Prosecution Insights
Last updated: November 20, 2026
Application No. 18/811,614

ARTIFICIAL INTELLIGENCE SYSTEMS AND ASSOCIATED METHODS FOR INTERACTING WITH APPLICATION WORKFLOWS

Non-Final OA §103
Filed
Aug 21, 2024
Examiner
ACHILLE, ASHMEED CARCIA
Art Unit
2198
Tech Center
2100 — Computer Architecture & Software
Assignee
Ema Unlimited Inc.
OA Round
1 (Non-Final)
Grant Probability
Favorable
1-2
OA Rounds

Examiner Intelligence

Grants only 0% of cases
0%
Career Allowance Rate
0 granted / 0 resolved
-55.0% vs TC avg
Minimal +0% lift
Without
With
+0.0%
Interview Lift
resolved cases with interview
Typical timeline
Avg Prosecution
11 currently pending
Career history
24
Total Applications
across all art units
This examiner has no resolved cases yet (career too new); statute-level performance unavailable. The Grant Probability card shows Tech Center averages instead.

Office Action

§103
DETAILED ACTION This communication is response to the application filed on August 21, 2024, in which claims 1-20 are pending in the application. Claims 1 and 11 are in independent form. 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 . Priority Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, 365(c), or 386(c) is acknowledged. Applicant has not complied with one or more conditions for receiving the benefit of an earlier filing date under 35 U.S.C. 120, 121, 365(c), or 386(c) as follows: Applicant has identified the relationship between the application in the specification in respect to the Continuation-in-Parts. However, no documentation has been submitted regarding Application No. 18/610,276, filed March 20, 2024, which the instant Application No. 18/811,614 filed August 21, 2024, is in cross-reference. Applicant is reminded that, in order to claim priority to an earlier application, the proper documentation must be submitted in the Application Data Sheet. Claim Rejections - 35 USC § 103 In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status. The following is a quotation of 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. Claim(s) 1, 2, 10-12 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child (US 20230121261 A1). As per claim 1 Mishchenko discloses receiving an input at a user interface associated with a large language model (LLM), wherein the input describes a task to be performed; [col 6:25-30] “As yet a further example, in some embodiments, in response to a conversation with a user (e.g., including through interpreting digital natural language input), a natural language model and/or associated system may, based on instructions received from the user, automate tasks for the user across one or more external applications” Mishchenko does not disclose automatically generating, by the LLM, a generative workflow for completing the task described in the input, wherein the generated generative workflow includes a plurality of steps; determining that a particular step, from the plurality of steps of the generative workflow, is to be processed using an external application; in response to determining that the particular step, from the plurality of steps of the generative workflow, is to be processed using an external application executing an application programming interface (API) call to a selected first external application, from a plurality of external applications; and based on instructions within the API call, causing the first external application to process the particular step of the generative workflow, wherein the first external application utilizes an external application workflow, which is distinct from the generative workflow, to process the particular step; receiving a result from the first external application via the API to process the particular step, wherein the result is associated with the processing of the particular step of the generative workflow by the first external application using the external application workflow. However, Dunn discloses automatically generating, by the LLM, a generative workflow for completing the task described in the input, [0024] “A user may interact with an app and/or web application running or otherwise executing on a mobile device 102 to accomplish one or more tasks or actions.”[0028] “ The application graph 232 is generated based on the application model 212 and requisite data from the repository or storage location 236; that is, the application graph 232 assembles processes and data together for executing the task intent 220 based on the task model 224.”[0035] “In examples, the input 620 may be received as a task input 604, where the input 620 may be processed at 608 to determine a task intent such that an appropriate task model may be generated and/or retrieved.” wherein the generated generative workflow includes a plurality of steps; [0027] “The task model 224 may include one or more subtasks. For example, the task model 224 may break a task intent of “send flowers” to the subtask of (1) determine who the flowers are to be delivered to; (2) determine an address or location to send the flowers to; (3) determine when the flowers are to be delivered; (4) determine what kind of flowers to send; (5) determine how much to spend on the flowers; (6) determine a vendor app to use to purchase the flowers; and (7) determine if there are to be any additions to the flower order, such as a card, vase, etc. For each subtask, a task intent 220 and/or a task model 224 maybe determined.” Mishchenko and Dunn are in similar field of endeavor, as they are both in software arrangement using machine learning models and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify the system of Mishchenko to include automatically generating, by the LLM, a generative workflow for completing the task described in the input, wherein the generated generative workflow includes a plurality of steps as