DETAILED ACTION
Acknowledgements
This Office Action is in response to Applicant’s correspondence filed on 1/30/26.
The Examiner notes that citations to United States Patent Application Publication paragraphs are formatted as [####], #### representing the paragraph number.
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
Status of Claims
Claims 1-20 are currently pending.
Claim 20 is withdrawn.
Claims 1-19 are rejected as set forth below.
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 1/30/26 has been entered.
Response to Arguments
Claim Rejections - 35 U.S.C. § 101
Applicant’s arguments with respect to claim(s) 1, 11 have been fully considered and are persuasive. The rejection (and corresponding rejections to its dependent claims, if applicable) is withdrawn.
Claim Rejections - 35 U.S.C. § 103
Applicant’s arguments with respect to claims 1-19 have been considered but are moot because the arguments do not apply to any of the references being used in the current rejection.
The Examiner notes that, during the 12/30/25 interview, agreement was reached regarding how the cited prior art did not teach a machine learning model receiving an additional reward signal having a strength greater than the first reward signal (emphasis placed). There was no agreement regarding the cited prior art not teaching a machine learning model that alters respective internal decision-making models to receive an additional reward signal for a subsequent output list. See the Interview Summary filed on 1/2/26.
Doran teaches one or more machine learning models that alters respective internal decision-making models to receive a reward signal for an output list. See the Paragraph 17 of the 35 USC 103 rejection below. In addition, one of ordinary skill in the art would recognize that the machine learning model is trained more than once since the database is not fixed and new data is routinely being added into the database. See [0029] of Doran.
Claim Rejections - 35 USC § 112(b)
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-19 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
As per claims 1, 11, the limitation “when a second output list of accepted contract clauses substantially matches a second correct output list” includes a relative term which renders the claim indefinite. The term “substantially” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention.
By virtue of dependence, the dependent claims are similarly rejected.
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 1-7, 9-17, 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over United States Patent Application Publication No. 20210125297 to Doran in view of United States Patent Application Publication No. 20180012315 to Erbe, United States Patent Application Publication No. 20230015535 to Smeltzer, and Non-Patent Literature Title ‘Reinforcement Learning Tutorial’ to Arumugam.
As per claims 1, 11, Doran teaches:
A clause review system for reviewing contract clauses in a contract, the clause review system comprising at least one processor in communication with at least one memory device, wherein the at least one processor is programmed to: execute one or more machine learning models configured to detect clause identifiers in a contract, wherein each clause identifier identifies a corresponding regulation; ([0045], “At step 308, the input data may be provided to at least one machine learning model. At step 308, the input data may be identified and certain features extracted and grouped together to form patterns. As a specific example, the machine learning model may compare the identified and extracted contract terms of the received input data to at least one model trained on previously received contract provisions with similar terms and target attributes that are associated with specific contract term identifiers.”)
train the one or more machine learning models using a first training dataset including first set of processed contracts, wherein an output list of accepted contract clauses created by one or more of the machine learning models is compared to a correct output list of accepted contract clauses, the one or more machine learning models are trained to refine one or more rules, each respective rule of the one or more rules corresponding to respective contract clause, and the one or more rules are each configured to aid a determination of whether the one or more machine learning models should accept a respective contract clause of a respective contract, based on one or more factors of the respective contract; retrain the one or more machine learning models using a second training dataset including a second set of processed contracts, wherein at least one of the processed second contracts has a different file format and/or a different sender entity from at least one of the first set of contracts, and the one or more machine learning models are retrained using a reinforcement mechanism, the reinforcement mechanism being configured to provide a first reward signal to the one or more machine learning models when a second output list of accepted contract clauses substantially matches a second correct output list, an additional reward signal being provided for a subsequent output list of accepted contract clauses generated based on a subsequent set of processed acquisition order, wherein the subsequent output list matches a subsequent correct output list based on the subsequent set of processed acquisition orders; ([0043]-[0045], “At step 304, at least one machine learning model may be trained on the input data from the database. The at least one machine learning model may be trained using a variety of machine learning algorithms. Targets (or target attributes) may be mapped to the contract provisions. For example, a certain keyword in a contract provision (e.g., a heading or term-of-art) may be associated with a standard contract clause. In other words, the certain term-of-art may be mapped to the “target” contract clause. The machine learning model(s) is trained to find and establish patterns and/or confirm the frequency of recognized patterns that have been established by at least one machine learning model. The machine learning model, once trained, will be able to identify future patterns in input data based on the previously mapped contract features and previously recognized patterns. At step 306, new input data is collected. Input data may include but is not limited to a user selection of consent, a user selection of rejection, a user non-selection, a contract, a contract provision, a contract term-of-art, parties to the contract, and other information associated with contracts between businesses and users. In at least one aspect, input data may comprise a contract provision and an NLP summary of that contract provision. At step 308, the input data may be provided to at least one machine learning model. The pattern that may be established by the user in rejecting certain contract provisions combined with a user's overall privacy profile (indicating a user's preferences as to how his/her PII is shared or distributed among third parties) may be used to train at least one machine-learning model. The model may subsequently be applied to future contract provisions, offering a recommendation of accepting or rejecting to the user.”, The Examiner notes that training the machine learning model to confirm the frequency of recognized patterns that have been established by the model is equivalent to comparing a correct output list of accepted contract clauses to an output list created by the machine learning model and retraining the model the using a reinforcement mechanism. In addition, contracts are not limited to a single business or entity. See [0020]. In addition, one of ordinary skill in the art would recognize that the machine learning model is trained more than once since the database is not fixed and new data is routinely being added into the database. See [0029].)
