recited at a high-level of generality (i.e., as a generic processor performing a generic computer function of processing, communicating (e.g., transmitting and receiving), and displaying) such that it amounts to no more than mere instructions to apply the exception using one or more general-purpose computers and/or one or more generic computer components. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional technical elements above do not integrate the abstract idea/judicial exception into a practical application because it does not impose any meaningful limits on practicing the abstract idea. More specifically, the additional elements fail to include (1) improvements to the functioning of a computer or to any other technology or technical field (see MPEP 2106.05(a)), (2) applying or using a judicial exception to effect a particular treatment or prophylaxis for a disease or medical condition (see Vanda memo), (3) applying the judicial exception with, or by use of, a particular machine (see MPEP 2106.05(b)), (4) effecting a transformation or reduction of a particular article to a different state or thing (see MPEP 2106.05(c)), or (5) applying or using the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception (see MPEP 2106.05(e) and Vanda memo).
Rather, the limitations merely add the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (see MPEP 2106.05(f)), or generally link the use of the judicial exception to a particular technological environment or field of use (see MPEP 2106.05(h)).
Thus, the claim is “directed to” an abstract idea (i.e., “PEG” Revised Step 2A Prong Two=Yes)
When considering Step 2B of the Alice/Mayo test, the claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the claims do not amount to significantly more than the abstract idea.
More specifically, as discussed above with respect to integration of the abstract idea into a practical application, the additional elements of using a device with a processor executing instructions including a large language model (LLM) and a machine learning (ML) model to perform the claimed functions amounts to no more than mere instructions to apply the exception using one or more general-purpose computers and/or one or more generic computer components.
“Generic computer implementation” is insufficient to transform a patent-ineligible abstract idea into a patent-eligible invention (See Affinity Labs, _F.3d_, 120 U.S.P.Q.2d 1201 (Fed. Cir. 2016), citing Alice, 134 S. Ct. at 2352, 2357) and more generally, “simply appending conventional steps specified at a high level of generality” to an abstract idea does not make that idea patentable (See Affinity Labs, _F.3d_, 120 U.S.P.Q.2d 1201 (Fed. Cir. 2016), citing Mayo, 132 S. Ct. at 1300). Moreover, “the use of generic computer elements like a microprocessor or user interface do not alone transform an otherwise abstract idea into patent-eligible subject matter (See FairWarning, 120 U.S.P.Q.2d. 1293, citing DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1256 (Fed. Cir. 2014)). As such, the additional elements of the claim do not add a meaningful limitation to the abstract idea because they would be generic computer functions in any computer implementation. Thus, taken alone, the additional elements do not amount to significantly more than the above-identified judicial exception (the abstract idea). Looking at the limitations as an ordered combination adds nothing that is not already present when looking at the elements taken individually. There is no indication that the combination of elements improves the functioning of the computer or improves any other technology. Their collective functions merely provide generic computer implementation.
The Examiner notes simply implementing an abstract concept on a computer, without meaningful limitations to that concept, does not transform a patent-ineligible claim into a patent-eligible one (See Accenture, 728 F.3d 1336, 108 U.S.P.Q.2d 1173 (Fed. Cir. 2013), citing Bancorp, 687 F.3d at 1280), limiting the application of an abstract idea to one field of use does not necessarily guard against preempting all uses of the abstract idea (See Accenture, 728 F.3d 1336, 108 U.S.P.Q.2d 1173 (Fed. Cir. 2013), citing Bilski, 130 S. Ct. at 3231), and further the prohibition against patenting an abstract principle “cannot be circumvented by attempting to limit the use of the [principle] to a particular technological environment” (See Accenture, 728 F.3d 1336, 108 U.S.P.Q.2d 1173 (Fed. Cir. 2013), citing Flook, 437 U.S. at 584), and finally merely limiting the field of use of the abstract idea to a particular existing technological environment does not render the claims any less abstract (See Affinity Labs, _F.3d_, 120 U.S.P.Q.2d 1201 (Fed. Cir. 2016), citing Alice, 134 S. Ct. at 2358; Mayo, 132 S. Ct. at 1294; Bilski v. Kappos, 561 U.S. 593, 612 (2010); Content Extraction & Transmission LLC v. Wells Fargo Bank, Nat’l Ass’n, 776 F.3d 1343, 1348 (Fed. Cir. 2014); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355 (Fed. Cir. 2014).
