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 .
Detailed Action
In amendments dated 6/23/26, Applicant amended claims 1, 12, and 20, canceled no claims, and added no new claims. Claims 1-20 are presented for examination.
Rejections under 35 U.S.C. 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to mental processes without significantly more. Independent clams 1, 12, and 20 each recites generating, via the one or more processors, a prompt for input into the generative Al model based upon the prompt criteria; evaluating, via the one or more processors, classification performance of the prompt at classifying documents in the corpus of documents with respect to the relevancy requirement; generating, via the one or more processors, an updated prompt based upon the updated prompt criteria; evaluating, via the one or more processors, classification performance of the updated prompt at classifying documents with respect to the relevancy requirement; and based on the evaluation, approving, via the one or more processors, the updated prompt to classify additional documents in the corpus of documents. Generating a prompt and an updated prompt is each generating data and a mental process accomplishable in the human mind or on paper. Evaluating performance of the prompt and performance of the updated prompt are each mental processes, and approving the updated prompt is also evaluating and a mental process. Each claim recites additional elements of obtaining, via one or more processors, prompt criteria associated with a document review of a corpus of documents, wherein the prompt criteria defines at least (i) a relevancy requirement for an inquiry and (ii) a description of an issue; and obtaining, via the one or more processors, an updated prompt criteria including an updated description of the issue, which are both data gathering steps and insignificant extra-solution activity. Claim 12 recites one or more processors and one or more non-transitory memories and claim 20 recites a non-transitory computer readable medium, which are each generic components of a computer. Examiner notes specification paragraphs 0004 and 0032 discuss drawbacks of using machine learning models in the eDiscovery process, for example using machine learning models may be cumbersome and inefficient, different attorneys working on the same case may use the machine learning models in different ways, and machine learning processes require training using a significant number of manually-labeled examples before a classifier can be sufficiently useful. Paragraph 0033 discusses modifying a prompt to classify documents with a generative AI model instead of training a machine language classifier with further details discussed elsewhere in the specification such as in figures 6-9 begging in paragraph 0078. The steps recited in the claim do not recite a particular improvement in any technology or function of a computer per MPEP 2106.04(d) and do not recite any unconventional steps in the invention per MPEP 2106.05(a). Therefore, the recited mental processes are not integrated into a practical application. Taking the claim as a whole, the data gathering steps are recited broadly and amounts to receiving data across a network per specification paragraphs 0045-0046 and figures 1 and 10, which is routine and conventional activity per the list of such activities in MPEP 2106.05(d) part II. The one or more processors, one or more non-transitory memories, and non-transitory computer readable medium are each still generic components of a computer. Thus the claims do not include additional elements that are sufficient to amount to significantly more than the recited mental processes.
Claims 2 and 13 each recites inputting, via the one or more processors, the prompt and each document of a sample of documents from the corpus of documents into the generative Al model to obtain a set of respective classifications of the sample of documents, and inputting the prompt and a document into an AI model is recited broadly and amounts to receiving data across a network per specification paragraphs 0045-0046 and figures 1 and 10, which is routine and conventional activity per the list of such activities in MPEP 2106.05(d) part II, also Examiner notes applying an AI model is not significantly more than a mental process per Recentive Analytics v. Fox Broadcasting Corp. (134 F.4th 1205, 2025 U.S.P.Q.2d 628); obtaining, via the one or more processors, review data associated with the sample of documents including ground truth data associated with the relevancy requirement, which is recited broadly and amounts to receiving data across a network per specification paragraphs 0045-0046 and figures 1 and 10, which is routine and conventional activity per the list of such activities in MPEP 2106.05(d) part II; and applying, via the one or more processors, the review data to determine classification performance of the prompt with respect to the relevancy requirement, and determining classification performance is evaluating and a mental process. Claims 3 and 14 each recites comparing, via the one or more processors, the classification performance associated with the prompt to classification performance associated with the updated prompt, and comparing performances is evaluating and a mental process.
