FIELD

[0001] The present disclosure relates generally to online banking, more particularly to predicting financial transaction decisions, and even more particularly to predicting financial transaction decisions based on behavioral data.

BACKGROUND

[0002] The Internet in combination with financial technology (FinTech) enables users to perform online banking tasks using a website, a mobile banking application, or the like. Online banking tasks may include financial transactions such as transferring money to other persons, payment for services or purchases, moving money between accounts, etc. An example of such a mobile banking application known in the art is Royal Bank of Canada’s NOMI™ App. The NOMI™ App may learn to identify when a customer will accumulate excess funds eligible for automatic transfer to a savings account. The NOMI™ App may also learn to detect and automatically alert customers of unusual increases in vendor fees.

[0003] Another mobile application known in the art is Intuit’s Mint App. The Mint App aggregates a customer’s financial account information, from different financial institutions, into a single dashboard, enabling a customer to holistically view all of their accounts. The Mint App also analyzes financial account information to provide insight on customer spending habits, and further provides financial tips for increasing savings and financial wellbeing.

[0004] Modern online banking systems have further incorporated artificial intelligence (Al) techniques, in the form of chatbots. Chatbots may, for example, automatically provide customers with answers to customer queries relating to online banking. One such chatbot known in the art is Bank of America’s Erica™ chatbot. Erica™ may leverage a customer’s financial account information to reply to voice or text prompts. For example, a customer may query Erica™ “Did I have any grocery purchase over $100?” to quickly receive insight from Erica™ into past financial transactions. Moreover, customers can prompt Erica™ to initiate financial transactions with commands such as “Transfer or Send Money” or “Make a bill payment”.

[0005] Other examples of FinTech includes International Patent Application No.

WO 2009/044396 A2 which discloses a system and method for predicting future cash flow and balances based on predicting periodic payments, whether recurring or not. The system and method disclosed therein enable customers to modify payment schedules for a selected period of time. The system and method further provide a balance sheet for predicting periodic transactions with predicted dates and amounts. [0006] It remains commercially desirable to develop further improvements and advancements in financial technology, to overcome shortcomings of known techniques, and to provide additional advantages.

SUMMARY OF INVENTION

[0007] Disclosed herein is an online banking system and method for predicting a user’s next most likely financial operation selection. The banking system collects, transforms, and analyzes data relating to financial operations to generate a financial score indicative of a likelihood of a user selecting a financial operation at a given time. Such data may be securely stored in a database stored in a remote server using techniques such as HTTPS encryption protocols. The remote server may further comprise analytic applications, including artificial intelligence methods, configured to receive, store, and analyze the data relating to financial operations. The banking system may further expedite performing a financial operation in response to receiving a single input from a user.

[0008] In other embodiments, the online banking system may include a banking application configured to run on a user device; a database for storing a plurality of financial operations, each financial operation based on a combination of transaction data; a behaviour prediction application configured to run on the user device or a remote server, the behaviour prediction application communicatively coupled to each of the banking application and the database; the behaviour prediction application to determine a financial operation score for each of the plurality of financial operations based on a transaction patterns associated with each of the plurality of financial operations, the financial operation score for use in predicting a likelihood of the user selecting a corresponding financial operation, and the banking application configured to display each of the plurality of financial operations on the user device for selection by a user, a display location of each of the plurality of financial operations based on the corresponding financial operation score.

[0009] In a further embodiment, the banking application may provide, to the behaviour prediction application, financial operation feedback in response to a financial operation selection, the financial operation feedback for use in adjusting the financial operation score of each of the plurality of financial operations.

[0010] The online banking system may include making a selection on the display location selects the corresponding financial operation for execution.

[0011] The online banking system may include the mobile banking application displaying the financial operations based on the corresponding financial operation score being above a financial operation score threshold.

[0012] In an embodiment, the combination of transaction data may include a customer identifier. [0013] In an embodiment, the behaviour prediction application may lower the financial score of the corresponding financial operation having a transaction date within a penalty window.

[0014] In an embodiment, the transaction pattern may include transaction frequencies associated with each of the plurality of financial operations.

