DETAILED ACTION
This communication is in response to the Application filed on April 22, 2026. Claims 1-20 are pending and have been examined.
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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on March 23, 2026 has been entered.
Response to Arguments
The reply filed on March 23, 2026 has been entered. Applicant’s arguments with respect to claims 1-20 have been considered but are not persuasive/moot in view of new ground(s) of rejection caused by the amendments.
With respect to the applicant’s arguments to claim rejections under 35 U.S.C § 101, Applicant has amended each of the independent claims and asserts that “Applicant again submits that the subject matter recited by amended independent claims 1, 16 and 20 is not something which can be practically performed in the human mind as required per the Memorandum titled "Reminders on evaluating subject matter eligibility of claims under 35 U.S.C. 101" released by the USPTO on August 4, 2025. It is not practical for a human mind to generate the candidate output or the subsequent candidate output which would be generated by a large language model based on the claimed inputs. The human mind cannot practically replicate the tokenization, embedding and probabilistic determinations performed by the claimed large language model.” The examiner respectfully disagrees with these assertions. The applicant claims that a human mind cannot practically replicate certain features of a large language model such as tokenization, embedding and probabilistic determinations, but the examiner fails to see such elements claimed within the amended claim language itself. The claimed “large language model” is not accompanied by any further claimed processes such as tokenization, embedding, or probabilistic determinations, and as such, it is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). When viewing the amended claim language under the broadest reasonable interpretation, the recited “large language model” is merely used as a tool to generate candidate prompts from input data. This limitation cannot amount to anything more than an additional element because of the generality in which it is recited.
Applicant further asserts that “Applicant submits that the improvement in machine learning technology actually arises from using a large language model to generate a prompt for itself (or a different target large language model in some embodiments), which leverages the large language model to iterate upon prompts for itself … the skilled artisan will appreciate that different large language models may generate different outputs even when receiving the same prompt and that one particular large language model may generate different outputs at different times even when receiving the same prompt. Human prompt engineers may not understand or appreciate why such differences occur (see paragraph [0003] of the specification as filed). A skilled artisan will also appreciate that this is an inherent technical issue of large language models.” The examiner respectfully disagrees with these assertions. First, the applicant never directly states what “technical issue” iterative prompting is actually solving. It can be inferred from the Applicant’s remarks that the technical issue lies in creating good prompts for an LLM to process data, but such an activity is considered an abstract idea. As per MPEP 2106.05(a), “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. For example, in Trading Technologies Int’l v. IBG, 921 F.3d 1084, 1093-94, 2019 USPQ2d 138290 (Fed. Cir. 2019), the court determined that the claimed user interface simply provided a trader with more information to facilitate market trades, which improved the business process of market trading but did not improve computers or technology.” Accordingly, iteratively generating prompts for a large language model using a feedback loop is considered an abstract idea because a human can just as similarly brainstorm prompts for a large language model and improve the prompts overtime based on a second human reviewing and ranking said prompts and their associated outputs. Improvement in an abstract idea cannot be considered equivalent to an improvement in machine learning technology, as stated above by MPEP 2106.05(a). As amended, there is no language in the independent claims that would prevent a human from performing these steps, as addressed in further detail below with respect to claim rejections under 35 USC § 101.
With respect to the applicant’s arguments to claim rejections under 35 U.S.C § 103, the applicant’s arguments with respect to claims 1-20 have been considered but are moot in view of new ground(s) of rejection caused by the amendments.
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 an abstract idea without significantly more. All of the claims are method claims (1-15), apparatus/machine claims (16-20) or manufacture claim under (Step 1), but under Step 2A all of these claims recite abstract ideas and specifically mental processes. These mental processes are more particularly recited in claims 1, 16, and 20 as:
inputting at least the candidate prompt into a large language model…
receiving, from the large language model and responsive to input of at least the candidate prompt, at least one candidate output generated by the large language model…
outputting the candidate prompt and the at least one candidate output…
receiving user input directed to one or both of the candidate prompt and the at least one candidate output…
inputting at least the user input and one or both of the candidate prompt and the at least one candidate output into the large language model…
receiving, from the large language model and responsive to input of at least the user input and one or both of the candidate prompt and the at least one candidate output, a subsequent candidate prompt generated by the large language model…
Under Step 2A Prong One, claims 1, 16, and 20 are directed to an abstract idea and specifically a mental process. As detailed above, the steps of inputting, receiving, outputting, etc. may be practically performed in the human mind with the use of a physical aid such as a pen and paper. For example, a human could receive instructions from their boss to process company data, brainstorm multiple summaries of various writing styles, receive input from their boss selecting a particular summary, and write a contextual instruction for summarizing the company data that integrates the boss’ preferences based off the chosen summary.