Applicant herein only requires one or more general-purpose computers and/or one or more generic computer components (as evidenced from paragraphs 20-27 of the applicant’s specification with discloses that the processor is part of a general-purpose computer; Crossplag (What is Aigiarism?, December 23, 2022, https://web.archive. org/ web/20221223172631/https://crossplag.com/what-is-aigiarism/, pgs. 1-7) which discloses on at least page 1, lines 1-13 that Large Language Models (LLM) were well known by at least December 23, 2022; Langley et al. (“Approaches to Machine Learning”, Journal of the American Society for Information Science, February 16, 1984, pgs. 1-28) which discloses on at least page 7, lines 23-29 of that discloses that machine learning models are well-known by at least February 16, 1984); therefore, there does not appear to be any alteration or modification to the generic activities indicated, and they are also therefore recognized as insignificant activity with respect to eligibility. Finally, the following limitations, if removed from the abstract idea and considered additional elements, would be considered insignificant extra solution activity as they are directed to merely receiving, storing and/or transmitting data:
communicating, over the network, the curated data structure to a set of network resources (transmitting data).
Thus, taken individually and in combination, the additional elements do not amount to significantly more than the above-identified judicial exception (the abstract idea) (i.e., “PEG” Step 2B=No).
The dependent claims 4-8, 12-14, and 18-20 appear to merely further limit the abstract idea by adding the additional steps of analyzing the attributes from the output, determining an effectiveness of the data structure which are considered part of the abstract idea and further limiting the modification of the data structure which is considered part of the abstract idea (Claims 4, 12, and 18); further limiting the data structure and the determination of effectiveness which is considered part of the abstract idea (Claims 5, 6, 13, and 19); adding the additional steps of retrieving data from a previous data structure and making the determination based on the previous data structure which are considered part of the abstract idea (Claims 7, 14, and 20); further limiting the data structure and the intended user action which are considered part of the abstract idea (Claim 8), and therefore only further limit the abstract idea (i.e. “PEG” Revised Step 2A Prong One=Yes), does/do not include any new additional elements that are sufficient to amount to significantly more than the judicial exception, and as such are “directed to” said abstract idea (i.e. “PEG” Step 2A Prong Two=Yes); and do not add significantly more than the idea (i.e. “PEG” Step 2B=No)..
Thus, based on the detailed analysis above, claims 1, 4-9, 12-15, and 18-20 are not patent eligible.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1, 4-9, 12-15, and 18-20 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Bent, III et al. (PGPUB: 2024/0126997).
Claims 1, 9, and 15: Bent discloses a method, a device, and a non-transitory computer-readable storage medium comprising:
a processor of the device (Paragraph 56) configured to:
identifying a data structure (content campaign) on a platform that comprises at least one intended user action (e.g., goal or intent with regards to clicks, views, conversions, likes, shares etc.: applicant’s specification paragraphs 45 and 58) and a plurality of implementation parameters, the data structure executing activities over a network related to at least one of generation, manipulation, or delivery of digital content (Paragraph 103: obtaining user input regarding the generation of a new content item such as an advertisement or advertisement campaign, wherein the input includes a content item group; Paragraphs 106 and 110: input includes landing page URL; Paragraph 108: input includes information about advertisers business and information about the product or service to advertise; Paragraph 131: input includes text, documents, images, handwriting, audio, and content item strategy associated with a campaign; Paragraph 194: input includes a desired audience (e.g., targeting), an amount to pay for content being displayed and measures of performance (e.g., bidding), or the content that can be displayed (e.g., creatives));
determining, based on historical facet data, a set of facets related to the data structure and at least one of a media type or a media format, the set of facets corresponding to intended activity related to the data structure on a network (Fig. 9-10 and Paragraph 107-117 and 152-154: The content assistant component can be a progressive disclosure field or a conversational interface where the user first provides original facet data (historic facet data), then in response to a query from the content assistant provides additional facet information (historic facet data); the computing system extracts information from the user input including business name and descriptive terms and phrases; the user then provides user input indicative of a desire for modifications of the suggested content (historic facet data), wherein after the user provides the last desired modification (current facet data) and input indication of a satisfactions; Paragraph 52: The user input data can be associated with a current user session and/or include historical user data; Figure 4 and Paragraph 103: user provides input regarding the type of new content item and the content item group they wish to have created (e.g., advertisement, advertisement campaign); Paragraph 133: transforming the user input into a feature vector or some other data structure to be ingested by the first machine learned model; Paragraphs 195-196: monitoring performance metrics, determining statistically significant change in performance; Paragraph 77: the content is served over a network and interactions are obtained over the network);