disclosed by Dunn automating task based on an application. Modification would increase the system efficiency by transforming the user intent into a structured workflow for performing the task, by providing an automated mechanism for orchestrating tasks. (Dunn [0005]) Mishchenko and Dunn do not disclose determining that a particular step, from the plurality of steps of the generative workflow, is to be processed using an external application; in response to determining that the particular step, from the plurality of steps of the generative workflow, is to be processed using an external application executing an application programming interface (API) call to a selected first external application, from a plurality of external applications; and based on instructions within the API call, causing the first external application to process the particular step of the generative workflow, wherein the first external application utilizes an external application workflow, which is distinct from the generative workflow, to process the particular step; receiving a result from the first external application via the API to process the particular step, wherein the result is associated with the processing of the particular step of the generative workflow by the first external application using the external application workflow. However, Child discloses determining that a particular step, from the plurality of steps of the generative workflow, is to be processed using an external application; [0114] “the action sequence system 105 performs a platform action in response to detecting the event trigger at the act 402 and (optionally) detecting that one or more conditions are satisfied at the act 404….the action sequence system 105 can perform one or more of acts 408-418 as applicable and as depicted in FIG. 4.” [0120] “As further shown in FIG. 4, the act 418 comprises transmitting digital integration data to a third-party server. As described above, the action sequence system 105 can integrate a variety of different data across platforms.” Examiner note: Child disclose this limitation by teaching that upon detecting a trigger event the system will perform a platform action selecting among different action types and when the “applicable action” is available the system then transmit the information. The system has determined a particular step in need to be processed from the plurality of steps. in response to determining that the particular step, from the plurality of steps of the generative workflow, is to be processed using an external application [0114] “At an act 406, the action sequence system 105 performs a platform action in response to detecting the event trigger at the act 402 and (optionally) detecting that one or more conditions are satisfied at the act 404.” executing an application programming interface (API) call to a selected first external application, from a plurality of external applications; and [0074] “the digital-action-sequence templates 312 comprise a specially curated selection of event triggers and corresponding platform actions relevant to a specific user associated with the computing device 300” [0100] “the platform action 346k comprises an option for generating a request via an API (e.g., to call data within an internal survey management platform or an external web service).”[0110] “event triggers may include creation of a digital content item, a digital event from an external platform, or a user opt-out selection. Accordingly, in one or more embodiments, the act 402 comprises detecting one or more parameters associated with the event trigger. [0114] “the action sequence system 105 performs a platform action in response to detecting the event trigger at the act 402 and (optionally) detecting that one or more conditions are satisfied at the act 404.” Examiner note: Child teaches this limitation by having the system detecting an event trigger, which is executed by the platform action, by using an API request/call to the external application to perform particular step. based on instructions within the API call [0100] “the platform action 346k comprises an option for generating a request via an API (e.g., to call data within an internal survey management platform or an external web service).” , causing the first external application to process the particular step of the generative workflow, [0121] “For example, the action sequence system 105 can transmit the digital action sequence to the third-party application for complete external integration rather than detecting an event trigger from a third-party application and subsequently performing a platform action internally.” wherein the first external application utilizes an external application workflow, which is distinct from the generative workflow, to process the particular step; and [0121] “To illustrate an example, a digital sales platform (e.g., a Salesforce® platform) may detect its own event trigger occurring within the digital sales platform. The digital sales platform may then perform the platform action internally— by requesting data (e.g., survey response data) from a survey management platform, such as a Qualtrics® platform. In this manner, digital action sequences can be embedded into third-party applications for single-workflow implementation as opposed to connecting two separate workflows (one externally and one internally).” Examiner Note: Child teaching read on the limitation by disclosing the API call perform the required instruction in response from a request with an external application internal process, to execute the particular step. receiving a result from the first external application via the API to process the particular step, [0107] “the field value 358 can include or reference data associated with another platform action (e.g., a prior platform