System and method for providing a digital transaction channel

Abstract

A system and method for providing a digital transaction channel are provided. The method includes extracting data from raw telecommunication event detail records of a group of consumers, building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs and training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger. The method includes providing real time assessment of a consumer at a given time and / or at a given trigger including inputting a consumer's current status to the propensity model, providing a push of one or more action recommendations to the consumer based on the output of the propensity model and in response to receiving a consumer selection of an action recommendation, executing, against the consumer's account, a telecommunication action relating to the selected action recommendation.

Description

[0001] SYSTEM AND METHOD FOR PROVIDING A DIGITAL TRANSACTION CHANNEL

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application claims priority from South African provisional patent application number 2024 / 06181 filed on 13 August 2024, which is incorporated by reference herein.

[0004] FIELD

[0005] This disclosure relates to a system and a method for providing a proactive digital transaction channel in particular to users of a mobile communication service hosted on a mobile telephone network.

[0006] BACKGROUND

[0007] A consumer may be any user or subscriber of a mobile communication service hosted on a mobile telephone network. The consumer may have a communication device with which the consumer may be able to conduct telephonic transactions such as making phone calls, sending short messaging service (SMS) messages, using mobile data and the like. The conducting of transactions over a mobile telephone network incurs charges for the consumer. These charges may be debited against an airtime account, or a bundle account associated with the consumer. An airtime account may be a prepaid account with a positive balance colloquially referred to as “airtime”. A bundle account may be an account associated with a particular category of mobile telephone network utilisation, such as sending SMSs, sending and receiving data packets or making telephone calls. Mobile bundles are typically associated with a usage limit, such as 30 SMSs, 500 MB of data or 100 voice minutes, and charges may be deducted from a bundle account associated with the respective category of utilisation.

[0008] Once the balance of the consumer’s airtime account approaches or reaches zero or once the mobile bundle’s usage limit has been reached, the consumer may be barred from conducting further telephonic transactions. Should the consumer wish to conduct further telephonic transactions, the consumer may perform an “airtime top-up” so as to increase the balance of the consumer’s airtime account or may purchase a mobile bundle such as SMS, data or voice minute bundles.

[0009] There may be cases where a consumer is unaware of his airtime account balance or that his mobile bundles have been depleted as the consumer continues to utilise the mobile telephone network. Such a consumer may be slow in recharging his account or making a decision on what amount of airtime or bundle to purchase.

[0010] The preceding discussion of the background is intended only to facilitate an understanding of the present disclosure. It should be appreciated that the discussion is not an acknowledgment or admission that any of the material referred to was part of the common general knowledge in the art as at the priority date of the application.

[0011] SUMMARY

[0012] In accordance with an aspect of the disclosure there is provided a computer-implemented method for providing a proactive digital transaction channel, comprising: extracting data from raw telecommunication event detail records of a group of consumers; building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; providing real time assessment of a consumer at a given time and / or at a given trigger including: inputting a consumer’s current status to the propensity model; providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and, in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0013] Telecommunication activity may include any occurrence, behaviour, or operation initiated by a consumer, network entity, or external system that takes place within, or in relation to, a mobile telephone network. This may include signalling, data transmission, service usage, system operations, and user interactions. A telecommunication event may be a discrete and observable instance of a telecommunication activity, identifiable in time and typically associated with a specific consumer, device, or system process. Events may be recordable and may often generate logs or trigger system responses. A telecommunication action may be a specific, intentional operation initiated by a consumer or automated agent that triggers a well-defined functional response from the mobile network. Actions are the most granular level of interaction, often representing service requests or transactions.

[0014] The method may include using the action recommendations and selected action recommendation to update the propensity model for the consumer. Providing real time assessment of a consumer may include inputting historical metadata of the consumer to the propensity model.

[0015] The consumer’s current status may relate to an account status (e.g., the current balance in their account) and historical metadata may relate to historical telecommunication activity for the consumer. Historical telecommunication activity may include one or more telecommunication events and, for each telecommunication event, an account status at the time of the telecommunication event.

[0016] The method of building metadata may include obtaining information relating to action prompt messages received by consumers and related to events (e.g., marketing information from third- party vendors or service providers which may be sent to the consumer’s communication device). The historical metadata of a consumer may include information relating to action prompt messages received by the consumer and may include the consumer’s reaction to the action prompt messages. The consumer’s reaction may include a telecommunication action executed in response to and relating to the action prompt message.