Under Step 2A Prong Two, this judicial exception is not integrated into a practical application because claims 1-20 do not recite additional elements that integrate the exception into a practical application. In particular, claims 1, 16, and 20 recite the additional elements of a computer (¶ [0072]), a processor (¶ [0067]), memory storing instructions (¶ [0148]), and a large language model (¶ [0055]). These additional elements are recited at a high level of generality and merely equate to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). Further, claims 1, 16, and 20 recite the additional elements of “obtaining…” which amounts to insignificant extra-solution activities which are not indicative of integration into a practical application as per MPEP 2106.05(g). Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Under Step 2B, the claims do not recite additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements of using a computer is noted as a general computer {computer (¶ [0072]); processor (¶ [0067]); memory storing instructions (¶ [0148]); large language model (¶ [0055])}. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. Further, the additional limitations in the claims noted above are directed towards insignificant extra-solution activities. The claims are not patent eligible.
With respect to claims 2, the claim relates to user input comprising a score. This relates to a boss rating each of the summary reports produced by the human. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 3, the claim relates to user input comprising a comment or flag. This relates to a boss providing feedback for each summary report produced by the human. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 4, the claim relates to inputting the subsequent candidate prompt back into the large language model, and then receiving a subsequent output based off of the subsequent candidate prompt. This relates to a human following their created instructions on how they would process the data, and then producing a modified summary report based on those instructions. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 5, the claim relates to receiving subsequent user input directed towards the subsequent candidate output, and using both the subsequent user input and subsequent candidate output to produce further subsequent candidate prompts. This relates to a human receiving further comments from their boss directed towards their modified summary report, integrating their boss’s feedback into the report, and then creating further instructions about how they would better process the data. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 6, the claim relates to inputting the original candidate prompt into the model responsive to receiving the subsequent candidate prompt. This relates to a human re-processing the data in accordance to their boss’ original instructions. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 7 and 17, the claim relates to inputting iterative candidate prompts into the large language model. This relates to a human receiving from their boss iterative instructions for processing the data. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 8 and 18, the claim relates to receiving and inputting iterative user input into the large language model. This relates to a human receiving from their boss iterative feedback for each summary report they create in quick succession, and generating iterative reports based on the boss’ feedback. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 9, the claim relates to inputting prompt templates into a large language model. This relates to a human formatting their boss’ instructions into an easily followable TODO list, and then utilizing the TODO list to create summary reports about the input data. The additional element of a “large language model” is recited at a high level of generality (¶ [0055]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 10, the claim relates to the prompt template containing a prompt region and an output region. This relates to a human creating, using instructions from their boss, an instruction form template which utilizes a dedicated instructions region for the boss’ instructions, an input data region for the input data, and a dedicated output region for the generated summary report. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 11, the claim relates to the outputting the candidate prompt and candidate output to a user device. This relates to a human presenting their final summary report to their boss. The additional elements of a “user device” and “display” are recited at a high level of generality {user device (¶ [0071]); display (¶ [0069])} and merely equate to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 12 and 19, the claim relates to storing a plurality of nodes and edges related to the prompt engineering method. This relates to a human drawing a graph where each node represents a particular iteration of the boss’ instruction and their corresponding summary report. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 13, the claim relates to displaying a progress region to the user that comprises the plurality of nodes and edges. This relates to a human presenting their hand-drawn graph to their boss in order to show them the iterative process of creating the summary report. The additional elements of a “user device” and “display” are recited at a high level of generality {user device (¶ [0071]); display (¶ [0069])} and merely equate to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 14, the claim relates to displaying a particular prompt and associated output when selected by the user. This relates to a human receiving a request from their boss to view a particular iteration and then displaying the boss’ entire instruction and corresponding summary report side-by-side for that particular iteration. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 15, the claim relates to receiving user input directed at a plurality of candidate outputs. This relates to a human generating a plurality of summary reports, to which the boss selects their favorite summary report out of the plurality of summary reports. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
For all of the above reasons, taken alone or in combination, claims 1-20 recite a non-statutory mental process.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 3, 9, 11, 15, 16, and 20 are rejected under 35 U.S.C. 103 as obvious over US Patent Publication 20240303424 A1 (Reddy et al.) embodiment 1 in view of US Patent 12424209 B1 (Guo et al.).