action in a digital action sequence). For example, the action sequence system 105 can pipe (or pull) in data from other datasets relating to other platform actions to associate such data with the additional platform action being defined in FIG. 3M” note: result is received from the platform action which is available for use in the sequences. wherein the result is associated with the processing of the particular step of the generative workflow by the first external application using the external application workflow. [0107] “embedded data field 356 of “TicketId” can correspond to a field value of a ticket key identifier for a digital ticket to be generated based on the platform action 346a.” the result of the input associated to the outcome of the platform action [0121] “the digital sales platform may then perform the platform action internally” note: based on the digital action of the sequence the result is then produce to the external application in its workflow. Mishchenko, Dunn and Child are in similar field of endeavor, as they all use using machine learning models to interact with application workflows and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, and Dunn to include determining a particular step and in response to determine that particular steps executing an application programming interface (API) call to a selected first external application, from a plurality of external applications; and based on instructions within the API call, causing the first external application to process the particular step of the generative workflow, wherein the first external application utilizes an external application workflow, which is distinct from the generative workflow, to process the particular step and receiving a result from the external application of the particular step as disclosed by Child to a provide a sequence- builder. Modification would provide a more efficient and optimized way to enabling generated workflow by allowing the external application to perform the intended task. (Child [0121]) As per claim 11, it has similar limitations as claim 1, therefore it is rejected under the same rationale. As per claim 2, Mishchenko, Dunn and Child disclose the method of claim 1 detailed above. Child further discloses “ wherein the external application workflow is any one of a) embedded workflow, b) customer owned workflow, and c) real-time generated workflow.” [0021] “In addition to options for event triggers and/or selectable conditions, the action sequence system provides a set of platform actions within the sequence-builder-user interface to continue building a digital action sequence (e.g., a digital workflow) based on user selections.”[0022] “For example, in certain implementations, the action sequence system identifies multiple user selections corresponding to multiple platform actions (e.g., aggregated platform actions) that can be performed upon detection of one or more event triggers.”[0078] “As an additional example, an event trigger 314b includes or represents one or more events identified in relation to a digital sales platform (e.g., the Salesforce® platform). For instance, the event trigger 314b may include the digital sales platform performing an act or workflow operation” As per claim 12, it has similar limitation as claim 2, therefore it is rejected under the same rationale. As per claim 10, Mishchenko, Dunn and Child disclose the method of claim 1 detailed above. Child further discloses determining whether a subsequent step sequentially follows the particular step in the generative workflow or if the particular step is the last step in the generative workflow; [0034] “If there remain specific steps that the electronic processor 200 has not performed, the electronic processor 200 analyzes a webpage using the second machine learning system 220 to identify GUI element associated with a subsequent specific step in the order of specific steps (returns to block 430). After the first specific step has been performed by the electronic processor 200, the webpage that is accessed and analyzed in block 430 may be a different webpage than the webpage specified in the general natural language command. For example, the webpage that is analyzed may be a webpage accessed as a result of a previous specific step being performed.” and in response to determining that a) the subsequent step sequentially follows the particular step of the generative workflow, using the received result for the particular step as input into the subsequent step that sequentially follows the particular step and b) the particular step is the last step in the generative workflow, computing a final result for completing the task to be performed. [0034] “If there remain specific steps that the electronic processor 200 has not performed, the electronic processor 200 analyzes a webpage using the second machine learning system 220 to identify GUI element associated with a subsequent specific step in the order of specific steps (returns to block 430). After the first specific step has been performed by the electronic processor 200, the webpage that is accessed and analyzed in block 430 may be a different webpage than the webpage specified in the general natural language command. For example, the webpage that is analyzed may be a webpage accessed as a result of a previous specific step being performed.” As per claim 20, it has similar limitations as claim 10, therefore it is rejected under the same rationale. Claims 3 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child (US 20230121261 A1), in view of Ghoche (US 20240386214 A1) As per claim 3, Mishchenko, Dunn and Child disclose the method of claim 2 detailed above. Mishchenko, Dunn and Child do not disclose automatically generating the real-time generated workflow, which relates to the external