[0017] The method may include preprocessing of large-scale data records including filtering and transforming the data records to standardized formats and storing the transformed data records into an appropriate data store that supports a large number of records.

[0018] The method may include obtaining training input data including training input data of trigger states of consumers, consumer profiles, consumers historical recharge data and / or action prompt messages received by consumers from a mobile network operator and / or third parties and training the propensity model.

[0019] The method may include applying the propensity model at regular intervals and / or trigger states and obtaining model output as real time assessment of a consumer.

[0020] A trigger may include an account status of a consumer account meeting a first predefined criterion (for example the balance of an account associated with the consumer falling to a predetermined low level); the account status of the consumer account meeting a second predefined criterion which is different from the first predefined criterion (for example, a zero balance of an account associated with the consumer); or an action execution failure notification relating to failure to execute a telecommunication action against the consumer account (for example, a failed transaction notification from a mobile network operator (MNO) indicating that a balance of an account associated with the consumer is insufficient for a consumer requested transaction). The method may include providing a push of one or more action recommendations to the consumer as push notifications. As examples, the push notifications may be Unstructured Supplementary Service Data (USSD), push SMS, an in-application notification, or a push notification over an over-the-top (OTT) technology (for example, WhatsApp). The action recommendation may be an airtime or a mobile bundle purchase (such as a specific quantity of airtime or a specific data bundle) and the method may include executing the telecommunication action relating to the selected action recommendation with a non-airtime digital payment method.

[0021] In accordance with a further aspect of the disclosure there is provided a system for providing a proactive digital transaction channel, the system comprising: a non-transitory computer-readable storage medium; and one or more processors coupled to the non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium comprises program instructions that, when executed on the one or more processors, cause the system to perform operations comprising: extracting data from raw telecommunication event detail records of a group of consumers; building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; providing real time assessment of a consumer at a given time and / or at a given trigger including: inputting a consumer’s current status to the propensity model; providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and, in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0022] In accordance with a further aspect of the disclosure there is provided a system for providing a proactive digital transaction channel including a memory for storing computer-readable program code and a processor for executing the computer-readable program code, the system comprising: a data extracting component for extracting data from raw telecommunication event detail records of a group of consumers; a metadata building component for building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; a propensity model training component for training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; an assessment component for providing real time assessment of a consumer at a given time and / or at a given trigger including: a consumer status component for inputting a consumer’s current status to the propensity model; a push component for providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and, a digital purchase component for in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0023] The system may include an action prompt obtaining component for obtaining information relating to action prompt messages received by consumers and related to events. The system may include a data preprocessing component for preprocessing of large-scale data records including a filtering component for filtering the data records and a transforming component for transforming the data records to standardized formats and storing the transformed data records into an appropriate data store that supports a large number of records. The system may include a training input data component for obtaining training input data including training input data of trigger states of consumers, consumer profiles, consumers historical recharge data and / or action prompt messages received by consumers from an MNO and / or third parties and training the propensity model. The system may include a model applying component for applying the propensity model at regular intervals and / or trigger states and a model output obtaining component for obtaining model output as real time assessment of a consumer.

[0024] In accordance with a further aspect of the disclosure there is provided a computer program product for providing a proactive digital transaction channel comprising a computer-readable medium having stored computer-readable program code for performing the steps of: extracting data from raw telecommunication event detail records of a group of consumers; building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; training a propensity model to output one or more probable action recommendations for a consumer in a given time period and / or at a given trigger; providing real time assessment of a consumer at a given time and / or at a given trigger including: inputting a consumer’s current status to the propensity model; providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0025] Further features provide for the computer-readable medium to be a non-transitory computer- readable medium and for the computer-readable program code to be executable by a processing circuit. Embodiments of the technology will now be described, by way of example only, with reference to the accompanying drawings.

[0026] BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In the drawings:

[0028] Figure 1 is a schematic diagram which illustrates an exemplary system for providing a digital transaction channel according to aspects of the present disclosure;

[0029] Figures 2A and 2B illustrates an exemplary method for providing a digital transaction channel according to aspects of the present disclosure;

[0030] Figure 3 illustrates exemplary screenshots of a high-level consumer journey according to aspects of the present disclosure;

[0031] Figures 4A to 4C illustrates flow diagrams of a detailed consumer journey according to aspects of the present disclosure;

[0032] Figure 5 is a schematic diagram of a general overview of a training and user of a machine learning model as implemented by aspects of the disclosure;

[0033] Figure 6 is a block diagram which illustrates exemplary components of a system for providing a digital transaction channel account according to aspects of the present disclosure; and,

[0034] Figure 7 illustrates an example of a computing device in which various aspects of the disclosure may be implemented.