Claim 1
Regarding claim 1, Reddy et al. embodiment 1 disclose a computer-implemented method comprising:
obtaining a candidate prompt including input data and candidate instructions for processing the input data (Reddy et al. ¶ [0081]-[0082], "User interface 307 transmits the user input to prompt engine 305 which generates a prompt based on the input. ... Prompt engine 305 also includes contextual information in the prompt. ... For example, for a general inquiry about the contents of data table 502, prompt engine 305 configures the prompt to include all the column headers in data table 502 along with the table name and the first five rows of data");
inputting at least the candidate prompt into a large language model (Reddy et al. ¶ [0083], "Continuing with FIG. 5B, with the prompt configured, prompt engine 305 sends the prompt to LLM 330.");
receiving, from the large language model and responsive to input of at least the candidate prompt, at least one candidate output generated by the large language model (Reddy et al. ¶ [0083], "LLM 330 generates a reply to the prompt and transmits the reply to prompt engine 305.");
outputting the candidate prompt (Reddy et al. ¶ [0089], “Task pane 503 displays chat history 505 that has accumulated thus far, and which may be submitted by prompt engine 305 for context in any subsequent requests." See Figure 5G, which shows that the entire conversation history is output to the user. Chat history, when used as context by prompt engine 305 for subsequent requests, is considered analogous to a candidate prompt. Thus, Outputting the accumulated chat history as context for subsequent requests is considered analogous to outputting the candidate prompt) and the at least one candidate output (Reddy et al. ¶ [0083], "Prompt engine 305 processes the reply, including extracting the natural language explanation enclosed in the appropriate tags, and generates a response for display by user interface 307 in task pane 503.");
receiving user input directed to one or both of the candidate prompt and the at least one candidate output (Reddy et al. ¶ [0084], "Next, in FIG. 5C, the user submits a second input in task pane 503 requesting suggestions for making “it” better.");
inputting at least the user input and one or both of the candidate prompt and the at least one candidate output into the large language model (Reddy et al. ¶ [0083]-[0084], "LLM 330 generates a reply to the prompt and transmits the reply to prompt engine 305. … Next, in FIG. 5C, the user submits a second input in task pane 503 requesting suggestions for making “it” better. ... Prompt engine 305 adds context information to the prompt: relevant spreadsheet data, such as the column headers and some rows of data, and the chat history thus far." Context information is considered analogous to the candidate prompt. For example, output chat history is considered analogous to a candidate prompt since it is added as context information to a prompt for subsequent requests. The LLM-generated reply received in ¶ [0083] is considered analogous to the at least one candidate output.); and
receiving, from the large language model and responsive to input of at least the user input and one or both of the candidate prompt and the at least one candidate output, a subsequent candidate prompt generated by the large language model (Reddy et al. ¶ [0085], "Continuing operational scenario 500 in FIG. 5C, prompt engine 305 submits the prompt to LLM 330. Upon receiving the prompt, LLM 330 generates three suggestions" Suggestions are considered analogous to a subsequent candidate prompt), wherein the subsequent candidate prompt includes [the input data and] modified candidate instructions for processing the input data which are different from the candidate instructions (Reddy et al. Figure 5C illustrates the three suggestions, each of which differ from the original candidate instructions ("What is this data about?")).
Reddy et al. embodiment 1 do not explicitly disclose all of the subsequent candidate prompt including the original input data.