application workflow generated in real-time, wherein the automatic generation of the real-time generated workflow is in response to determining that customer owned workflow is either not available or does not fit a requirement for executing the particular step. However, Ghoche discloses automatically generating the real-time generated workflow[0183] “FIG. 28A is a high-level flowchart of an example of a method of automating the generation of template answers and/or workflow for a selection topic., which relates to the external application workflow generated in real-time,[0239] “workflow policy is automatically built based on an analysis of representative answers for a given intent. An AI model can be used to infer, from representative answers”[0248] “AI technique generate this inferred policy as natural language, but it can also detect any required API calls or actions needed to execute the policy.” wherein the automatic generation of the real-time generated workflow is in response to determining that customer owned workflow is either not available or does not fit a requirement for executing the particular step. [0176] “For example, in one implementation this may include identifying topics for which a workflow has not yet been automated.” [0183] “In block 2800, topics are automatically discovered. This may be implemented, for example, using the discover module. In block 2802, a recommendation is made to automate a topic not previously automated. For example, the most frequently encountered un-automated topics may be recommended for automation. In block 2804, the text portions of template answers are automatically generated for selected topics.”[0194] “In block 3310, a template answer/workflow is automatically generated based on the detected intent.” Examiner note: Ghoche teaches this limitation by disclose that the workflow policy is automatically created in real-time from the user intent, which further combine with the API calls needed to execute the particular step. Mishchenko, Dunn, Child ,and Ghoche are in similar field of endeavor, as they are all in managing application programming interfaces (APIs) and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, and Child to include automatically generating the real-time generated workflow, which relates to the external application workflow generated in real-time, wherein the automatic generation of the real-time generated workflow is in response to determining that customer owned workflow is either not available or does not fit a requirement for executing the particular step as disclosed by Ghoche by automatic creating a workflow. Modification would improve the system reliability by reduce the time that an admin needs to create the workflow by automatically generating a workflow based on the intent task. (Ghoche [0180]) As per claim 13, it has similar limitation as claim 3, therefore it is rejected under the same rationale. Claims 4 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child(US 20230121261 A1), in view of Mackey (US 20120079457 A1) As per claim 4, Mishchenko, Dunn and Child disclose the method of claim 2 detailed above. Mishchenko, Dunn and Child do not disclose causing the first external application to process the particular step of the generative workflow using the plurality of selected external application workflows; obtaining a plurality of sets of results from the first external application for processing the particular step, wherein each set of results are obtained for each of the plurality of selected external application workflows used by the first external application for processing the particular step; and selecting a set of results, from the plurality of sets of results, to process the particular step. However, Mackey discloses causing the first external application to process the particular step of the generative workflow using the plurality of selected external application workflows; [0033] “For example, the action broker (230) may send a first action request (260) and first implementation objects (262), i.e., objects that are formatted to be used by a first automation implementation (264), to the first automation implementation (264). For example, this may be done by the action broker (230) making a call to the first automation implementation (264). obtaining a plurality of sets of results from the first external application for processing the particular step, [0033] “The first automation implementation (264) can return first implementation results (268), which can be objects returned from running the action. The first implementation results (268) can be returned in a format that is specific to the first execution path.” wherein each set of results are obtained for each of the plurality of selected external application workflows used by the first external application for processing the particular step; [0036] “The input/output information of these actions is generally similar, but not the same, between implementations. In the test environment (200), Action B could be executed in the UI path and action C could be executed in the API path because the output results can be converted through the broker to allow communication across implementations. [0042] “For example, choosing (322) can include using a priority list indicating one or more action execution paths for the first action, as well as determining availability of one or more automation implementations for executing the first action through one or more action execution paths indicated by the priority list. and selecting a set of results, from the plurality of sets of results, to process the particular step. [0037] “The components can also include the logical layer (220) including indications of available automation implementations (264, 274) for executing actions through a plurality of action