[0035] DETAILED DESCRIPTION WITH REFERENCE TO THE DRAWINGS

[0036] Telecommunication activity as disclosed herein may refer to any occurrence, behaviour, or operation initiated by a consumer, network entity, or external system that takes place within, or in relation to, a mobile telephone network. This may involve the transmission, reception, or processing of signals or data over a telecommunication system. This may include signalling, data transmission, service usage, system operations, and user interactions. For example, a consumer making a call; a base station handing off a call to another tower; and a system-wide software update on network infrastructure.

[0037] A telecommunication event as disclosed herein may refer to a discrete and observable instance of a telecommunication activity, identifiable in time and typically associated with a specific consumer, device, or system process. Events may be recordable and may often generate logs or trigger system responses. For example, call setup initiation; short messaging service (SMS) delivery receipt; and authentication request during network attachment.

[0038] A telecommunication action as disclosed herein may refer to a specific, intentional operation initiated by a consumer or automated agent that triggers a well-defined functional response from the mobile network. Actions are the most granular level of interaction, often representing service requests or transactions. For example, a consumer purchasing a data bundle; a consumer dialling a Unstructured Supplementary Service Data (USSD) code to check balance; and a consumer sending a command via an app to enable roaming.

[0039] A system and a method for providing a digital transaction channel are provided. The system and method may include extracting data from raw telecommunication event detail records of a group of consumers and building metadata relating to telecommunication activity of consumers from the extracted data. The metadata may include information on a trigger type, consumer profiles, consumers historical recharge data and / or action prompt messages received by the consumers from third parties. The metadata may be stored in a data store for later retrieval and input into a propensity model.

[0040] A propensity model is provided by training it on the extracted data and the built metadata of telecommunication activity. The model may be trained to output one or more action recommendations for the consumer in a given time period (e.g., 24 hours or 48 hours) and / or at a given trigger (e.g., an account status of a consumer account meeting a first predefined criterion; the account status of the consumer account meeting a second predefined criterion which is different from the first predefined criterion, and the like). The propensity model may be a machine learning model.

[0041] The system and method may include providing a real time assessment of the consumer at a given time and / or at a given trigger including, inputting a consumer’s current status to the propensity model and providing a push of one or more action recommendations to the consumer based on the output of the propensity model. The one or more action recommendations may be an airtime or a mobile bundle purchase and the push may include transmitting an action prompt message (e.g., and offer to purchase airtime, or a mobile bundle). The consumer may select an action recommendation of the one or more action recommendations, and the system and method may include executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation. This may have the advantage in that the system and method may infer a consumer’s need to select an action recommendation of the one or more action recommendations based on the consumer’s current status and may subsequently execute, against the consumer’s account, a telecommunication action relating to the selected action recommendation. For example, the system and method may infer a consumer’s need to purchase an airtime or mobile bundle based on the consumer’s current status and may transmit in real time action recommendations most likely to be selected by the consumer at a given time and / or at a given trigger.

[0042] Figure 1 is a schematic diagram which illustrates an exemplary system (100) for providing a digital transaction channel. The system (100) may include a provisioning platform (102) and a mobile network operator (MNO) platform (104). The MNO platform (104) may be provided by a mobile network operator (MNO). The system (100) may further include a communication device (106), such as a mobile handset associated with a consumer (108), who typically purchases airtime and mobile data bundles to make use of a mobile telephone network (1 10) provided by the MNO. Although only one communication device (106) is illustrated, it should be appreciated that in a practical implementation there may be a plurality of these.

[0043] The MNO platform (104) and mobile telephone network (1 10) may collectively provide the infrastructure by way of which communication devices can communicate with each other and the broader public switched telephone network (PSTN) and the Internet. The mobile telephone network (1 10) may for example include one or more base station subsystems while the MNO platform (104) may provide back-end systems such as an Unstructured Supplementary Service Data (USSD) / General packet radio service (GPRS) gateway, Short Message Service Center (SMSC) and associated components (Visitor Location Register (VLR), Home Location Register (HLR), etc.).