detect clause identifiers in an unprocessed acquisition order using the one or more trained machine learning models; ([0045], “At step 308, the input data may be provided to at least one machine learning model. At step 308, the input data may be identified and certain features extracted and grouped together to form patterns. As a specific example, the machine learning model may compare the identified and extracted contract terms of the received input data to at least one model trained on previously received contract provisions with similar terms and target attributes that are associated with specific contract term identifiers.”)
determine whether to accept contract clauses corresponding to the clause identifiers based on the one or more rules and one or more factors of the unprocessed acquisition order; and output a list of accepted contract clauses in the unprocessed acquisition order, wherein the list of accepted contract clauses is presented in a standard format irrespective of a file format of the unprocessed acquisition. ([0045], “In some aspects, the comparison results may include a confidence indicator (e.g., 70% confidence that the contract clause is similar to other contract clauses the user has seen before) as to whether a user should agree to a certain contract provision or not. For example, the comparison results may include information indicating how confident the model is that one or more contract terms are related to a specific type of contract provision, and based on historical user actions with similar types of contract provisions, the machine-learning model may indicate that the user should not agree to that particular provision.”; Fig 5, [0038], “At step 208, the NLP summary may be presented. In examples, the NLP summary may be displayed adjacent to the original provision of the contract, allowing a user to easily read the NLP summary but also check the NLP summary against the original language of the contract provision. In further examples, as illustrated in FIG. 5, the NLP summary may be displayed in proximity to at least one switch that may prompt the user to either consent or not consent to a particular provision.”)
Doran does not explicitly teach, but Erbe teaches:
acquisition order; ([0003], “In one embodiment, a method is provided for processing a variation of a construction project contract between a contractor and a second party for work to be performed on a construction project by the contractor for the second party.”)
acquisition regulations; ([0003], “The request object includes a first new contract component item and, in some embodiments, multiple new contract component items. The first new contract component item of the request object includes an item ID field identifying an item ID of the first new contract component item, a target field identifying a memory location for a new contract component in a contract budget data structure, and a proposed value field identifying a proposed value for the new contract component.”)
One of ordinary skill in the art would have recognized the substitution of a known prior art element of Erbe for another known prior art element of Doran would have yielded predicable results. Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject and the prior art rests not on any individual element or function but in the very combination itself- that is in the substitution of the acquisition order and the acquisition regulations in the secondary reference for the contract and the regulations in the primary reference. It would have been recognized by one of ordinary skill in the art that the results of the substitution would have been predictable because an acquisition order and an acquisition regulation are merely specific types of a contract and a regulation does not change the functionality of Doran.
Doran as modified does not explicitly teach, but Smeltzer teaches:
government acquisition regulations; ([0016], “In one embodiment, the present disclosure is directed to a method, which comprises: (i) receiving, by a computing system, a digital solicitation from a government entity for a public works project being offered by the government entity for competitive bidding, wherein the digital solicitation includes: (a) an electronic contract document containing procurement terms and conditions of the government entity for the public works project.”)