execution paths. The components can also include the priority list (250) indicating an execution path priority for one or more of the actions to be performed.” Mishchenko, Dunn, Child and Mackey are in similar field of endeavor, as they are all automating task across application workflow and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, and Child to include causing the first external application to process the particular step of the generative workflow using the plurality of selected external application workflows; obtaining a plurality of sets of results from the first external application for processing the particular step, wherein each set of results are obtained for each of the plurality of selected external application workflows used by the first external application for processing the particular step; and selecting a set of results, from the plurality of sets of results, to process the particular step as disclosed by Mackey by providing a test automation execution. Modification would be to improve the reliability of the system by allowing steps to be executed through several paths and by doing so it would prevent steps from not being executed properly.(Mackey [0037], [0042]) As per claim 14, it has similar limitation as claim 4, therefore it is rejected under the same rationale. Claims 5-6 and 15-16 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child(US 20230121261 A1), in view of Lim (US 20130152038 A1) As per claim 5, Mishchenko, Dunn, and Child disclose the method of claim 1 detailed above. Mishchenko further discloses selecting the API from a plurality of API options, wherein the selected API is to be used to call the selected first external application” [col 9:65-67 , col 10 :1-7]“According to some embodiments, a natural language model may be trained to call functions in response to the first or second input. For example, a natural language model may be provided, in the model's system prompt, with a description of functions specific to one or more particular APIs via the manifest. From the description of functions, the natural language model may select one or more of the functions which it may determine based on the desired external API, a natural language input from a user of the natural language model user interface,…” Mishchenko, Dunn, and Child do not disclose selecting the external application workflow from a plurality of external application workflow options; selecting the first external application from the plurality of external applications, wherein the selected external application workflow is to be used with the selected first external application; selecting API parameters, from a plurality of API parameters, to execute an API call using the selected API to the selected first external application. However, Lim discloses selecting the external application workflow from a plurality of external application workflow options; [0024] Workflow design application 106 provides an initial user interface for creating project workflows. User interface may display at least one selectable project workflow platform type upon verifying a connection between the workflow design application and a project server. For example, FIG. 2 illustrates an example of an initial user interface 200 for creating a project workflow according to one embodiment. In order to provide project workflow creation, a web application platform tool 102 may be capable of establishing a connection to the project server selecting the first external application from the plurality of external applications, wherein the selected external application workflow is to be used with the selected first external application; [0025] “A project server platform selection type menu 204 may then be displayed in the user interface 200. A project server platform type may be displayed as a selectable item among other, general purpose workflow platform types that are unrelated to projects or project servers (e.g., "Workflow Platform 1," "Workflow Platform 2," etc. in FIG. 2). Workflow design application 106 may then receive a project server platform type selection from a user (e.g., a "Workflow Platform--Project Server" platform type option 206).” selecting API parameters, from a plurality of API parameters, to execute an API call using the selected API to the selected first external application. [0036] “Project field data 122 including one or more project workflow parameters and one or more project workflow variables may be requested, received by and stored in the project data lookup module 114… “[0037] Project workflow lookup user interface 600 may also present to the user and receive a selection of a source field 606 in the form of the one or more variables or parameters available from the project server, such as project server 104 through API 116. The user may also specify a "return field" that indicates the workflow field to which the returned variable or parameter should be written. Data source, field retrieval, and variable or parameter selection may be repeated to set variables and parameters as desired.” Mishchenko, Dunn, Child and Lim are in similar field of endeavor, as they are all in managing application programming interfaces (APIs) and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, and Child to include selecting the external application workflow from a plurality of external application workflow options; selecting the first external application from the plurality of external applications, wherein the selected external application workflow is to be used with the selected first external application; and selecting API parameters, from a plurality of API parameters, to execute an API call using the selected API to the selected first external application as disclosed by Lim to provide a workflow design application. Modification would reduce