[0044] The MNO platform (104) may include a front-end Application Programming Interface (API) (112) by way of which the provisioning platform (102) may interface with components or nodes of the MNO platform (104) and a communication gateway (1 14) which may for example include the GPRS / USSD gateway which facilitates the exchange of USSD messages between consumers and various applications. The communication gateway (114) and / or GPRS / USSD gateway may expose an API (e.g. one or more of Short Message Peer-to-Peer Protocol (SMPP), ParlayX, a proprietary Hypertext Transfer Protocol (HTTP)-based API or the like). The provisioning platform (102) may be provided by any suitable computing device or devices. The provisioning platform (102) may be configured to interface with the MNO platform (104), for example, via an appropriate integration component (116).

[0045] The provisioning platform (102) may include a digital transaction channel (120) for interfacing with the mobile telephone network (1 10) and in turn the communication device (106). The provisioning platform (102) may transmit and receive data and / or messages to and from the communication device (106) via the digital transaction channel (120) and the mobile telephone network (1 10). The provisioning platform (102) may for example be configured to receive a given trigger and to provide a push of one or more action recommendations to the communication devices of consumers.

[0046] The digital transaction channel (120) may for example include an SMSC integration component (122) and a USSD integration component (124), which may provide a USSD gateway adapter which may be configured to connect with a communication gateway (114) of the mobile operator platform (104) to send and receive USSD messages.

[0047] The provisioning platform (102) may have access to one or more consumer databases (130) in which records (132) associated with consumers are stored (e.g. raw telecommunication event detail records (EDR), which may include call detail records (CDRs) and records of other events) and may include or have access to a data preprocessing system (140) configured for large-scale processing of data records. The data preprocessing system (140) may include a data extracting component (142) for extracting data from the EDR / CDRs of consumers, a metadata building component (144) for building metadata from the extracted data, and a data store (146) for storing each consumer’s metadata. The data preprocessing system (140) may be configured to process data records continuously to update the data store (146) or may be configured to process data records at a given time and / or at a given trigger.

[0048] The data preprocessing system (140) may include: a an action prompt obtaining component (148) for obtaining information relating to action prompt messages received by consumers and related to events; and, a data preprocessing component (150) for preprocessing of large-scale data records. The data preprocessing component (150) may include a filtering component (152) for filtering the data records and a transforming component (154) for transforming the data records to standardized formats and storing the transformed data records into the appropriate data store (146) that supports a large number of records.

[0049] The provisioning platform (102) may include a transaction modelling system (160) and a real-time assessment component (170) for providing, training, and applying the propensity model as described herein.

[0050] The communication device (106) may be any suitable computing device with a communication functionality, such as a mobile phone or mobile handset. The communication device (106) may be associated with a device identifier which uniquely identifies the communication device (106), such as an International Mobile Equipment Identity (IM El) or the like.

[0051] The various gateways of the mobile operator platform (104) interface with corresponding components of the provisioning platform (102) so as to provide a level of integration between the two. Accordingly, the provisioning platform (102) may interact with the mobile operator platform (104), for example, to access and / or update or manipulate account information of an account associated with the consumer.

[0052] The system (100) described above may implement a method for providing a digital transaction channel. An exemplary method for providing a digital transaction channel is illustrated in the flow diagram of Figures 2A and 2B. The method may be conducted by the provisioning platform (102) or one or more computing devices thereof.

[0053] The method may include pre-processing (210) of consumer data, training (220) a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger, inferring (230) a consumer’s need to select an action recommendation of the one or more action recommendations, and executing (240) against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0054] Pre-processing (210) of consumer data may include extracting (211 ) primary data of transactions from the EDR / CDRs of consumers and pre-processing (212) of large-scale data records, including filtering and transforming the data records to standardized formats and storing the transformed data records into an appropriate data store that supports a large number of records.

[0055] The pre-processing (210) may include building (213) metadata of transaction data to determine metadata relating to telecommunication activity of consumers. The historical metadata of a consumer may include information relating to action prompt messages (such as marketing information, or the like) received by the consumer and the consumer’s reaction or absence thereof to the action prompt messages. The building (213) of metadata may include obtaining (214) the historical metadata relating to the action prompt messages, including for example one or more of: message content; message type (such as call to action, informational, etc.); time and date of transmission; time and date of receipt; context of message (including, e.g., number of consumers to whom message was sent, data or indication relating to personalization of message to the consumer, etc.); and the like.