However, Guo et al. disclose obtaining a candidate prompt including input data (Guo et al. ¶ (33), "The plan prompt generation component 140 may further receive context data 205 representing various contextual signals associated with the user input data 127, such as weather information, time of day, device information associated with the device that sent the user input data 127 (e.g., device ID, device states, historical device interaction data, etc.)." Context data 205 is considered analogous to a candidate prompt) and candidate instructions for processing the input data (Guo et al. ¶ (30), "As illustrated in FIG. 1, the user input data 127 may be received by the LLM orchestrator component 130 at the plan generation component 135, which may be configured to generate (e.g., using the plan generation language model 145) a list (e.g., one or more) of tasks (e.g., steps/sub-actions) that are to be completed in order to perform an action responsive to the user input and select (e.g., using the task selection language model 155) a task of the list of the tasks that is to be completed first (e.g., in a current iteration of processing by the system 100), as described in detail herein below with respect to FIG. 2." User input 127 is considered analogous to candidate instructions);
inputting at least the candidate prompt into a large language model (Guo et al. ¶ (33), "the plan prompt generation component 140 processes the user input data 127 to generate prompt data 220 representing a prompt for input to the plan generation language model 145." ¶ (21), "The language models 145, 155, 180 are large models. In some embodiments, the language models 145, 155, 180 may be a LLM.") ;
receiving, from the large language model and responsive to input of at least the candidate prompt, at least one candidate output generated by the large language model (Guo et al. ¶ (44), "The plan generation language model 145 processes the prompt data 220 to generate model output data 225 representing one or more predicted tasks to be completed in order to perform the action responsive to the user input.");
outputting the candidate prompt and the at least one candidate output (Guo et al. ¶ (45), "The model output data 225 is sent to the task selection prompt generation component 150, which processes the model output data 225 to generate prompt data 230 representing a prompt for input to the task selection language model 155. In some embodiments, such prompt data 230 may be generated based on combining the user input data 127, the context data 205, the personalized context data 215, the prompt data 220, and/or the model output data 225." Context data 205 is considered analogous to a candidate prompt. Model output data 225 is considered analogous to a candidate output); …
inputting at least [the user input and] one or both of the candidate prompt and the at least one candidate output into the large language model (Guo et al. ¶ (45), "The model output data 225 is sent to the task selection prompt generation component 150, which processes the model output data 225 to generate prompt data 230 representing a prompt for input to the task selection language model 155." ¶ (21), "The language models 145, 155, 180 are large models. In some embodiments, the language models 145, 155, 180 may be a LLM."); and
receiving, from the large language model and responsive to input of at least the user input and one or both of the candidate prompt and the at least one candidate output, a subsequent candidate prompt generated by the large language model (Guo et al. ¶ (51)-(52), "The task selection language model 155 processes the prompt data 230 to generate model output data representing the task to be completed first and/or a prioritization of the one or more tasks. … the plan generation component 135 (or another component of the plan generation component 135) may process the model output data of the task selection language model 155 to determine task processing data 157 … The shortlister prompt generation component 175 processes the task processing data 157 to generate prompt data 320 representing a prompt for input to the shortlister language model 180." Prompt data 320 is considered analogous to a subsequent candidate prompt), wherein the subsequent candidate prompt includes the input data and modified candidate instructions for processing the input data (Guo et al. ¶ (51-52), "the plan generation component 135 (or another component of the plan generation component 135) may process the model output data of the task selection language model 155 to determine task processing data 157 representing the user input data 127, the context data 205, the personalized context data 215, and/or the task selected by the task selection language model 155 to be completed first. … prompt data 320 may be generated based on combining the task processing data 157… and relevant API data 310 representing one or more APIs associated with the user input data 127 and/or the current task." Context data 205 is considered analogous to input data. Prompt data 320 is considered analogous to modified candidate instructions.) which are different from the candidate instructions (Guo et al. Figures 2 and 3 illustrate user input 127 (candidate instructions) and prompt data 320 (modified candidate instructions) respectively).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Reddy et al. embodiment 1’s prompt generation method to incorporate Guo et al.’s inclusion of input data in the subsequent candidate instruction because such a modification is the result of simple substitution of one known element for another producing a predictable result. More specifically, Reddy et al. embodiment 1’s subsequent candidate instruction and Guo et al.’s subsequent candidate instruction perform the same general and predictable function, the predictable function being specifying a task for downstream data processing. Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject matter and the prior art rests not on any individual element or function but in the very combination itself - that is in the substitution of Reddy et al. embodiment 1’s subsequent candidate instruction by replacing it with Guo et al.’s subsequent candidate instruction. Thus, the simple substitution of one known element for another producing a predictable result renders the claim obvious.
Claim 3
Regarding claim 3, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 further disclose wherein the user input further comprises at least one of a comment associated with one or both of the candidate prompt and the at least one candidate output (Reddy et al. ¶ [0084], "Next, in FIG. 5C, the user submits a second input in task pane 503 requesting suggestions for making “it” better."); or a flag associated with one or both of the candidate prompt and the at least one candidate output, and wherein inputting the user input into the large language model comprises:
inputting at least one of the comment or the flag into the large language model (Reddy et al. ¶ [0084], "For an ambiguous or underspecified input, the prompt template may include an instruction to LLM 330 to interpret the input in multiple ways and to generate multiple alternative suggestions based on the interpretations. Here, “it” may refer to, for example, the formatting of data table 502, the organization of data table 502, to the contents of data table 502, or to some other aspect of data table 502.").