One of ordinary skill in the art would have recognized the substitution of a known prior art element of Smeltzer for another known prior art element of Doran as modified would have yielded predicable results. Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject and the prior art rests not on any individual element or function but in the very combination itself- that is in the substitution of government acquisition regulations in the secondary reference for the acquisition regulations in the primary reference. It would have been recognized by one of ordinary skill in the art that the results of the substitution would have been predictable because a government acquisition regulation is merely a specific type of an acquisition regulation and does not change the functionality of Doran.
Doran as modified does not explicitly teach, but Arumugam teaches:
the one or more machine learning models alter respective internal decision- making models to receive an additional reward signal having a strength greater than the first reward signal, the additional reward signal being provided to the one or more machine learning models; (Slide 12, The reward function of a reinforcement learning process is defined as R: S x A -> R, indicating that the reward feedback is a numerical real number and can vary in degree.)
One of ordinary skill in the art would have recognized that applying the known technique of Arumugam to the known invention of Doran as modified would have yielded predictable results and resulted in an improved invention. It would have been recognized that the application of the technique would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such reinforcement learning features into a similar invention. Further, it would have been recognized by those of ordinary skill in the art that modifying the one or more machine learning models to alter respective internal decision- making models to receive an additional reward signal having a strength greater than the first reward signal, the additional reward signal being provided to the one or more machine learning models, results in an improved invention because applying said technique increases the granularity of the reward signal, thus allowing the machine learning models to be better fine-tuned and therefore improving the overall accuracy of the invention.
As per claims 2, 12, Doran teaches:
wherein the at least one processor is further programmed to: determine the one or more rules corresponding to the contract clauses based on prescriptions in the contract clauses. ([0043], “Targets (or target attributes) may be mapped to the contract provisions. For example, a certain keyword in a contract provision (e.g., a heading or term-of-art) may be associated with a standard contract clause. In other words, the certain term-of-art may be mapped to the “target” contract clause. The machine learning model(s) is trained to find and establish patterns and/or confirm the frequency of recognized patterns that have been established by at least one machine learning model. The machine learning model, once trained, will be able to identify future patterns in input data based on the previously mapped contract features and previously recognized patterns.”)
As per claims 3, 13, Doran does not explicitly teach, but Erbe teaches:
determine whether the contract clauses corresponding to the clause identifiers apply to a subcontractor; and output a list of flow down contract clauses if a recipient of the acquisition order is subcontracting. ([0003],” In some embodiments, the second party is a general contractor for the construction project and the contractor is a subcontractor hired by the general contractor.”; [0063], “Each SOV line item in the GC SOV 157 identifies either a direct cost incurred by the general contractor (i.e., materials and labor provided directly by the GC) or corresponds to a subcontract between the general contractor and one or more subcontractors or materials suppliers (e.g., subcontract 159, subcontract 161, and subcontract 163). The work and materials associated with a particular subcontract are further defined by one or more subcontract line items 165, 167. A subcontractor is linked to the project through one or more particular subcontracts and the subcontractor is then able to create a schedule of values for each subcontract line item—for example, SOV 169 is created by the subcontractor for subcontract line item 165.”)
One of ordinary skill in the art would have recognized that applying the known technique of Erbe to the known invention of Doran would have yielded predictable results and resulted in an improved invention. It would have been recognized that the application of the technique would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such contract processing features into a similar invention. Further, it would have been recognized by those of ordinary skill in the art that modifying the invention to determine whether the contract clauses corresponding to the clause identifiers apply to a subcontractor; and output a list of flow down contract clauses if a recipient of the acquisition order is subcontracting, results in an improved invention because applying said technique ensures that the clause that is being processed is assigned to the correct entity before being approved or denied, thus improving the overall accuracy of the invention.
As per claims 4, 14, Doran does not explicitly teach, but Erbe teaches:
receive a list of one or more factors of the acquisition order; and determine whether to accept the contract clauses and/or whether the contract clauses apply to a subcontractor based on the list of one or more factors and the one or more rules. ([0062]-[0063], “For example, a general contractor may divide the work associated with the project into work types and define separate general contract line items for electrical, plumbing, materials, etc. and define a value (i.e., a portion of the overall project value) to each category. The GC schedule of values 157 corresponding to the “electrical” contract line item 153 may further include multiple SOV line items as shown in FIG. 1C. Alternatively, the GC may contract with an electrical subcontractor to provide all electrical work for a project.”)