Claims 4 and 15 each recites determining, via the one or more processors, that classification performance of the updated prompt with respect to the issue has improved over the classification performance of the prompt with respect to the issue, and determining is evaluating and a mental process. Claims 5 and 16 each recites determining, via the one or more processors, that classification performance of the updated prompt with respect to the relevancy requirement has not degraded over the classification performance of the prompt with respect to the relevancy requirement, and determining is evaluating and a mental process. Claims 6 and 17 each recites analyzing, via the generative Al model, the prompt criteria to determine that no contradiction exists between the relevancy requirement and the description of the issue, and analyzing prompt criteria is recited broadly and a mental process accomplishable in the human mind or on paper, and using the generative AI model is applying a generative AI model and is not significantly more than a mental process per Recentive Analytics v. Fox Broadcasting Corp. (134 F.4th 1205, 2025 U.S.P.Q.2d 628), and determining no contradiction exists is evaluating and a mental process; and in response to determining that a contradiction exists, generating an alert, which is recited broadly and is a mental process accomplishable in the human mind or on paper. Claim 7 recites wherein classifying documents in the corpus of documents includes classifying a document as one of: junk; responsive; not responsive; likely responsive; or likely not responsive, and classifying a document is applying a generative AI model per specification 0031 and is not significantly more than a mental process per Recentive Analytics v. Fox Broadcasting Corp. (134 F.4th 1205, 2025 U.S.P.Q.2d 628).
Claims 8 and 18 each recites generating, via the one or more processors, the prompt by supplementing the prompt criteria with additional context, which is recited broadly and is a mental process accomplishable in the human mind or on paper. Claims 9 and 19 each recites wherein at least a portion of the additional context is a set of rules that that instruct the generative Al model to reach intermediate conclusions before outputting a classification for a document, and supplementing with rules is recited broadly and is a mental process accomplishable in the human mind or on paper. Claim 10 recites wherein the intermediate conclusions include one or more of: citations to documents in the corpus of documents that support the intermediate conclusions; rationales behind the intermediate conclusions; and considerations accounted for when making the intermediate conclusions, which are each data and a mental process accomplishable in the human mind or on paper. Claim 11 recites wherein the additional context includes additional prompt criteria defining one or more of: case summary; a description of relevant entities; or identification of key documents, and prompt criteria are data and a mental process accomplishable in the human mind or on paper.
Rejections under 35 U.S.C. 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 1-5, 7-16, and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Natarajan et al (US 20240428937) hereafter Natarajan, in view of Heller et al (US 12,067,366), hereafter Heller.
With respect to claims 1, 12, and 20, Natarajan teaches:
obtaining, via one or more processors, prompt criteria associated with a document review of a corpus of documents, wherein the prompt criteria defines at least (i) a relevancy requirement for an inquiry (paragraph 0281 data for a prompt associated with a corpus of medical research (review of documents) and includes query relevancy (relevancy to an inquiry));
generating, via the one or more processors, a prompt for input into the generative Al model based upon the prompt criteria (paragraph 0281 generating a prompt from the data);
evaluating, via the one or more processors, classification performance of the prompt at classifying documents in the corpus of documents with respect to the relevancy requirement (paragraph 0023 tuning the prompt from a query to tweak model behavior);
obtaining, via the one or more processors, an updated prompt criteria including an updated description of the issue (paragraph 0036 updating parameters for prompt input);
generating, via the one or more processors, an updated prompt based upon the updated prompt criteria (paragraph 0036 figure 1 generate an updated prompt);
evaluating, via the one or more processors, classification performance of the updated prompt at classifying documents with respect to the relevancy requirement (paragraph 0044 prompt values iteratively updated, performance of downstream model using the prompt evaluated); and
based on the evaluation, approving, via the one or more processors, the updated prompt to classify additional documents in the corpus of documents (paragraph 0044 updated prompt approved since iteratively used in model).
Natarajan does not teach obtaining, via one or more processors, prompt criteria associated with a corpus of documents, wherein the prompt criteria defines at least (ii) a description of an issue. Heller teaches this in the prompt created from a prompt template using a question summaries as citations from legal documents as data in a prompt template (column 18 lines 45-60). It would have been obvious to have combined the function of prompt criteria including a description of a legal issue in Heller with the creation and updating prompt techniques in Natarajan to provide more information as Narajan also describes obtaining data for prompts including data from or review of output by legal representatives or advocates (various examples in paragraphs 0299-0330) and including such information for generating a prompt makes for a more accurate uses of the model.