Claim 9
Regarding claim 9, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 further disclose further comprising a candidate template (Reddy et al. ¶ [0099], "To configure the prompt, prompt engine 305 identifies a prompt template according to the type of request in the input in an implementation. Prompt templates can include a prompt configuration for suggesting a calculated column to be added to workbook data 320, for a general inquiry about workbook data 320, for analyzing workbook data 320 to project a result, such as for a hypothetical scenario, and so on.") for iteratively receiving different candidate prompts including at least one of different candidate instructions or different input data (Reddy et al. ¶ [0036], “The chat interface displays a turn-based conversation through which the user can iterate or step through multiple revisions of the spreadsheet which are responsive to the user's inquiries” Using a prompt template to respond to queries inputted through the chat interface is considered analogous to using a candidate template for iterative prompting), and wherein obtaining the candidate prompt, inputting at least the candidate prompt into the large language model and receiving the at least one candidate output comprises:
iteratively inputting the candidate template including the different candidate prompts into the large language model (Reddy et al. Figures 5B-5F illustrate inputting iterative templates from the user interface to the LLM); and
iteratively receiving, from the large language model and responsive to the inputting of the candidate template, corresponding different at least one candidate outputs generated by the large language model (Reddy et al. Figures 5B-5F illustrate receiving outputs from the LLM corresponding to the inputted iterative prompt templates).
Claim 11
Regarding claim 11, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 further disclose wherein outputting the candidate prompt and the at least one candidate output comprises outputting for display, on a display of a user device, the candidate prompt and the at least one candidate output (Reddy et al.¶ [0023], "The application displays the reply provided by the LLM in a task pane in the user interface of the application. In some implementations, the application may display the reply with graphical input devices by which the user can interact with the reply" See Figure 5G, which illustrates displaying the entire chat history to the user via user interface).
Claim 15
Regarding claim 15, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 further disclose wherein the at least one candidate output comprises a plurality of candidate outputs (Reddy et al.¶ [0024], "the user may ask, “How can I get better results?” The application may include in its prompt to the LLM an instruction to interpret the inquiry in multiple ways and to generate suggestions based on the interpretations. The application presents the suggestions generated by the LLM to the user in the task pane.") and receiving the user input comprises receiving user input directed to different candidate outputs of the plurality of candidate outputs (Reddy et al.¶ [0024], "The application may submit a follow-up prompt based on the user's selection of a suggestion and including in the contextual information of the prompt the selection made by the user thereby to receive a more focused reply or suggestion from the LLM.").
Claim 16
Regarding claim 16, Reddy et al. embodiment 1 disclose a system, comprising:
at least one processor (Reddy et al.¶ [0110], "processing system 1002 may comprise a micro-processor and other circuitry that retrieves and executes software 1005 from storage system 1003."); and
a memory storing processor-executable instructions (Reddy et al.¶ [0110], "processing system 1002 may comprise a micro-processor and other circuitry that retrieves and executes software 1005 from storage system 1003.").
The remaining limitations of claim 16 are similar in scope to the limitations of claim 1 and therefore are rejected for similar reasons as described above.
Claim 20
Regarding claim 20, Reddy et al. embodiment 1 disclose a non-transitory computer-readable storage medium having stored thereon processor-executable instruction (Reddy et al.¶ [0110], "processing system 1002 may comprise a micro-processor and other circuitry that retrieves and executes software 1005 from storage system 1003.”).
The remaining limitations of claim 20 are similar in scope to the limitations of claim 1 and therefore are rejected for similar reasons as described above.
Claim 2 is rejected under 35 U.S.C. 103 as obvious over Reddy et al. embodiment 1 in view of Guo et al. as applied to claim 1 above, and further in view of US Patent Publication 20240346254 A1 (Liu et al.).
Claim 2
Regarding claim 2, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 in view of Guo et al. do not explicitly disclose all of user input comprising a score.