[0056] EDR / CDRs are of a transactional nature, they describe an event that took place in a specific moment in time and its attributes (e.g., timestamp) are immutable. Primary data of transactions which may be extracted from the EDR / CDRs may include:

[0057] • Airtime purchases, which are events where a consumer spends an amount of money in order to increase his airtime account balance with the Mobile Network Operator.

[0058] • Bundle purchases, which are events where a subscriber buys a product offered by the Operator that combines a number of services (voice, data, etc.) with specific volumes or usage limits for each product. Such products are commonly referred to as bundles.

[0059] • The price and time at which a bundle was purchased.

[0060] • The status of the consumers (e.g., account balance) at the time a bundle was purchased.

[0061] Large-scale raw data records and may have no single or well-defined format. The data records may be in the form of telecommunication records including a first type of data records being consumer data records with static values of consumer attributes such as Know Your Customer (KYC) records and a second type of data records relating to transactional records with timestamps of events being EDR / CDRs. Pre-processing (212) of large-scale data records may include receiving raw data records including the first type of data records and the second type of data records relating to transactional records with timestamps of events; filtering and transforming the data records to standardized formats; persisting the transformed data records into two different data stores, including a first data store for static data relating to entities for retrieval based on static values including a unique identifier of a consumer, and a second data store for transactional data for retrieval based on timestamps of events; grouping different categories of transactional records according to the unique identifiers of entities and persisting the categories to different tables including maintaining a table of stable size for efficient access of a category of transactional records that is most frequently accessed; and retrieving features relating to the consumer from multiple tables for processing to provide an output relating to the consumer.

[0062] Filtering and transforming the data records to standardized formats may include applying filtering rules to extract relevant records from transactional records including one or more of the group of: records relating to events including: calls, messages, monetary events, data usage events; lifecycle events; advance usage events; loan events; mobile wallet events. Grouping different categories of transactional records according to the unique identifiers of an entity may include categories of open credit advances and closed credit advances, and a category of transactional records that is most frequently accessed are open credit advances available for subsequent analysis.

[0063] Action prompt messages from the MNO platform (104) or from third-party vendors or service providers may be sent to the consumer’s communication device (106) and the method may include obtaining (214) the action prompt messages and any associated metadata.

[0064] Training (220) the propensity model may include obtaining (221 ) training input data of metadata, obtaining (222) training input data of action prompt messages, obtaining (223) training input data of trigger states, obtaining (224) training output data of action recommendations and training (225) the model on training datasets of input and output data with defined weights of variables. Each action recommendation may describe a telecommunication action that can be initiated by the consumer, for example, a consumer purchasing airtime or a data bundle.

[0065] Obtaining (221 ) training input data to model of metadata, may include retrieving consumers’ metadata from the data store (146). Obtaining (222) training input data of action prompt messages, may include the type of action prompt messages sent to and / or responded to by a consumer. For example, the type of marketing activities sent to and / or responded to by a consumer. Obtaining (223) training input data of trigger states, may include obtaining information of a trigger such as an account status of a consumer account meeting a first predefined criterion (e.g., a low balance); or the account status of the consumer account meeting a second predefined criterion which is different from the first predefined criterion (e.g., a zero balance); or an action execution failure notification relating to failure to execute a telecommunication action against the consumer account (e.g., a failed transaction notification from a mobile network operator indicating that a balance of an account associated with a consumer is insufficient for a consumer requested transaction (such as trying to make a phone call or send a message)). Obtaining (224) training output data of action recommendations may include data on consumer historical recharge, and the method may include training (225) the model on training datasets of input and output data with defined weights of variables.

[0066] The method may provide real time assessment of a consumer at a given time and / or at a given trigger in the inference stage (230) using the trained propensity model. Inferring (230) a consumer’s need to select an action recommendation of the one or more action recommendations may include providing (231 ) the trained model, and applying (232) the trained model at a regular interval and / or trigger state, inputting (233) a trigger state of a consumer (such as consumer account balance statuses), running (234) the model with historical recharge data of the consumer and outputting (235) one or more action recommendations for a consumer. Each action recommendation may describe a telecommunication action that can be initiated by the consumer, for example, a consumer purchasing airtime or a data bundle.