[0015] In an embodiment, making a selection on the display location may be conducted using a swipe.

[0016] In an embodiment, the combination of transaction data may include at least one of: a card number, a transaction type, an outbound account, and an inbound account.

[0017] In a yet further embodiment, a method for performing a financial transaction, may include: receiving, at a user device, a plurality of financial operations each financial operation having a corresponding operation score; displaying, on a display of the user device, the plurality of financial operations positioned based on the corresponding financial operation score; receiving, at the user device, indication of a financial operation selection made by the user, and providing, in response to the financial operation selection, a combination of transaction data, to a banking transaction system, for enabling execution of the selected financial operation.

[0018] In an embodiment, the method may include: determining the corresponding operation score for each of the plurality of financial operations based on a transaction patterns associated with each of the plurality of financial operations. [0019] In an embodiment, the indication of the financial operation selection may be provided by a making a selection on the display location of the corresponding selected financial operation.

[0020] In an embodiment, the transaction patterns may include transaction frequencies associated with each of the plurality of financial operations.

[0021] In an embodiment, the transaction patterns may include transaction patterns that are word patterns.

[0022] In an embodiment, making a selection on the display location may be conducted using a swipe.

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Embodiments will now be described, by way of example only, with reference to the attached Figures.

[0024] Figure 1 illustrates components of an online banking system according to an embodiment disclosed herein.

[0025] Figure 2 illustrates steps of an online banking method according to an embodiment disclosed herein.

[0026] Figure 3 illustrates steps of a behavioural analysis algorithm according to an embodiment disclosed herein.

[0027] Figure 4 illustrates an example financial operation based on a combination of transaction data.

[0028] Figure 5 is a table categorizing financial operation transactions by type and channel.

[0029] Figure 6 is a table categorizing financial operation transactions as repeated and non-repeated. [0030] Figure 7 is a diagram illustrating time deltas between different transactions of a financial operation.

[0031] Figure 8 is a distribution model of time deltas for a financial operation having a plurality of transactions.

[0032] Figure 9 is a table depicting duration and score penalization factors for use in adjusting financial operation scores.

[0033] Figure 10 is a diagram illustrating calculation of a penalization window.

[0034] Figure 11 is a table illustrating improvements in predication accuracy based on adjusting financial operation scores by a penalization factor.

[0035] Figure 12 is an example GUI for a banking application according to an embodiment disclosed herein.

[0036] Figure 13 is an example of another GUI for a banking application according to an embodiment disclosed herein.

[0037] Throughout the drawings, sometimes only one or fewer than all of the instances of an element visible in the view are designated by a lead line and reference character, for the sake only of simplicity and to avoid clutter. It will be understood, however, that in such cases, in accordance with the corresponding description, that all other instances are likewise designated and encompassed by the corresponding description.

DETAILED DESCRIPTION

[0038] The online banking system disclosed herein generally relates to analyzing financial data with behavioural analytics to predict which financial operation(s) a customer is most likely to select at a particular point of time. In a preferred embodiment, the online banking system allows a user to select and perform a financial operation in a single step, thereby expediting the financial operation.

[0039] Financial operations are based on a particular combination of transaction data executed each time the financial operation is selected to be performed. In a preferred embodiment, the transaction data includes a customer card number, a type of transaction, an outbound account number, and an inbound account number. As an illustrative example, a financial operation may pertain to a particular combination of transaction data, such as card number 1234, a transfer type transaction, outbound account 5678, and inbound account 9999; while a different financial operation pertains to a different combination of transaction data, such as card number 1234, a transfer type transaction, outbound account 5678, and inbound account 0000. Types of transactions may comprise virtually any online banking transaction known in the art, including transactions that may otherwise be performed in-person at a financial institution. Types of transactions include, but are not limited to renewing a guaranteed investment certificate (GIC), transferring funds between accounts, paying bills, transferring funds to a third party, etc. Types of transactions having a known transaction date, such as automatically recurring payments, are excluded from consideration as there is no benefit to predicting a financial transaction when the transaction date is already scheduled and known.