However, Liu et al. disclose wherein the user input comprises a score of one or both of the candidate prompt and the at least one candidate output (Liu et al. ¶ [0047], “The external source and/or user can subsequently provide external feedback 346 rating various aspects of the third natural language output 344 such as coherence, fluency, and so forth.” ¶ [0041], "external feedback 230 can include a numerical score for various aspects of the natural language output 216 such as readability, coherence, and language fluency."), and wherein inputting the user input into the generative language model comprises:
inputting the score into the generative language model (Liu et al. ¶ [0047], “The external feedback 346 can be analyzed by the natural language generation system 304 and/or the large language model 302 to modify various parameters to improve both augmentation and training over time.”).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Reddy et al. embodiment 1 in view of Guo et al. to incorporate Liu et al.’s numerical user scoring.
The suggestion/motivation for doing so would have been that, “The external feedback 346 can be analyzed by the natural language generation system 304 and/or the large language model 302 to modify various parameters to improve both augmentation and training over time,” as noted by the Liu et al. disclosure in paragraph [0047].
Claims 4-6 are rejected under 35 U.S.C. 103 as obvious over Reddy et al. embodiment 1 in view of Guo et al. as applied to claim 1 above, and further in view of Reddy et al. embodiment 2.
Claim 4
Regarding claim 4, Reddy et al. embodiment 1 in view of Guo et al. disclose all the elements of the claimed invention as stated above.
Reddy et al. embodiment 1 in view of Guo et al. does not explicitly disclose all of inputting the subsequent candidate prompt back into the generative language model.
However, Reddy et al. embodiment 2 disclose inputting at least the subsequent candidate prompt back into the generative language model (Reddy et al. ¶ [0097]-[0098], "the user has selected one of the suggested actions of output 712 (“Break down of sales looking at subscription and newsletter”). The application receives the selection and configures a prompt based on the selection for submission to the LLM." Selecting a suggested action generated by the LLM is considered analogous to inputting the subsequent candidate prompt back into the LLM.); and
receiving, from the generative language model and responsive to input of the subsequent candidate prompt, at least one subsequent candidate output generated by the generative language model (Reddy et al. ¶ [0098], "The LLM generates and sends a reply to the prompt to the application which the application processes to produce a response to the input. The response to the input from the application based on the reply from the LLM is shown in output 715.").
It is recognized that the citations and evidence provided above are derived from potentially different embodiments of a single reference. Nevertheless, it 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 to employ combinations and sub-combinations of these complementary embodiments, because Reddy et al. explicitly motivates doing so at least in paragraph [0079], [0114] and [0119] including “those skilled in the art will also appreciate that the features described above may be combined in various ways to form multiple embodiments” and otherwise motivating experimentation and optimization. Additionally, doing so merely combines prior art elements according to known methods to yield predictable results.
Claim 5
Regarding claim 5, the rejection of claim 4 is incorporated.
Reddy et al. embodiment 2 further disclose receiving subsequent user input directed to one or both of the subsequent candidate prompt and the at least one subsequent candidate output (Reddy et al. ¶ [0098], "The user, in response to output 715, can select another of the suggested actions presented in output 717 or submit another input in textbox 713. Continuing the exemplary scenario, the user submits a natural language input in textbox 713, “How can I get better results?”");
inputting the subsequent user input and one or both of the subsequent candidate prompt and the at least one subsequent candidate output into the generative language model (Reddy et al. ¶ [0099], "Based on the input received from the user in textbox 713, the application generates a prompt for submission to the LLM."); and
receiving, from the generative language model and responsive to input of at least the subsequent user input and one or both of the subsequent candidate prompt and the at least one subsequent candidate output, a further subsequent candidate prompt generated by the generative language model (Reddy et al. ¶ [0099], "In FIG. 7C, task pane 710 is updated by the application to show input 716 and output 717 based on the LLM's reply to the prompt based on input 716. Output 717 includes three suggestions received from the LLM in response to input 716 according to parameters provided by the application in the prompt." Suggestions generated by an LLM are considered analogous to a subsequent candidate prompt), wherein the further subsequent candidate prompt includes further modified candidate instructions for processing the input data which is different from the modified candidate instructions (Reddy et al. Figure 7C illustrates the three suggestions, each of which differ from the modified candidate instructions).
Claim 6
Regarding claim 6, the rejection of claim 4 is incorporated.