generating, based on the set of facets, a large language model (LLM) prompt, the LLM prompt generation comprising processing, via a natural language processing (NLP) model, the set of facets into plain-text facet data (Paragraph 42: analyzing the user input can be performed by machine learned models (e.g., natural language processing models); Paragraph 70: input to the machine-learned model(s) of the present disclosure can be text or natural language data; the machine-learned model(s) can process the text or natural language data to generate an output such as a language encoding output, a latent text embedding output, a translation output, a classification output, a textual segmentation output, a semantic intent output comprising at least one word or phrase determined from the text or natural language data, an upscaled text or natural language output (e.g., text or natural language data that is higher quality than the input text or natural language, etc.) and/or a prediction output; Paragraphs 79 and 158: the content assistant component can include a plurality of models that use a large language model to better understand an intent of the obtained user input data; Paragraph 169: generation of advertisements, using the system, requires understanding of nuances of businesses, business goals, products, and the like and the system uses a large language model to tune the models to provide higher quality output; Paragraph 178: the computing system can determine a semantic intent through natural language processing (e.g., via a large language model);
generating via an LLM, using the LLM prompt, an LLM output comprising information related to attributes indicating the activity related to the data structure (Paragraph 134: output data generated indicative of one or more content item components; Paragraph 146: output includes suggested headlines, descriptions, images, videos, taglines, etc.; Paragraph 155: output includes both the suggested content items and predicted performance increases associated with the suggested content items; Paragraph 159: system determines intent of obtained user input and provides recommendations based on the intent, wherein recommendations include recommendations for analyzing historical content performance data (e.g., advertisement performance, cost per click, return on advertisement spend), recommendations for adjusting content campaign parameters based on analyzed data (e.g., adjust spend across different media channels, adjust maximum or minimum bidding parameters), or any other recommendation associated with a content campaign; Paragraph 165: the system generates predictions and/or inferences based on the features in the feature vector; Paragraph 169: the advertisement is generated based an understanding obtained from inputs such as nuances of businesses, business goals, products, and the like; Paragraphs 196-197: updating recommendations based on performance metrics)
curating, based on the LLM output comprising information related to attributes indicating the activity related to the data structure, the data structure (Paragraph 112: the computing system can generate suggested content such as a summary of a strategy for an advertisement campaign comprising one or more advertisements, requirements for display, time for display, expected performance, etc.; Paragraph 118: obtaining user input regarding the suggestions; Paragraph 137: upon user approval, generating the advertisement which may comprise a plurality of content items such as headlines, descriptions, videos, images, taglines, etc.), the curation comprising:
analyzing, by a machine learning (ML) model, the attributes from the LLM output; determining an effectiveness of the data structure, wherein the curation of the data structure is based on the determined effectiveness (Paragraphs 172-173, 181, and 218-221: determining, by a content assistant component, a content score for suggested content and providing additional recommendations and obtaining additional inputs until a predicted score of the generated content exceeds a threshold; Paragraph 170: the content assistant component can be a machine-learned model); and