configuration of API calls errors and simplifies the creation of the workflow by generating a workflow of the external application based on the APIs parameters. (Lim [0005], [0006], [0036]) As per claim 15, it has similar limitations as claim 5, therefore it is rejected under the same rationale. As per claim 6, Mishchenko, Dunn, Child and Lim disclose the method of claim 5 detailed above. Mishchenko further discloses “wherein the selection of the external application workflow, first external application, API, and API parameters is automatically performed by an LLM without user intervention.” [col 5 28:34] “Furthermore, the model may then provide a resulting output from the external tool or application to the user, all without the user needing to access the external tool or application on their own. Moreover, all of this may be performed without any involvement from users, developers, or administrators associated with the natural language model or associated systems” [col 5 49-56] “As an example, if a publisher of an external tool or application desires to allow that external tool or application to be accessible by (and thus integratable with) a natural language model via an API associated with the external tool or application, the publisher may follow one or more of the following steps (without requiring any additional interaction or involvement of any user associated with the natural language model):” As per claim 16, it has similar limitation as claim 6, therefore it is rejected under the same rationale. Claims 7 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child(US 20230121261 A1), in view of Dolby (US 20180121320 A1) As per claim 7, Mishchenko, Dunn, and Child disclose the method claim 1 detailed above. Mishchenko, Dunn, and Child do not disclose “wherein the instructions and API parameters associated with the API call are obtained from a catalog, wherein the instructions and API parameters are used for performing the API call and troubleshooting any failures encountered during the API call to the first external application.” However, Dolby disclose “wherein the instructions and API parameters associated with the API call are obtained from a catalog, [0024] “The static web API invocation debugging tool 100 statically analyzes source code (such as JavaScript) before run-time for errors related to generation of web service requests (i.e., invocation of web APIs). wherein the instructions and API parameters are used for performing the API call and troubleshooting any failures encountered during the API call to the first external application.” [0023] Some embodiments of the disclosure provide a debugging tool for analyzing source code (e.g., JavaScript), specifically to statically troubleshoot invocations of web APIs in the source code. [0027] “the static web API invocation debugging tool 100 receives source code files 150 from the code storage 115 and performs static debugging of web API invocations in the source code by referencing web API specifications” [0028] “The static web API invocation debugging tool 100 has an extractor 155 for extracting a set of request information 157 for each web API invocation in the source code 150. A set of request information for a web API invocation may include a URL, an HTTP method, and request data. The static web API invocation debugging tool 100 also has a specification interpreter 165, which identifies web API endpoints 167 in the web API specifications 160” Mishchenko, Dunn, Child ,and Dolby are in similar field of endeavor, as they are all in managing application programming interfaces (APIs) and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, and Child to include wherein the instructions and API parameters associated with the API call are obtained from a catalog, wherein the instructions and API parameters are used for performing the API call and troubleshooting any failures encountered during the API call to the first external application as disclosed by Dolby analyzing API specification. Modification would increase the efficiency within the system by reducing run-time errors and failed API calls to the external application to further identified API calls errors; and by identifying API call errors it will then go through a check point comparing the call to the specification. (Dolby [0022] [0040]) As per claim 17, it has similar limitation as claim 7, therefore it is rejected under the same rationale. Claims 8 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child(US 20230121261 A1), in view of Manzano (US 20240412157 A1) As per claim 8, Mishchenko, Dunn, and Child disclose the method claim 1 detailed above. Mishchenko, Dunn, and Child do not disclose searching a catalog for approved external applications that can be used to process the particular step; determining that the first external application is listed as an approved external application in the catalog; selecting the first external application based on its approved listing in the catalog to process the particular step of the generative workflow. However, Manzano discloses searching a catalog for approved external applications that can be used to process the particular step; [0046] “In 206, integration application generator system 110 may employ token analyzer 112 to match a source and target in API catalog 121 to the source token and target token received in and parsed from the natural language input. For example, token analyzer 112 may reference connector repository 122 by engaging connector metadata service 115 to determine APIs in API catalog 121 that match a source token and retrieve the API specifications for the matching source as needed to complete subsequent steps. determining that the first external application is listed as an approved external application in the catalog;[0015] “ The inference relies upon API specifications in an API catalog. After determining the requisite sources, targets, connectors, operations, entities, and data mappings, the technique may generate and deploy an integration application to accommodate the natural language request.”