[0040] The banking system further leverages behavioural data relating to when a financial operation is performed to generate a financial operation score indicative of a predicted likelihood that a user will select the corresponding financial operation at a particular point of time. In an embodiment, the financial operation score is generated based on behavioural data including transaction frequency, time between subsequent transactions, and time since the most recent transaction. In an embodiment, a total financial operational score for a corresponding financial operation is determined based on aggregating a financial operation score for each occurrence of the corresponding financial operation occurring within a previous number of days. In a preferred embodiment, the total financial operation score is determined based on each financial operation occurrence within the previous 100 days.

[0041] In an embodiment, financial operation scores are further adjusted based on penalization factors. For example, a financial operation having a recent transaction may be indicative that a subsequent transaction with respect to the same financial operation is less likely to occur until a specified amount of time has passed. Accordingly, in an example embodiment, a financial operation score may be adjusted by a penalization factor based on a recently occurring financial operation. Consequently, lowering the financial operation score of a given financial operation results in a relative increase in other final operation scores, advantageously improving prediction accuracy.

[0042] In an embodiment, a plurality of financial operations are illustrated as icons in a graphical user interface (GUI) displayed on a screen of a user device, such as a mobile phone screen or a laptop screen. In a preferred embodiment, the plurality of financial operations are ranked for display in the GUI based on a corresponding financial operation score. In an embodiment, the financial operation score is a predicted likelihood that a user will select the corresponding financial operation. Accordingly, in such an embodiment, the plurality of financial operations are displayed in the GUI according to the predicted likelihood of being selected by a user. Advantageously, presenting financial operations in a manner in which they are likely to be selected by a user provides an improved user experience whereby financial operations are selected in fewer steps and expedited for execution based on using the associated combination of transaction data. In another embodiment, only financial operations having a financial operation score above a financial operation score threshold are displayed in the GUI.

[0043] In a one step embodiment, a user may select a financial operation using a single action, such as clicking or swiping an icon corresponding to a financial operation. In an embodiment, selecting a financial operation results in automatically submitting the combination of transaction data used each time the financial operation is performed, thereby expediting performing the financial operation.

[0044] Further embodiments as disclosed herein may relate to financial operations unrelated to previously executed transactions. In an example embodiment, a financial operation may be automatically presented for selection by a user, based on predicting that a predetermined financial condition will be achieved. In an example embodiment, a financial operation comprises transferring funds to a higher interest account and the financial condition is a predicted total account balance exceeding a predetermined total account balance. In such an example embodiment, a confirmation prompt may be further provided to the user to confirm whether the user wants to perform the financial operation or not. In an embodiment, the financial operation may relate to a third party products. In an embodiment, the financial operation may relate to pre-approved financial services.

[0045] In an embodiment security measures required to reset a password are reduced for users identified as having a high risk of a comprised account.

[0046] Figure 1 illustrates an online banking system according to an embodiment disclosed herein. The online banking system includes a financial solution 110 running on a mobile device communicatively coupled to a remote server 120. In an embodiment, the financial solution 110 is a banking application configured to run on a user device, such as a mobile device, a laptop, or other electronic device operable to run the banking application. In an embodiment, remote server 120 comprises a behaviour prediction application and a database, such as database 230 illustrated in Figure 2. In an embodiment the behaviour prediction application is configured to run on a user device concurrently with the banking application.

[0047] The banking application provides a graphical user interface (GUI) and includes a login prompt enabling a user to log in and access their financial accounts, including bank accounts, saving accounts, chequing accounts, investment accounts, and other accounts associated with banks, financial institutions, and other similar organizations. In an embodiment, logging in generates a set of user data elements, further transmitted to remote server 120. A behaviour prediction application may use the set of user data elements to aid in generating a financial operation score. [0048] The banking application is configured to display a plurality of financial operations, enabling a user of the banking application to select a financial operation to perform. The plurality of operations are displayed as at least one of text, an image, or an icon. In an embodiment, the banking application is configured to retrieve or receive, a financial operation score from the behaviour predication application, for each of the plurality of financial operations. In a further embodiment, a display position associated with each of the plurality of financial operations is based on the corresponding financial score. In a preferred embodiment, the plurality of financial operations are ranked for display according to the corresponding financial score. In such an embodiment, the financial operation having the highest financial operation score may be displayed at a topmost portion of the banking application GUI, with subsequently lower scored financial operations being displayed in sequentially descending order below the highest scoring financial operation.