determining whether the effectiveness of the data structure satisfies a performance threshold, wherein: when the performance threshold is satisfied, the curation comprises maintaining the data structure in its current form, and when the performance threshold is not satisfied, the curation comprises modifying the data structure by changing the format of content within the data structure and changing the quantity of content items within the data structure; (Paragraphs 172-173, 181, and 218-221: determining, by a content assistant component, a content score for suggested content and providing additional recommendations and obtaining additional inputs until a predicted score of the generated content exceeds a threshold, wherein the modifications include displaying additional content such as images; Paragraphs 196-197: the system can suggest changes to parameters including the reason for the change and/or automatically adjust parameters of the content based on performance metrics, wherein parameters include adding or removing content elements (e.g., descriptions, keywords, images, audio, video), adjusting bidding strategy, or other adjustments; Paragraphs 113-114: first an advertisement content is created and then, after said creation, the content is modified; Figures 18-20: adding the image in figure 20, reformats that way in which the text is displayed such that before the image was added the text spanned the horizontal distance of the advertisement, and after the image was added the text now only spans about half of the distance of the advertisement); and
communicating, over the network, the curated data structure to a set of network resources. (Paragraph 164: performing pilot and/or live traffic experiments to capture online metrics with regards to the suggested content)
Claims 4, 12, and 18: Bent discloses the method of claim 1, the device of claim 9, and the non-transitory computer-readable storage medium of claim 15, wherein the modification of the data structure comprises at least one of adding content, removing content, changing target information, changing frequency, changing timing, changing location, and changing platforms. (Paragraphs 172-173, 181, and 218-221: modifications include displaying additional content such as images; Paragraphs 196-197: the system can suggest changes to parameters including the reason for the change and/or automatically adjust parameters of the content based on performance metrics, wherein parameters include adding or removing content elements (e.g., descriptions, keywords, images, audio, video), adjusting bidding strategy, or other adjustments))
Claims 5, 6, 13, and 19: Bent discloses the method of claim 1, the device of claim 9, and the non-transitory computer-readable storage medium of claim 15,
wherein when the data structure is an existing, launched data structure on the network, the determined effectiveness corresponds to a realized effectiveness (Paragraphs 195-197: the content assistant component can monitor one or more performance metrics, continually review the content campaign performance metrics and provide recommendations and updates in real time), and
wherein when the data structure is a new data structure to the network, the determined effectiveness corresponds to a predicted effectiveness (Paragraph 55: the computing system can use user context data predicted performance increases (e.g., predicted performance metrics); Paragraph 65: the training data can include, for example, past performance metrics (e.g., predicted performance increase(s)); Paragraph 77: generating predicted performance increases associated with suggested content items and providing an updated user interface including the suggested content items and the respective predicted performance increase; Paragraph 181: providing recommendations to update a content creation interface until a predicted score of the generated content exceeds a threshold Paragraphs 159: analyzing historical content performance data (e.g., advertisement performance, cost per click, return on advertisement spend), adjusting content campaign parameters based on analyzed data (e.g., adjust spend across different media channels, adjust maximum or minimum bidding parameters) .
Claims 7, 14, and 20: Bent discloses the method of claim 1, the device of claim 9, and the non-transitory computer-readable storage medium of claim 15, further comprising: retrieving data from a previous data structure that implemented parameters from within the plurality of implementation parameters, wherein the set of facets determination is further based on the previous data structure. (Paragraph 65: The training data can include, for example, past performance metrics (e.g., predicted performance increase(s))
Claim 8: Bent discloses the method of claim 1, wherein the data structure corresponds to a content campaign from a demand side platform (DSP), wherein the at least one intended user action corresponds to an interaction with content associated with the content campaign. (Paragraphs 129 and 144: a content creation flow can be associated with a third party that provides a platform for content creators to generate customized content items (e.g., search results for display that link to a website, an advertisement, generated constructed content items).)
Response to Arguments
Applicant's arguments filed May 28, 2026 have been fully considered but they are not persuasive.
The applicant asserts that the claim amendment has overcome the 35 USC 112(a) rejections. The examiner disagrees. The amendment has only overcome the 112(a) rejections directed towards the term “serializing”. As detailed in the rejection above, it did not overcome the other 112(a) rejections raised in the Office Action dated March 4, 2026 and, has introduced new 112(a) rejections.