[0018] “Once an associated API for the source is identified, the system may determine an entity or entities associated with a specified entity token or tokens using an API Specification for the particular API. The system may then identify operation(s) associated with the entity from the API Specification that match an operation token.” [0029] “Token analyzer 112 may reference connector repository 122 by engaging connector metadata service 115 to determine APIs in an API catalog that match a source token and retrieve the API specifications as needed.” selecting the first external application based on its approved listing in the catalog to process the particular step of the generative workflow. [0049] “In 212, integration application generator system 110 may determine if matches were found for the source, target, entity, and operations. If matches were found for each requirement of the integration application, method 200 may proceed to 214. If ambiguities remain, i.e., if no source token was not found in the API catalog and/or no entity token or operation token is found in the API specification, a map entry may be set to null. Null entities produce an exception trigger and user 102 may be informed that the requested data is not available in the requested service. In such an instance, method 200 may return to 202 to receive additional input from user 102.” Mishchenko, Dunn, Child and Manzano are in similar field of endeavor, as they are all in managing application programming interfaces (APIs) and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, and Child to include selecting the first external application based on its approved listing in the catalog to process the particular step of the generative workflow, determining that the first external application is listed as an approved external application in the catalog and selecting the first external application based on its approved listing in the catalog to process the particular step of the generative workflow as disclosed by Manzano by bridging communication through applications. Modification would be to provide a more efficient way to process a task and to minimize failed API requests by establishing a proper way of processing an application that exist with the specification based on the API call by identify the request to the source and target application. (Manzano [0017], [0046] [0052]) As per claim 18, it has similar limitation as claim 8, therefore it is rejected under the same rationale. Claims 9 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Mishchenko (US 11922144 B1), in view of Dunn (US 20210303342 A1), in view of Child(US 20230121261 A1), in view of Manzano (US 20240412157 A1), in view Iyer (US 20210109487 A1) As per claim 9 Mishchenko, Dunn, and Child disclose the method of claim 1 detailed above. Mishchenko, Dunn, and Child do not disclose searching for approved external applications that can be used to process the particular step; determining that first external application is not listed as an approved external application; and in response to determining that first external application is not listed as an approved external application: automatically generating a first action using the LLM; transmitting the generated first action for approval; and listing the first action in a catalog upon approval. However, Manzano discloses searching for approved external applications that can be used to process the particular step; [0046] “In 206, integration application generator system 110 may employ token analyzer 112 to match a source and target in API catalog 121 to the source token and target token received in and parsed from the natural language input. For example, token analyzer 112 may reference connector repository 122 by engaging connector metadata service 115 to determine APIs in API catalog 121 that match a source token and retrieve the API specifications for the matching source as needed to complete subsequent steps. determining that first external application is not listed as an approved external application; [0049] “if no source token was not found in the API catalog and/or no entity token or operation token is found in the API specification,” and in response to determining that first external application is not listed as an approved external application [0049] “if no source token was not found in the API catalog and/or no entity token or operation token is found in the API specification,” Mishchenko, Dunn, Child and Manzano are in similar field of endeavor, as they are all in managing application programming interfaces (APIs) and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, and Child to include searching for approved external applications that can be used to process the particular step; determining that first external application is not listed as an approved external application, as disclosed by Manzano by bridging communication through applications. Modification would be to provide a more efficient way to process a task and to minimize failed API requests by establishing a proper way of processing an application that exist with the specification based on the API call by identify the request to the source and target application. (Manzano [0017], [0046] [0052]). Mishchenko, Dunn, Child and Manzano do not disclose automatically generating a first action using the LLM; transmitting the generated first action for approval; and listing the first action in a catalog upon approval. However, Iyer disclose automatically generating a first action using the LLM; [0089] “After entering the text in