One of ordinary skill in the art would have recognized that applying the known technique of Erbe to the known invention of Doran would have yielded predictable results and resulted in an improved invention. It would have been recognized that the application of the technique would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such contract processing features into a similar invention. Further, it would have been recognized by those of ordinary skill in the art that modifying the invention to receive a list of factors of the acquisition order; and determine whether to accept the contract clauses and/or whether the contract clauses apply to a subcontractor based on the list of factors and the one or more rules, results in an improved invention because applying said technique ensures that the clause that is being processed is assigned to the correct entity before being approved or denied, thus improving the overall accuracy of the invention.
As per claims 5, 15, Erbe teaches:
receive the list of one or more factors including a contract value of the acquisition order; and determine whether to accept the contract clauses based on the contract value and the one or more rules; ([0064], “If, during the course of the project, the subcontractor discovers that their costs will exceed the agreed upon value in a particular subcontract 159 or if additional work or materials are necessary, the subcontractor may request a variation of the terms of the subcontract. In the examples described herein, an approved variation issued by the general contractor causes a new subcontract line item 171 with a defined value to be added to a subcontract 157. The Subcontractor is then able to define a new SOV 173 corresponding to the new subcontract line item 171. Because, in most cases, the addition of this new subcontract line item 171 alters the total value of the subcontract 159, the general contractor may decide to request a corresponding variation to the project contract 151.”)
As per claims 6, 16, Doran teaches:
receive the list of one or more factors including a type of subject matter of the acquisition order; and determine whether to accept the contract clauses based on the type of subject matter and the one or more rules. ([0043])
As per claims 7, 17, Erbe teaches:
receive the list of one or more factors including a contractor tier of a recipient of the acquisition order; and determine whether the contract clauses apply to a subcontractor based on the contractor tier and the one or more rules; ([0062]-[0063])
As per claims 9, 19, Doran teaches:
wherein the one or more machine learning models are configured to determine the one or more rules based on the contract clauses. ([0043])
As per claims 10, Doran teaches:
comprising a client device configured to communicate with a user, wherein the client device is provided on an application. ([0028], “Client devices 102, 104, and 106 may be configured to access a web portal that implements the systems and methods described herein. In other example aspects, the systems and methods described herein may be a standalone executable software that is downloaded to a client device.”)
Claims 8, 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over United States Patent Application Publication No. 20210125297 to Doran in view of United States Patent Application Publication No. 20180012315 to Erbe, United States Patent Application Publication No. 20230015535 to Smeltzer, and Non-Patent Literature Title ‘Reinforcement Learning Tutorial’ to Arumugam, and further in view of United States Patent Application Publication No. 20160171635 to Senzee.
As per claims 8, 18, Doran as modified does not explicitly teach, but Senzee teaches:
determine whether to accept the contract clauses by: communicating with a website that maintains acquisition regulations. ([0037]-[0038], “The seller's terms are stored on e-commerce website 120 and include shipping, payment, arbitration, choice of law, risk of loss, and other contract terms as determined by a person of ordinary skill in the art. STP 200 then determines whether the parties agree with the terms of the contract (216).”)
One of ordinary skill in the art would have recognized that applying the known technique of Senzee to the known invention of Doran as modified would have yielded predictable results and resulted in an improved invention. It would have been recognized that the application of the technique would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such contract processing features into a similar invention. Further, it would have been recognized by those of ordinary skill in the art that modifying the invention to determine whether to accept the contract clauses by: communicating with a website that maintains acquisition regulations results in an improved invention because applying said technique allows the regulations to be maintained at a single location to ensure that the regulations are up-to-date and be easily accessible to all relevant entities from any location with Internet access, thus improving the overall reliability of the invention.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
United States Patent Application Publication No. 20220092711 to Venkatesan discloses a method for extracting data from contracts using a contract data extraction system. The method includes following steps of: obtaining a contract from which the data to be extracted; processing the contract to identify one or more sections of the contract; scanning for data within the identified one or more sections of the contract; and identifying and extracting the data from the corresponding one or more sections of the contract using natural language processing (NLP). The one or more sections include at least one of clauses, obligations, signature and tabular data. The one or more sections are identified using a predefined library. The identified one or more sections are demarcated by matching with the predefined library. The identified one or more sections are tagged across existing and new contract repositories using the Artificial Intelligence (AI) technique.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JAY HUANG whose telephone number is (408)918-9799. The examiner can normally be reached 9:00a - 5:30p PT.
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, Anita Coupe can be reached at (571) 270-3614. 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.
/JAY HUANG/Primary Examiner, Art Unit 3619
Prosecution Insights