The applicant asserts that the claims overcome the 35 USC 101 rejection under Step 2a, Prong 2 because, when considered as a whole and in light of the specification, they recite a specific technical improvement to the functioning of computer-implemented content curation systems – namely, a particular pipeline by which structed facet data is converted, via natural language processing, into a large language model prompt; the LLM output is then analyzed by a machine learning model; and the underlying data structure is automatically modified in two specifically defined ways when a performance threshold is not satisfied. The examiner disagrees. It appears that the applicant is misconstruing both MPEP 2106 and the Desjardins decision. Ex parte Desjardins has not been incorporated into MPEP 2106.04(d). In fact, the Desjardins decision did not change the way in which MPEP 2106 has always required a 101 analysis to be performed. In order to overcome a 101 rejection under Step 2a, Prong 2, the improvement must be rooted in the additional elements of the claims in a manner other than merely applying the abstract idea using the “additional elements” of the claim as tool. “Additional elements” are defined as those elements which are not part of the abstract idea itself. In the instant claims, the “additional elements” are a device with a processor executing instructions including a large language model (LLM) and a machine learning (ML) model which is merely a general-purpose computer with generic computer components. The claims merely require that the identified abstract idea be applies using said general-purpose computer with generic computer components (i.e., device with a processor executing instructions including a large language model (LLM) and a machine learning (ML) as a tool which is incapable of transforming an abstract idea into a practical application under Step 2a, Prong 2 and/or incapable of being considered significantly more under Step 2b. Thus, the instant clams bear no resemblance to the claims of the Desjardins decision and, instead, are much closely aligned to the claims of the Recentive Analytics decision. The Desjardins decision and Recentive Analytics decision are both precedential decisions. As such, when analyzing claims under MPEP 2106 both decisions must always be true. In the Recentive Analytics decision we learn that the disclosure of an "already available [technology] with [its] already available basic functions, to use as [a] tool[] in executing the claimed process" is still an abstract idea; claims that do no more than apply established methods of machine learning to a new data environment are not patent eligible; and the requirements that the machine learning model be “iteratively trained” or dynamically adjusted in the Machine Learning Training patents do not represent a technological improvement. As such, the instant claims which merely use a generic large language model and a generic machine learning model as tools to implementing a content curation process are not a technological improvement. Whereas, in the Desjardins decision we find that claims which recite a applicant invented machine learning model, whose inner workings, operate in a manner different from traditional machine learning models is an improvement in technology because it is not a generic machine learning model. In traditional machine learning models, when a model is trained to perform a desired task, it uses the training data to generate a first algorithm for accomplishing the desired task. When said machine learning model is then trained to perform a second task, it discards the first algorithm and generates an new algorithm for accomplishing the second task. This discarding of the first algorithm is called catastrophic forgetting. In the claims of the Desjardins decision, the applicant invented a new machine learning model which, internally, does not discard the first algorithm. Instead, when trained on a second task, it generates a new algorithm, “while protecting performance of the machine learning model on the first machine learning task”, thereby, overcoming the technical problem of catastrophic forgetting. As such, the claims in the Desjardin invented a new machine learning model which, internally, worked in a manner different from traditional machine learning model. This newly invented machine learning model was an technological improvement that overcame the 101 rejections. The decision is very similar to the Enfish decision. In Enfish, the applicant invented a new type of database called a self-referential database that operated in a manner different from traditional database and, as such, was an “additional element” of the claims that resulted in an improvement that overcame the 101 rejections. In both of these decisions, the claims recited an “additional element” that was a new technology (i.e., new type of machine learning model and new type of database) that overcame the 101 rejections. In the instant claims, the “additional elements” of the claims do not represent any such new technologies. Instead, the recite a generic large language model and a generic machine learning model which are merely used as tools to implement the abstract idea. Thus, any purported improvement obtained from practicing the claimed invention is an improvement rooted solely in the abstract idea which is an improvement in ineligible subject matter (see MPEP 2106.05(a) - “It is important to note, the judicial exception alone cannot provide the improvement”; and MPEP 2106.05(a)(II) - “However, it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology”; and the SAP v Investpic decision - Page 2, line 22 through Page 3, line 13 - Even assuming that the algorithms claimed are groundbreaking, innovative or even brilliant, the claims are ineligible because their innovation is an innovation in ineligible subject matter because they are nothing