With respect to claims 2 and 13, all the limitations in claims 1 and 12 are addressed by Natarajan and Heller above. Natarajan also teaches:
inputting, via the one or more processors, the prompt and each document of a sample of documents from the corpus of documents into the generative Al model to obtain a set of respective classifications of the sample of documents (paragraph 0280 input into model includes a prompt an documents);
obtaining, via the one or more processors, review data associated with the sample of documents including ground truth data associated with the relevancy requirement (paragraph 0279 getting review data for example documents and test questions from professors to orient the model to a domain of medical literature); and
applying, via the one or more processors, the review data to determine classification performance of the prompt with respect to the relevancy requirement (paragraph 0279 applying the review data with the prompt to evaluate the prompt data to help orient the model to the domain).
With respect to claims 3 and 14, all the limitations in claims 1, 2, 12, and 13 are addressed by Natarajan and Heller above. Natarajan also teaches comparing, via the one or more processors, the classification performance associated with the prompt to classification performance associated with the updated prompt (paragraph 0068 compare output of model with reference to update prompts).
With respect to claims 4 and 15, all the limitations in claims 1, 2, 3, 12, 13, and 14 are addressed by Natarajan and Heller above. Natarajan also teaches determining, via the one or more processors, that classification performance of the updated prompt with respect to the issue has improved over the classification performance of the prompt with respect to the issue (paragraph 0068 a decreasing difference between compared outputs means improvement).
With respect to claims 5 and 16, all the limitations in claims 1, 2, 3, 12, 13, and 14 are addressed by Natarajan and Heller above. Natarajan also teaches determining, via the one or more processors, that classification performance of the updated prompt with respect to the relevancy requirement has not degraded over the classification performance of the prompt with respect to the relevancy requirement (paragraph 0068 a decreasing difference between compared outputs means performance has not degraded).
With respect to claim 7, all the limitations in claim 1 are addressed by Natarajan and Heller above. Heller also teaches wherein classifying documents in the corpus of documents includes classifying a document as one of: junk; responsive; not responsive; likely responsive; or likely not responsive (column 18 lines 45-66 references from source documents for prompt are assigned numbers on a scale of 1 to 5 for relevancy).
With respect to claims 8 and 18, all the limitations in claims 1 and 12 are addressed by Natarajan and Heller above. Heller also teaches generating, via the one or more processors, the prompt by supplementing the prompt criteria with additional context (column 5 lines 13-22 adding workflow for example to aid prompt and the model, also example column 10 lines 29-41 figure 4 discussing step 410 supplementing a prompt based on input text).
With respect to claims 9 and 19, all the limitations in claims 1 and 12 are addressed by Natarajan and Heller above. Natarajan also teaches wherein at least a portion of the additional context is a set of rules that that instruct the generative Al model to reach intermediate conclusions before outputting a classification for a document (paragraph 0226 instruct a model to perform first task (intermediate), then perform a second task using output (text portions) from the first task).
With respect to claim 10, all the limitations in claims 1, 8, and 9 are addressed by Natarajan and Heller above. Natarajan also teaches wherein the intermediate conclusions include one or more of: citations to documents in the corpus of documents that support the intermediate conclusions; rationales behind the intermediate conclusions; and considerations accounted for when making the intermediate conclusions (paragraph 0226 first task instructions include citations (passages) to documents that support questions for lawyers).
With respect to claim 11, all the limitations in claims 1, 8, 9, and 10 are addressed by Natarajan and Heller above. Natarajan also teaches wherein the additional context includes additional prompt criteria defining one or more of: case summary; a description of relevant entities; or identification of key documents (paragraph 0226 passages as description of relevant entities).