Reddy et al. embodiment 2 further disclose reverting back to inputting the candidate prompt into the generative language model yielding the at least one candidate output (Reddy et al. ¶ [0099], "Based on the input received from the user in textbox 713, the application generates a prompt for submission to the LLM. The prompt includes the input or the substance of the input, along with contextual information such as the chat history and spreadsheet contextual data" Chat history includes the original candidate prompt.) responsive to receiving the at least one subsequent candidate output (Reddy et al. ¶ [0098]-[0099], "The user, in response to output 715, can select another of the suggested actions presented in output 717 or submit another input in textbox 713. ... Based on the input received from the user in textbox 713, the application generates a prompt for submission to the LLM.").
Claims 7, 8, 10, 17, and 18 are rejected under 35 U.S.C. 103 as obvious over Reddy et al. embodiment 1 in view of Guo et al. as applied to claims 1 and 16 above, and further in view of US Patent Publication 20250103621 A1 (Pan et al.).
Claim 7
Regarding claim 7, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 in view of Guo et al. do not explicitly disclose all of iterative prompting.
However, Pan et al. disclose iteratively inputting successive candidate prompts generated by the generative language model back into the generative language model yielding corresponding successive at least one candidate outputs (Pan et al. ¶ [0058], "In step 506, component 150 generates outputs based on the initial prompt and identified prompt variations. In various embodiments, component 150 generates and outputs prompt variations associated with the initial input from the user." Generating multiple outputs based on the initial prompt and variations of the initial prompt is considered analogous to iteratively inputting successive candidate prompts generated by a GLM).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Reddy et al. embodiment 1 in view of Guo et al. to incorporate Pan et al.’s iterative prompting.
The suggestion/motivation for doing so would have been that, “it adapts to the user's desire for iterative improvements in the outcomes and helps guide these refinements effectively,” as noted by the Pan et al. disclosure in paragraph [0014].
Claim 8
Regarding claim 8, the rejection of claim 7 is incorporated.
Pan et al. further disclose iteratively receiving user input directed to one or more of the successive candidate prompts and the corresponding successive at least one candidate outputs (Pan et al. ¶ [0059], "In step 508, component 50 records and retrieves iteration feedback from a user. ... Component 150 may collect the iteration feedback and utilizes the iteration feedback to update the heuristic to generate and organize the prompt variations."); and
iteratively inputting the user input and one or more of the successive candidate prompts and the corresponding successive at least one candidate outputs back into the generative language model to generate further candidate prompts (Pan et al. ¶ [0060], "In step 510, [component 150] updates the heuristic based on the iteration feedback from the user. ... The heuristic may evolve using interaction data, wherein component 150 utilizes the heuristic evolution to improve prompt variation recommendations." Improving prompt variation generation using a heuristic that evolves overtime from user interactions is considered analogous to iteratively inputting user input into a GLM to generate further candidate prompts).
Claim 10
Regarding claim 10, the rejection of claim 1 is incorporated.
Reddy et al. further disclose a prompt region for inputting the different candidate prompts, the prompt region including an instructions subregion for inputting the different candidate instructions and an input data subregion for inputting the different input data (Reddy et al. Figures 5B illustrates an exemplary prompt template. The "Content information" and "Input" sections are considered analogous to the instructions and input data subregions, respectively).
Reddy et al. embodiment 1 in view of Guo et al. do not explicitly disclose all of an output region for displaying corresponding different candidate outputs.
However, Pan et al. disclose a prompt region for inputting the different candidate prompts, the prompt region including an instructions subregion for inputting the different candidate instructions (Pan et al. Fig. 4 illustrates instructions subregion 202, which displays a candidate instruction) [and an input data subregion for inputting the different input data]; and
an output region for displaying the corresponding different at least one candidate outputs (Pan et al. Fig. 4 illustrates output region 206, which displays different candidate outputs corresponding to the candidate instruction).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Reddy et al. embodiment 1 in view of Guo et al. to incorporate Pan et al.’s output region.
The suggestion/motivation for doing so would have been to, “[allow] users to get inspired and review prompt variations more effectively,” as noted by the Pan et al. disclosure in paragraph [0034].
Claim 17
Regarding claim 17, the rejection of claim 16 is incorporated. The limitations of claim 17 are similar in scope to that of claim 7 and therefore are rejected for similar reasons as described above.
Claim 18
Regarding claim 18, the rejection of claim 17 is incorporated. The limitations of claim 18 are similar in scope to that of claim 8 and therefore are rejected for similar reasons as described above.
Claims 12, 13, and 19 are rejected under 35 U.S.C. 103 as obvious over Reddy et al. embodiment 1 in view of Guo et al. as applied to claims 1 and 16 above, and further in view of "PromptChainer: Chaining Large Language Model Prompts through Visual Programming" (Wu et al.).