but a series of mathematical algorithms based on selected information and the presentation of the results of those algorithms. Thus, the advance lies entirely in the realm of abstract ideas, with no plausible alleged innovation in the non-abstract application realm. An advance of this nature is ineligible for patenting; and Page 10, lines 18-24 - Even if a process of collecting and analyzing information is limited to particular content, or a particular source, that limitations does not make the collection and analysis other than abstract.). Finally, nothing the in 2024 AI SME update can be said to support the applicant’s position that claims reciting specific architectural, or operational improvements to an abstract idea such as AI/ML pipelines are eligible. The 2024 AI SME update allows an expert in the field to support the contention that an AI/machine learning model invention recites a newly invented AI/machine learning model that is an improvement over traditional AI/machine learning models. Affidavits from a SME that merely assert that an improvement is achieved to specific AI/ML pipeline by applying the “abstract idea” using a general purpose-computer executing instructions including a large language model (LLM) and a machine learning (ML) model as a tool will be unconvincing. In order for such an affidavit to overcome a 101 rejection, the affiant must indicate that the improvement is rooted in the “additional elements” of the claim. Otherwise, the affiant is merely swearing to the fact that the abstract idea is an improved abstract idea, which is an improvement ineligible subject matter. According to MPEP 2106.05(a), MPEP 2106.05(a)(II), and the SAP v. InvestPic decision (see citations above), improvements of this nature are not improvements in technology, incapable of transforming an abstract idea into a practical application under Step 2a, Prong 2 and incapable of being considered “significantly more” under Step 2b. As such, the applicant’s arguments are not convincing and the rejections have been maintained.
The applicant asserts that the claims overcome the 35 USC 101 rejection under Step 2b because they describe a specific multi-stage technical architecture – facet derivation, NLP-based prompt construction, LLM inference, ML-based effectiveness analysis, conditional restructuring of the underlying campaign (i.e., data structure), and network distribution. The examiner disagrees. In order to overcome a 101 rejection under Step 2b, the “additional elements” of the claim must be considered ‘significantly more”. A claim in which the “additional elements” are merely used as a tool to apply the abstract idea are incapable of being considered “significantly more” under Step 2b. What the applicant calls a specific multi-stage architecture is the abstract idea itself – facet derivation, NLP-based prompt construction, inference, effectiveness analysis, conditional restructuring of the underlying campaign (i..e., data structure), and network distribution. The LLM is a generic LLM merely used as a tool to perform the inferences. The ML is a generic machine learning model merely used as a tool to perform the effectiveness analysis. As such, it is clear that the applicant is merely claiming the performance of an abstract idea using a general purpose-computer executing instructions including a generic large language model (LLM) and a generic machine learning (ML) model as a tool which is insufficient to be considered “significantly more” under Step 2b. Thus, the applicant’s arguments are not convincing and the rejections have been maintained.
In regards to the 35 USC 102 rejection, the applicant argues that Bent does not disclose determining, based on historical facet data, a set of facets that includes at least one of a media type or a media format and that corresponds to intended activity on a network”. The examiner disagrees. The applicant appears to be reading the claims in a much narrower manner than required by the claims as currently written. The examiner notes that the broadest reasonable interpretation of the claim limitation does not require determining, based on historical facet data, a set of facets that includes at least one media type or a media format. The claim, as currently written, requires determining, based on historical facet data, “a set of facets related to the data structure” and “at least one of a media type or a media format”. Thus, as currently written, the determined set of facets is different from the determined at least one of a media type or a media format. As such, the set of facets cannot be said to include the at least one media type or media format as argued. Likewise, the applicant appears to place a much narrow interpretation of terms such as “historical facet data” and “set of facets”. The applicant’s specification does not define these terms and, as such, the common dictionary definitions of these terms have been used. The Merriam-Webster Online Dictionary defines a facet as: a particular part or aspect (as of something being contemplated or considered). Thus, the “set of facets” is merely a particular part or aspect of the content campaign (i.e., data structure). The Merriam-Webster Online Dictionary defines historic as: known or established in the past. Thus, the historic facet data is any data regarding the current content campaign or other content campaigns which was known or established prior to the determining step occurring. Thus, given the broadest reasonable interpretation of the claims, Bent must merely disclose determining, based on some type of previously known data, “a set of facets related to the content campaign (i.e., data structure)” and, also determining “at least one of a media type or a media format”. Bent clearly discloses obtaining “a set of facets related to the data structure based on historic facet data in at least Fig. 9-10 and Paragraph 107-117 and 