the text editor, the developer may click on a “GENERATE” button, for example. This triggers the AI model to build the workflow using the text and return the generated workflow in XAML format. See, for example, FIG. 10, which is an image 900 illustrating the RPA workflow created from the text of FIG. 8,” transmitting the generated first action for approval; and [0078] “If the predicted workflow(s) is/are approved or modified by the developer, RPA workflow development application 602 may send the approved or modified workflow to training database 606 for storage.” listing the first action in a catalog upon approval. [0086] “At 708, the workflow generation module loads the workflow model(s), and predicts the workflow using the preprocessed text from step 704. At 710, the workflow generation module transmits a list of possible XAML files to the RPA workflow development application based on the prediction(s).” Mishchenko, Dunn, Child, Manzano and Iyer are in similar field of endeavor, as they are all in using AI to managing APIs and, therefore, are combinable/modifiable. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed inventions to modify the system of Mishchenko, Dunn, Child, and Manzano to include automatically generating a first action using the LLM ;transmitting the generated first action for approval; and listing the first action in a catalog upon approval as disclosed by Iyer to provide a workflow generation. Modification would improve the build time of workflow by automatically creating actions when there is not an approve application by doing so it would reduce development for a developer/admin to build a workflow. (Iyer [0025], [0070]) As per claim 19, it has similar limitation as claim 9, therefore it is rejected under the same rationale . Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Mohanty; Bijan Kumar – (US 11775894 B2) discloses receiving at least one natural language input corresponding to at least one task and analyzing the at least one natural language input using one or more machine learning models to determine one or more parameters of the at least one task. FRAZZINGARO; Gennaro – (US 20180329957 A1) discloses executing one or more services of the electronic device, and a relevant output responsive to the user request can be returned to the user. Aghajanyan; Armen; - (US 20210117624 A1) discloses receiving a user input comprising a natural-language utterance by an assistant xbot from a client system associated with a user, determining a semantic representation of the user input based on a structural ontology Prathap; Sujith - (US 20220366331 A1) disclose aids users in completing software workflows using information about the context of the desired outcome Rodriguez Bravo; Cesar Augusto – (US 20220392444 A1) discloses one or more instructions associated with the first task request are contextually recognized. The first task request is completed based on the contextually recognized one or more instructions. Hugelmann; Pascal – (US 20230368103 A1) discloses method may include receiving a natural language command invoking a workflow of an enterprise software application. Bhargav; Pandravada – (US 20240054035 A1) discloses receiving, from one or more source point associated with one or more user device, a set of natural language instructions for performing a task, processes, via a language model (LM) engine. Raimondo; Stefania Evelina-( US 20240176958 A1) discloses a workflow to be implemented by a chatbot to complete a task-oriented dialog is received. Tripathy; Shyamalendu – (US 20250139417 A1) discloses a workflow template can be populated with generated parameter values and provided to a workflow system that generates a workflow based on the workflow template. Bhargava; Vivek -(US 20250307774 A1) discloses identifying one or more users to receive the at least one task approval request by processing data associated with the at least one task approval request using one or more artificial intelligence. El Hattami; Amine- ( US 20230385026 A1) discloses a machine learning inputs based at least in part on the text description and the context information are provided to a machine learning model to determine an implementation prediction for the desired workflow. STIFFLER; MARK A. –(US 20140337071 A1) disclose receiving a user input specifying a request type representing a framework of information to be provided by an end user for a request to be acted on. Yaobo Liang (TaskMatrix.AI: Completing Tasks by Connecting Foundation Models with Millions of APIs, March 29, 2023) discloses a foundation models to propose task solution outlines and then automatically match some of the sub-tasks in the outlines to the off-the-shelf models and systems with special functionalities to complete them. Yifan Song – (RestGPT: Connecting Large Language Models with Real-World RESTful APIs, August 27, 2023) disclose achieved remarkable progress API executor tailored for calling RESTful APIs, which can meticulously formulate parameters and parse API responses. Any inquiry concerning this communication or earlier communications from the examiner should be directed to ASHMEED ACHILLE whose telephone number is (571)272-9437. The examiner can normally be reached Monday-Friday 7am -4pm. 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, Pierre Vital can be reached at (571)272-4215. 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. /A.A./ Examiner, Art Unit 2198 /PIERRE VITAL/ Supervisory Patent Examiner, Art Unit 2198
Read full office action

Prosecution Timeline

Aug 21, 2024
Application Filed
Oct 16, 2026
Non-Final Rejection mailed — §103 (current)

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

1-2
Expected OA Rounds
Grant Probability
Low
PTA Risk
Based on 0 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month