[0049] User devices, such as a mobile device, a laptop, or other electronic devices operable to run a banking application, are configured to receive user inputs, including inputs received from a keyboard, a mouse, a touchpad, a touchscreen, or other means of receiving user input such as receiving a voice command. Such user input may select a particular financial operation that the user wishes to perform. Accordingly, in response to a user input, the banking application may take steps to ensure a selected financial operation is performed. In an embodiment, the banking application is configured to take steps to perform a financial operation in response to a user performing a selection action over a visual representation of the corresponding financial operation displayed on the user device.

[0050] Figure 2 illustrates a plurality of steps of a banking system method according to an embodiment disclosed herein.

[0051] Step 200 comprises a user logging in to their financial accounts through a banking application running on a mobile device. Logging in further generates user data elements to be provided to a remote server. [0052] Step 210 comprises the remote server receiving the user data elements

210 resulting from the user log in 200.

[0053] Step 220 comprises the remote server collecting other user data elements from a database 230, such as a plurality of financial operations associated with a user’s financial accounts. In an embodiment the remote server comprises a database 230 for storing the plurality of financial operations.

[0054] Step 240 comprises the remote server running behavioral analysis. In an embodiment, the remote server includes a behaviour prediction application configured to run the behaviour analysis. In an embodiment, running behaviour analysis 240 includes generating financial operation scores based on analyzing: transaction frequency, time between subsequent transactions, and time since the most recent transaction of a plurality of financial operations.

[0055] Step 250 comprises the remote server transmitting the most likely next user action to the mobile device. In an embodiment, the behaviour prediction application transmits the most likely next user action to the mobile device. In an embodiment, the most likely next user action corresponds to the financial operation having the highest financial operation score. In an embodiment, the behaviour prediction application transmits, to the mobile device, a plurality of financial operations having the highest financial operation scores indicative of a group of financial operations most likely to be selected by a user. As an illustrative example, the behaviour analysis outputs a financial operation score for each of a first financial operation comprising transfer $20 to Bob, a second financial operation comprising pay $100 to credit card, and a third financial operation comprising transfer $50 to savings account, wherein the first, second, and third financial operations are ranked according to their corresponding financial operation score.

[0056] Step 260 comprises the mobile device receiving the most likely next user action, or other outputs that may be generated in step 250, as disclosed herein. The banking application further displays the received outputs. In an embodiment, the banking application displays the plurality of financial operations in a ranked format according to their financial operation score.

[0057] Step 270 comprises the user selecting (or not selecting) a financial operation displayed on the mobile device. In an embodiment, the user can select more than one financial operation displayed on the mobile device.

[0058] Step 280 comprises transmitting behaviour feedback to the remote server, such as financial operation feedback generated based on a financial operation selection made by the user. In an embodiment the financial operation feedback is formatted as a duple wherein the duple comprises a first value for identifying the financial operation and a second value indicating whether the financial operation was selected. In an embodiment a second value of zero (0) indicates the corresponding financial operation was not selected, and a second value of one (1 ) indicates that the corresponding financial operation was selected. In an illustrative example, the financial operation feedback comprises a string of three duples: (1 , 0), (2, 1 ), and (3, 0). The string of duples indicating that a second financial operation was selected by the user, and that the first and third financial operations were not selected by the user.