Claim 12
Regarding claim 12, the rejection of claim 1 is incorporated.
Reddy et al. embodiment 1 in view of Guo et al. do not explicitly disclose all of nodes and edges.
However, Wu et al. disclose storing a plurality of nodes, each node comprising a prompt and at least one output generated by the generative language model based on the prompt (Wu et al. pg. 6, Figure 5, "P4’s chains depict iterative content processing" Scenario P4 includes Generic LLM node 10 and Generic LLM node 12. Each Generic LLM node contains an input and generated LLM output. See Figure 3 for node definitions), wherein a first node and a second node of the plurality of nodes are connected by an edge when the first node includes a first prompt and at least one first output and the second node includes a second prompt generated by the generative language model utilizing user input directed to one or both of the first prompt and the at least one first output (Wu et al. pg. 6, Figure 5. Scenario P4 illustrates a first Generic LLM node 10 that connects to a User Action node 11. User Action node 11 takes user input. The output of the User Action node is passed through a Processing node into the input (prompt) of a second Generic LLM node 12. See Figure 3 for node definitions).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Reddy et al. embodiment 1 in view of Guo et al. to incorporate Wu et al.’s node view.
The suggestion/motivation for doing so would have been that, “designers and developers may struggle to prototype realistic applications with only a single LLM prompt ... For designers and developers to apply chaining to their own prototypes, they need to … prompt within each individual LLM step,” as noted by the Wu et al. disclosure in pg. 1-2, Section 1, Paragraph 2.
Claim 13
Regarding claim 13, the rejection of claim 12 is incorporated.
Wu et al. further disclose outputting for display, on a display of a user device, a progress region including the plurality of nodes and edges between nodes of the plurality of nodes (Wu et al. pg. 6, Figure 5, Scenario P4 illustrates a user interface displaying the progress of a prompt through several nodes. The display includes a plurality of Generic LLM nodes and edges connecting the Generic LLM nodes to other nodes surrounding them.).
Claim 19
Regarding claim 19, the rejection of claim 16 is incorporated. The limitations of claim 19 are similar in scope to that of claim 12 and therefore are rejected for similar reasons as described above.
Claim 14 is rejected under 35 U.S.C. 103 as obvious over Reddy et al. embodiment 1 in view of Guo et al. in view of Wu et al. as applied to claim 13 above, and further in view of "ChainForge: A Visual Toolkit for Prompt Engineering and LLM Hypothesis Testing" (Arawjo et al.).
Claim 14
Regarding claim 14, the rejection of claim 13 is incorporated.
Reddy et al. embodiment 1 in view of Guo et al. in view of Wu et al. do not explicitly disclose all of selecting and outputting nodes.
However, Arawjo et al. disclose automatically outputting a particular prompt associated with a particular node and at least one particular output associated with the particular node in response to user selection of the particular node (Arawjo et al. pg. 7, Section 4.1, Paragraph 3, "To inspect responses, users open a pop-up Response Inspector (Figure 4). The inspector has two layouts: Grouped List, where users see LLM responses side-by-side for the same prompt and can organize responses by hierarchically grouping on input variables; and Table, with columns plotting input variables and/or LLMs by user choice." See Fig. 4, which illustrates "four LLMs’ responses side-by-side to the same prompt." See Fig. 1 and Fig. 5, both of which illustrate that a user must select the "Inspect response" button located within an LLM node).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Reddy et al. embodiment 1 in view of Guo et al. in view of Wu et al. to incorporate Arawjo et al.’s node selection.
The suggestion/motivation for doing so would have been to, “help users navigate [a] multiverse of outputs and reduce complexity to reach conclusions across them,” as noted by the Arawjo et al. disclosure in pg. 7, Section 4.1, Paragraph 3.
Reference Cited
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure.
US Patent Publication 20250131020 A1 to Gupta et al. discloses generating subsequent prompts based on user reactions to input data (e.g. video).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JACOB B VOGT whose telephone number is (571)272-7028. The examiner can normally be reached Monday - Friday 9:30am - 7pm 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, Paras D Shah can be reached at (571)270-1650. 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.
/JACOB B VOGT/ Examiner, Art Unit 2653
/Paras D Shah/ Supervisory Patent Examiner, Art Unit 2653
05/12/2026