152-154, where the content assistant component can be a progressive disclosure field or a conversational interface where the user first provides original facet data (historic facet data), then in response to a query from the content assistant provides additional facet information (historic facet data); the computing system extracts information from the user input including business name and descriptive terms and phrases; the user then provides user input indicative of a desire for modifications of the suggested content (historic facet data), wherein after the user provides the last desired modification (current facet data) and input indication of a satisfactions. Bent further discloses in Paragraph 52 that the user input data can be associated with a current user session and/or include historical user data. As such, Bent clearly discloses determining, based on historical facet data, “a set of facets related to the data structure”. Furthermore, Bent discloses in at least Figure 4 and Paragraph 103 the user provides input regarding the type of new content item and the content item group they wish to have created (e.g., advertisement, advertisement campaign), and in Paragraph 52 that the user input data can be associated with a current user session and/or include historical user data. As such, Bent is clearly disclosing determining, based on historical facet data “at least one of a media type or a media format”. Additionally, Bent discloses obtaining historic facet data from content providers (e.g., advertisers) websites and web documents (search results, advertisements, particular businesses) so the user does not need to manually input such data in at least paragraphs 41, and 95-97. Bent also discloses inputting audio and visual data in at least paragraphs 74, 87, 101, 131, 158, and 167. Therefore, given that the information obtained in Bent includes images and videos, it would be completely inaccurate to describe Bent as deriving it’s input from merely user-supplied natural language (e.g., a user describing their business in a chat interface) and data scraped from a landing page. As such, it is clear that Bent discloses the limitations of the claims, as currently written, and the rejections have been maintained.
In regards to the 35 USC 102 rejection, the applicant argues that Bent does not disclose generating, based on the set of facets, a large language model prompt, the prompt generation comprising processing, via a natural language processing model the set of facets into plain-text facet data. The examiner disagrees. First, the fact that the user is the current or previous provider of the natural language facets does not preclude the invention of Bent from processing the natural language facets into plain-text facet data. The invention of Bent, as indicated in paragraph 70, the NLP processes these natural language facets to generate a textual segmentation output which would be plain-text facet data; generate a semantic intent output comprise at least one word or phrase determined from the text or natural language data which would also be plain-text facet data; generate a translation output which also be a plain-text facet data; and/or generate an upscaled text or natural language output (e.g., text or natural language data that is higher quality than the input text or natural language, etc.). Which is the exact same type of plain text facet data disclosed in paragraphs 53-55 of the applicant’s specification. However, the claim as currently written would allow for the processing of other facets into plain-text facets. For example, the input in Bent need not be text-based natural language input by the user. Bent discloses the input being speech input, which is then converted into plain-text facet data as disclosed in at least paragraph 71. Additionally, as indicated in paragraph 52 the user input data can be associated with a current user session and/or include historical user data. Thus, the processing of the set of facets into plain-text facet data might include processing of currently inputted data and stored historic user input data which is taught by Bent. It might include processing historic user information associated with a plurality of user identifiers to aggregate the historic data when generating plain-text facet data, and formulating the LLM prompt using this data which is also taught by Bent. As such, it is clear that Bent discloses generating, by the device, based on the set of facets, a large language model (LLM) prompt, the LLM prompt generation comprising processing, via a natural language processing (NLP) model, the set of facets into plain-text facet data. Thus, the applicant’s arguments are not convincing and the rejections have been maintained.
In regards to the 35 USC 102 rejection, the applicant argues that Bent does not disclose “curating…the content campaign (i.e., data structure)”. The examiner disagrees. It appears that the applicant is reading limitations into the claims which are not required by the claim language itself. The Merriam-Webster Online Dictionary defines curate as to select and/or organize especially for presentation, distribution, or publication. As such when Bent discloses in at least paragraph 112, 118, and 137 that the computing system can generate suggested content such as a summary of a strategy for an advertisement campaign comprising one or more advertisements, requirements for display, time for display, expected performance, etc.; obtaining user input regarding the suggestions; and upon user approval, generating the advertisement which may comprise a plurality of content items such as headlines, descriptions, videos, images, taglines, etc. his is clearly “curating…the content campaign (i.e., data structure) and the limitations of the claims as currently written have been met. Thus, the applicant’s arguments are not convincing and the rejections have been maintained.
