Universal payment intent
In an example embodiment, a solution is provided that introduces a dynamic user interface that automatically identifies payment methods available for a transaction. The dynamic user interface also may optionally present the available payment methods in an ordering, as determined by a machine learning model. More particularly, a payment processing system may determine a payment intents object, with all available payment methods for a transaction and optionally an ordering for those payment methods.
1 . A server system for training and executing machine learning models to dynamically revise, via a document object container, a user interface for presenting request execution methods based on compatibility of available request execution methods, the server system comprising:
at least one hardware processor and non-transitory media storing instructions that, when executed by the at least one hardware processor, cause the at least one hardware processor to perform operations comprising:
generating the document object container, associated with a request type identifier, that is placed on a page of a graphical user interface associated with a computing infrastructure;
in connection with loading the page of the graphical user interface, and before request information provided via the graphical user interface is submitted by the server system to a processing system for processing, retrieving, from a data structure, one or more attributes for each of a plurality of request execution methods;
predicting, using a machine learning model, at least one incompatible request execution method of the plurality of request execution methods a presentation order of the plurality of request execution methods based on comparing the one or more attributes for each of the plurality of request execution methods to the request information; and
generating an intent data object containing a subset of the plurality of request execution methods that causes a presentation of the plurality of request execution methods according to the presentation order and excludes the at least one incompatible request execution method, and mounting the intent data object to the document object container with at least one credential associated with the request execution methods on the page of the graphical user interface to cause presentation of the subset of the plurality of request execution methods;
monitoring interactions with the intent data object, via the graphical user interface, comprising selections of a request execution method from the subset of the plurality of request execution methods to execute a request associated with the request type identifier; and
retraining the machine learning model using the monitored interactions comprising a threshold number of selections where a selected request execution method is a threshold distance from a first request execution method in the presentation order of the plurality of request execution methods.
2 . The server system of claim 1 , wherein the one or more attributes each pertain to a location, merchant, product, transaction, or customer.
3 . The server system of claim 1 , wherein the request information includes a location of a customer and the one or more attributes includes an indication that at least one of the plurality of request execution methods does not process transactions in a currency associated with the location.
4 . The server system of claim 1 , further comprising:
returning, to a client device, an instance of the intent data object with some of the request execution methods removed.
5 . The server system of claim 1 , wherein request execution via one of the subset of the plurality of request execution methods is performed using a request execution element containing an iframe that securely sends request execution information over a secure Hypertext Transfer Protocol (HTTPs) connection.
6 . The server system of claim 5 , wherein the request execution element renders a dynamic form that allows a customer to pick a request execution method from the subset of the plurality of request execution methods.
7 . A method comprising:
generating a document object container, associated with a request type identifier, that is placed on a page of a graphical user interface associated with a computing infrastructure;
in connection with loading the page of the graphical user interface, and before request information provided via the graphical user interface is submitted by a server system to a processing system for processing, retrieving, from a data structure, one or more attributes for each of a plurality of request execution methods;
predicting, using a machine learning model, at least one incompatible request execution method of the plurality of request execution methods a presentation order of the plurality of request execution methods based on comparing the one or more attributes for each of the plurality of request execution methods to the request information;
generating an intent data object containing a subset of the plurality of request execution methods that causes a presentation of the plurality of request execution methods according to the presentation order and excludes the at least one incompatible request execution method, and mounting the intent data object to the document object container with at least one credential associated with the request execution methods on the page of the graphical user interface to cause presentation of the subset of the plurality of request execution methods;
monitoring interactions with the intent data object, via the graphical user interface, comprising selections of a request execution method from the subset of the plurality of request execution methods to execute a request associated with the request type identifier; and
retraining the machine learning model using the monitored interactions comprising a threshold number of selections where a selected request execution method is a threshold distance from a first request execution method in the presentation order of the plurality of request execution methods.
8 . The method of claim 7 , wherein the one or more attributes each pertain to a location, merchant, product, transaction, or customer.
9 . The method of claim 7 , wherein the request information includes a location of a customer and the one or more attributes includes an indication that at least one of the plurality of request execution methods does not process transactions in a currency associated with the location.
10 . The method of claim 7 , further comprising:
returning, to a client device, an instance of the intent data object with some of the plurality of request execution methods removed.
11 . The method of claim 7 , wherein request execution via one of the subset of the plurality of request execution methods is performed using a request execution element containing an iframe that securely sends request execution information over a secure Hypertext Transfer Protocol (HTTPS) connection.
12 . The method of claim 11 , wherein the request execution element renders a dynamic form that allows a customer to pick a request execution method from the subset of the plurality of request execution methods.
13 . A non-transitory machine-readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising:
generating a document object container, associated with a request type identifier, that is placed on a page of a graphical user interface associated with a computing infrastructure;
in connection with loading the page of the graphical user interface, and before request information provided via the graphical user interface is submitted by a server system to a processing system for processing, retrieving, from a data structure, one or more attributes for each of a plurality of request execution methods;
predicting, using a machine learning model, at least one incompatible request execution method of the plurality of request execution methods a presentation order of the plurality of request execution methods based on comparing the one or more attributes for each of the plurality of request execution methods to the request information;
generating an intent data object containing a subset of the plurality of request execution methods that causes a presentation of the plurality of request execution methods according to the presentation order and excludes the at least one incompatible request execution method, and mounting the intent data object to the document object container with at least one credential associated with the request execution methods on the page of the graphical user interface to cause presentation of the subset of the plurality of request execution methods;
monitoring interactions with the intent data object, via the graphical user interface, comprising selections of a request execution method from the subset of the plurality of request execution methods to execute a request associated with the request type identifier; and
retraining the machine learning model using the monitored interactions comprising a threshold number of selections where a selected request execution method is a threshold distance from a first request execution method in the presentation order of the plurality of request execution methods.
14 . The non-transitory machine-readable medium of claim 13 , wherein the one or more attributes each pertain to a location, merchant, product, transaction, or customer.
15 . The non-transitory machine-readable medium of claim 13 , wherein the request information includes a location of a customer and the one or more attributes includes an indication that at least one of the plurality of request execution methods does not process transactions in a currency associated with the location.
16 . The non-transitory machine-readable medium of claim 13 , further comprising:
returning, to a client device, an instance of the intent data object with some of the plurality of request execution methods removed.
17 . The non-transitory machine-readable medium of claim 13 , wherein request execution via one of the subset of the plurality of request execution methods is performed using a request execution element containing an iframe that securely sends request execution information over a secure Hypertext Transfer Protocol (HTTPs) connection.
18 . The server system of claim 1 , wherein retraining the machine learning model comprises adjusting one or more parameters of the machine learning model based on an evaluated loss between predictions of the machine learning model and the monitored interactions.
19 . The method of claim 7 , wherein retraining the machine learning model comprises adjusting one or more parameters of the machine learning model based on an evaluated loss between predictions of the machine learning model and the monitored interactions.
20 . The non-transitory machine-readable medium of claim 13 , wherein retraining the machine learning model comprises adjusting one or more parameters of the machine learning model based on an evaluated loss between predictions of the machine learning model and the monitored interactions.