[0059] Step 290 comprises the remote server receiving the financial operation feedback, for updating financial information stored on the database 230. The financial operation feedback further improves the underlying behavioural analytic algorithm. [0060] Figure 3 illustrates steps of a behaviour analysis 310 according to an embodiment disclosed herein. The behavior analysis 310 according to an embodiment herein predicts and ranks a predefined number of most likely financial operations for display to a user upon the user logging into the banking application. In an embodiment, the behaviour analysis is prepared by a behaviour prediction application. In an embodiment the behaviour prediction application is run on one of a user device or a remote server. In an embodiment, the output of the behaviour analysis is a financial operation score. [0061] Step 320 comprises preparing user data elements to generate a financial operation having a combination of transaction data based on the user data elements. Figure 4 illustrates such an embodiment of financial operation having a combination of transaction data including a card number, type of transaction, from account (outbound account), and to account/recipient (inbound account). In an embodiment, the financial operation is provided as a string of characters, wherein a “+” character separates different strings of transaction data.

[0062] Step 330 comprises calculating transaction frequency for each financial operation. In an embodiment, calculating transaction frequency is based on classifying financial operations by transaction type and transaction channel. Transactions types include, but are not limited to, e-mail transfers, bank transfers, wire transfers, GIC renewals, etc. Transactions channels include, but are not limited to, banking applications, banking web portals, call centers etc. Figure 5 is a table classifying financial operations base on transaction type (e.g. Type A, Type B, Type C), and transaction channel (e.g. Channel A, Channel B, and Channel C). As illustrated in Figure 5, a transaction type C made on channel B has the highest transaction frequency, accounting for approximately 29.34% of all transactions.

[0063] In addition to calculating transaction frequency, or as an alternative to calculating transaction frequency, behaviour analysis may include categorizing financial operations as repeated and non-repeated transactions, further illustrated in Figure 6. In an embodiment, repeated transactions are those financial operations completed more than once within a predetermined timeframe. In an embodiment, only financial operations occurring within a previous 100 days are considered for classification. Accordingly, explanatory variables in behavioral analysis may include financial operation transaction frequency and financial operation transaction repetition. [0064] Step 340 comprises calculating a time delta, or time between, a plurality of transaction dates of a financial operation. As illustrated in Figure 7, the financial operation has four different transaction dates: Day -35, Day -25, Day -15, and Day 5. A delta is calculated between transaction Day 5, and each other transaction date. The time deltas corresponding to Day 5 are thus 20 days, 30 days, and 40 days. The process of calculating time deltas is repeated for each transaction date to generate a distribution of time deltas. Figure 8 further illustrates an example model of a time delta distribution.

[0065] Step 350 comprises determining a financial operation score for a corresponding financial operation. In an embodiment a total financial operation score is determined for a financial operation based on aggregating a financial operation score for each financial operation transaction that occurred within a predetermined number of days.

[0066] Step 360 comprises ranking each financial operation based on a corresponding total financial operation score.

[0067] Embodiments of a banking system as disclosed herein adjust a financial operation score base on a penalization factor. In an embodiment, the penalization factor comprises a duration factor and a score factor. Figure 9 is table illustrating duration and score penalization factors according to an embodiment disclosed herein In an embodiment the penalization factor is applied when the time since the most recent transaction is less than the historically lowest time between subsequent transactions of a given financial operation. In an embodiment a penalization factor is calculated for each of the plurality of financial operations based on augmenting the shortest time delta by a duration factor and lower the corresponding financial operation score by a score factor.

[0068] As illustrated in Figure 10, a financial operation comprises four previous transactions. The smallest time delta between the transactions is subsequently identified and divided by a duration factor to determine a penalty window. Financial operation transactions occurring within a penalty window of the most recent financial operation transaction are further subject to reducing a corresponding financial operation score based on a score factor. In an embodiment, the smallest time delta is 15 days, the duration factor is 1 .5 and the resulting penalty window is 15/1.5 = 10 days. Accordingly, any corresponding financial operation transaction occurring within 10 days of the most recent financial operation will have a corresponding financial operation score reduced by a score factor. In an embodiment, a financial operation score is 200, the score penalty factor is 2, and the resulting financial operation penalty score is 200/2 = 100.

[0069] Figure 11 illustrates how a penalization factor increases accuracy for repeated financial operation transactions, which in the embodiment disclosed in figure 11 , increases from 81 .6% accuracy to 82.7% accuracy; meaning, 82.7 times out of 100 times, a previous financial operation transaction that a particular user would like to select, will appear in the top five suggestions resulting from the behavioural analysis disclosed herein. Such financial operations may be displayed to users via a mobile banking application running on a user device.