Relevant Prior Aert
During his search for prior art, Examiner found the following references to be relevant to Applicant's claimed invention. Each reference is listed on the Notice of References form included in this office action:
Yang, Eugene et al, “Effectiveness Results for Popular e-Discovery Algorithms,” teaches text categorization algorithms for determining relevance of documents to a case, does not teach prompts or evaluating classification performance of prompts with the algorithms (Abstract, 2 E-Discovery Configuration pages 1-2 ); and
Krishnamurthy et al (US 202450135729) teaches classifying a document for straight-through processing using a machine learning model and a confidence score between said model and a model template, does not teach prompts for the machine learning model or evaluating the performance of classification using prompts (paragraphs 0008, 0023-0024).
Responses to Applicant’s Remarks
Regarding rejections of claims 1-20 under 35 U.S.C. 101 for reciting mental processes without significantly more, Applicant’s arguments have been considered but are not persuasive. On page 8 of his Remarks Applicant paraphrases MPEP 2106.04(d)(I) (“a claim integrates a judicial exception into a practical application when it improves the functioning of a computer or improves another technology or technical field. To evaluate such an improvement, (1) the specification should provide sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement, and (2) the claim itself must reflect the disclosed improvement.”). On pages 8-9 Applicant cites the problems described in specification paragraphs 0004 and 0032 and discusses aspects of the invention’s solution such as a two-branch iterative workflow for refining a prompt when classification performance has not met a threshold in paragraphs 0115-0116 which is not claimed. Applicant then discusses Example 47 claim 3, “which was found eligible because the specification described an improvement to network security technology and the claim reflected that improvement through specific steps.” Applicant then asserts “Similarly here, the specification describes an improvement to document classification technology through the specific iterative prompt refinement and approval/deployment workflow that replaces conventional machine learning classifier training and ensures uniform application of classification criteria across the corpus. Applicant respectfully submits that the claims reflect this improvement.” Examiner disagrees as the claims broadly recite steps of generating a prompt, evaluating classification performance of the prompt, generating an updated prompt, evaluating classification performance of the updated prompt, and approving the updated prompt without details showing how the invention performs any of these activities. While a BRI of the evaluating and approving steps include evaluating and are mental processes, the BRI of generating a prompt and an updated prompt include use of pen and paper and are also mental processes. Furthermore, unlike in Example 47 claim 3, the recited claims do not recite details of iterative prompt refinement. For example in addition to each evaluating step not reciting how the invention performs the evaluating, each evaluating step also does not recite a result of the evaluating and none of the following steps (obtaining updated prompt criteria, generating an updated prompt, approving the updated prompt) indicate they are performed in any way from the results of the evaluating steps. Thus the claims do not recite an improvement in document classification technology.
On page 11 Applicant discusses claims 6 and 17 and asserts they do not recite mental processes since the analyzing step is performed via a generative AI model. Examiner disagrees as the analyzing step is recited broadly and is merely performed on a computer per MPEP 2106.05(f)(2). For clarity, Examiner notes here and in the rejection that use of the generative AI model is conventional and not significantly more than a mental process per Recentive Analytics v. Fox Broadcasting Corp. (134 F.4th 1205, 2025 U.S.P.Q.2d 628).
Regarding rejections of claims 1-5, 7-16, and 18-20 under 35 U.S.C. 103 by Natarajan in view of Heller, Applicant’s arguments have been considered but are not persuasive. On pages 11-2 Applicant discusses the approving step of the claim and its support in specification paragraphs 0115-0116 and argues that Natarajan does not teach this functionality from the specification. Examiner notes paragraphs 0115-0116 and the Applicant discuss aspects of the invention that are not claimed such as evaluation against a classification metric threshold and actions taken related to said threshold being met or not. MPEP 2145(vi) states “Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims.” Examiner believes Natarajan teaches the approving limitation in paragraph 0044 as shown I the rejections above.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Inquiry
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRUCE M MOSER whose telephone number is (571)270-1718. The examiner can normally be reached M-F 9a-5p.
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, Boris Gorney can be reached at 571 270-5626. 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.
/BRUCE M MOSER/Primary Examiner, Art Unit 2154 8/28/26
Prosecution Insights