In regards to the 35 USC 102 rejection, the applicant argues that Bent does not disclose that when the performance threshold is not satisfied, modifying the data structure by changing a format of the digital content within the content campaign (i.e., data structure) and changing the quantity of content items within the content campaign (i.e., data structure). The examiner disagrees. The cited sections of Bent in the rejection above (i.e., Paragraphs 113-114, 172-173, 181, 196-197 and 218-221, as well as figures 18-20) discloses that the modification of the digital content (e.g., advertisement) continues until a predicted score of the generated content exceeds a threshold. Each time modified digital content is created, a content score for suggested content is determined and the suggested content is displayed along with recommendations for improving the score, and additional inputs are obtained, and the process repeats. The types of modifications include displaying additional content items such as images. The invention of Bent can automatically adjust the parameters of the digital content including adjusting parameters of the content based on performance metrics, wherein parameters include adding or removing content elements (e.g., descriptions, keywords, images, audio, video). As these modifications include adding or removing content elements, Bent clearly discloses the quantity of content items is changed within the digital content, when the performance threshold is not satisfied. Additionally, changing the quantify of content items by adding or removing content elements such as descriptions, keywords, images, audio, and/or video) changes the format of the digital content as made clear from at least figures 18-20 where adding the image in figure 20, reformats that way in which the text is displayed such that before the image was added the text spanned the horizontal distance of the advertisement, and after the image was added the text now only spans about half of the distance of the advertisement. As such, it is clear that Bent discloses that when the performance threshold is not satisfied, modifying the data structure by changing a format of the digital content within the content campaign (i.e., data structure) and changing the quantity of content items within the content campaign (i.e., data structure) and the limitations of the claims as currently written have been met. Thus, the applicant’s arguments are not convincing and the rejections have been maintained.
In regards to the 35 USC 102 rejection, the applicant argues that Bent does not disclose communicating, over the network, the curated data structure to a set of network resources. The examiner disagrees. Bent clearly discloses in at least paragraphs 84 and 164 that the evaluation component performs pilot and/or live traffic experiments to capture online metrics with regards to the suggested content which means the curated data structure is communicated to a set of network resources. As such, this limitation of the claims is taught by Bent. Therefore, the applicant’s arguments are not convincing and the rejections have been maintained.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Agrawal et al. (PGPUB: 2024/0312087) which discloses using a large language model to generate custom content for an advertising campaign (e.g., data structure) based on obtained product images, target themes, and target guidelines, wherein system selects a product image, modifies the product image to be consistent with a target theme and target guidelines and appends text appropriate for the image, target theme and target guidelines.
Data Science Dojo, LLM Use-Cases: Top 10 industries that can benefit from using large language models, September 15, 2023, https://medium.com/@data sciencedojo/llm-use-cases-top-10-industries-that-can-benefit-from-using-large-language-models-data-science-e3018f098d8#:~:text=A%20large%20language %20 model%20can,recommendations%20for%20products%20and%20services, pgs. 1-21 which discloses the use of Large Language Models in generating customized marketing content and measuring the effectiveness of marketing campaigns.
Reisenbichler et al., Applying Large Language Models to Sponsored Search Advertising, August 15, 2024, https://thearf-org-unified-admin.s3.amazonaws. com/MSI_Report_23-136.pdf, pgs. 1-24 which discloses the use of Large Language Models to create customized content that is optimized for search engine advertising, wherein both the created content and the bids associated with the created content are optimized for effectiveness.
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 JOHN W VAN BRAMER whose telephone number is (571)272-8198. The examiner can normally be reached Monday-Thursday 5:30 am - 4 pm EST.
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, Spar Ilana can be reached on 571-270-7537. 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.
/John Van Bramer/Primary Examiner, Art Unit 3622
Prosecution Insights