[0070] The following pseudocode is provide as yet a further example of an embodiment disclosed herein:

Dictionary; customer(Q)=Specific Customer transaction_Type(B)=Specific Transaction Type from account(V)=Specific Account where funds are out going to account(C)= Specific Recipient of funds

Transaction_Combination(X)= A concatenation of the variables customer(Q), transaction_Type(B), from account(V), and to account(C)

T=Specific time stamp P=Previous time interval

V= Specific Numeric Score associated with P

Duration_Penalization_Factor= A coefficient used to augment how long a Transaction should be lowered

Penalization Factor= A coefficient that lowers the score if the previous transaction occurred in a specific time interval

Transaction_Combination(X _n ) =A11 Transaction CombinationsAssociated with customer(Q) Final_Score= Score used to assign ranks to Transaction Combinations ranging from most likely to occur to least likely Step 1 Calculate time between each consecutive transaction for Transaction Combination(X):

Transaction_Combination(X)= Concatenate(customer(Q), transaction_Type(B), from account(V), to account(C))

For Transaction_Combination(X); Calculate; time_between(Transaction_Combination(X Transaction_Combination(X) ^t - ¹ )

Step 2 Calculate the Score for Transaction_Coinbination(X) for each predetermined day it landed on;

For Transaction_Combination(X);

Calculate; Sum(if(time_between(Transaction_Combination(X) ^t ,

Transaction_Combination(X) ) =P then V)) as score

Step 3 Determine if the total score for Transaction_combination(X) needs to be lowered by Optimized Penalization Coefficient;

For;

Transaction Combination(X)=Concatenate(customer(Q),transaction Type(B), from account(V), to account(C))

If; time between(Date of

Recommendation, Last Date Transaction combination(X) executed) >= minimum(time between(Transaction Combination(X) Transaction Combination(X) ^t_1 ))

/Duration Penaliztion Factor then score/Penalization Factor else score)as Final Score

Step 4 Rank All Transaction_Combination(Xn) for Customer(Q);

For Customer(Q); Sort and Rank Transaction Combination(X _n) by Final Score

[0071] Figures 12 and 13 further illustrate embodiments of a GUI according to an embodiment of a mobile banking application as disclosed herein. Figure 12 illustrates a GUI 1210 include three financial operations 1220, 1230, and 1240. Figure similarly illustrates three financial operations, namely send money to Yin, pay AT&T bill, and transfer to savings. [0072] In the preceding description, for purposes of explanation, numerous details are set forth in order to provide a thorough understanding of the embodiments. However, it will be apparent to one skilled in the art that these specific details are not required. In other instances, well-known electrical structures and circuits are shown in block diagram form in order not to obscure the understanding. For example, specific details are not provided as to whether the embodiments described herein are implemented as a software routine, hardware circuit, firmware, or a combination thereof. The scope of the claims should not be limited by the particular embodiments set forth herein, but should be construed in a manner consistent with the specification as a whole.

[0073] Embodiments of the disclosure can be represented as a computer program product stored in a machine-readable medium (also referred to as a computer-readable medium, a processor-readable medium, or a computer usable medium having a computer-readable program code embodied therein). The machine- readable medium can be any suitable tangible, non-transitory medium, including magnetic, optical, or electrical storage medium including a diskette, compact disk read only memory (CD-ROM), memory device (volatile or non-volatile), or similar storage mechanism. The machine-readable medium can contain various sets of instructions, code sequences, configuration information, or other data, which, when executed, cause a processor to perform steps in a method according to an embodiment of the disclosure. Those of ordinary skill in the art will appreciate that other instructions and operations necessary to implement the described implementations can also be stored on the machine-readable medium. The instructions stored on the machine-readable medium can be executed by a processor or other suitable processing device, and can interface with circuitry to perform the described tasks.

[0074] The above-described embodiments are intended to be examples only. Alterations, modifications and variations can be effected to the particular embodiments by those of skill in the art without departing from the scope, which is defined solely by the claims appended hereto.