[0067] The method may include executing (240) against the consumer’s account, a telecommunication action relating to the selected action recommendation, including obtaining (241 ) model output including one or more action recommendations for the consumer as real time assessment at a regular interval and / or trigger state, pushing (242) the one or more action recommendations to the consumer, receiving (243) consumer selection of an action recommendation of the one or more action recommendations, and in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing (244), against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0068] Obtaining the model output may include inputting a trigger state of the consumer into the propensity model and receiving one or more action recommendations as an output from the propensity model. Pushing the one or more action recommendations may include transmitting an action prompt message, including the one or more action recommendations, to a user communication device via the mobile telephone network. In some examples, the action recommendation is an offer relating to a network product for purchase and executing the action may include purchasing the offer with a non-airtime digital payment method. A non-airtime digital payment may include payments with a bank card, a bank account, a bank USSD, payment with a mobile money wallet account and the like.

[0069] Figure 3 illustrates exemplary screenshots (301 -305) of a high-level consumer journey according to aspects of the present disclosure. According to this embodiment, an account status of a consumer account may meet a second predefined criterion, for example, a consumer mobile data bundle account has been depleted. This zero-account balance state may provide a trigger for providing and applying the trained propensity model. The consumer may then be provided with a push, in this case, a transmitting an action prompt message including two action recommendations (301 ) based on the output of the propensity model i.e., a first action recommendation for the consumer to select 1 GB of mobile data bundle at ZAR50 and a second action recommendation for the consumer to select 2GB of mobile data bundle at ZAR75. The consumer may select the first action recommendation and the provisioning platform may execute (303) the telecommunication action (in this case payment of the mobile data bundle) using with a non-airtime digital payment method, in this embodiment, a mobile money wallet account. The propensity model is accordingly used to infer the consumer’s need to perform an action (in this case to purchase an airtime or mobile bundle) based on the consumer’s current status, determine the appropriate action recommendations to provide to the consumer, and the provisioning platform executes the telecommunication action in response to receiving a consumer selection, in this case the digital payments and activation of the telecommunication products selected by the consumer.

[0070] Figures 4A to 4C illustrate flow diagrams of a detailed consumer journey. According to this embodiment, an account status of a consumer account may meet a first predefined criterion, for example, a consumer mobile data bundle account may have fallen below a predetermined threshold or the account status of the consumer account may meet a second predefined criterion which is different from the first predefined criterion, for example, the consumer mobile data bundle account may have been depleted (401 ). This provides a trigger in the form of either: a TNP file for the low balance may be transmitted to the provisioning platform or an HTTP notification for the data depletion. The trigger state is received at the provisioning platform and the propensity model provides an assessment and prepares one or more action recommendations. The one or more action recommendations are then pushed via USSD to the communication (106) device of the consumer. If the USSD session expires (405) with no input from the consumer, a fallback SMS with a pull USSD long sting may alternatively be pushed to the communication (106) device. The consumer may then select an action recommendation and the consumer’s input is registered as shown in Figure 4A. Other push notifications may be used including an in-application notification, or a push notification over an over-the-top (OTT) technology (for example, WhatsApp).

[0071] The provisioning platform may facilitate execution of the telecommunication action, in this case the digital purchase of the selected offer by pushing a USSD menu with available payment channels. In this embodiment, the consumer may be provided with a menu (41 1 ) to select one of a bank USSD, bank account, mobile money wallet account or credit card. The consumer may use a bank USSD if his mobile communication service line is active on bank USSD. The consumer may select an option, and the provisioning platform may provide a bank list menu (413) for the consumer to select a bank account. Upon selecting a bank account, the consumer may be provided with a USSD confirmation message (415) and an SMS (417) with a payment USSD string to complete the purchase. The consumer may proceed with the purchase and the transaction log record may be saved as shown in Figure 4B. It is appreciated that in some embodiments, the consumer may be provided with other payment methods, for example, “Apple Pay”, “Google Pay”, “Samsung Pay” or the like.

[0072] Where the payment is unsuccessful (421 ), the transaction is marked as a failed transaction (423) and where the payment is successful (422), the consumer is provided with the airtime or bundle (425). If the provisioning is unsuccessful (427) the payment is refunded, and the refund transaction log record is saved (429). If the provisioning is successful (429) the product is activated (433) and the activation log record is saved as shown in Figure 4C.

[0073] The terms artificial intelligence (Al), machine learning, and deep learning may be used interchangeably throughout this disclosure when referring to trained Al models. Machine learning may be considered a sub-branch of Al and deep learning may be considered a sub-branch of machine learning. Deep learning is a form of machine learning that uses a layered network, referred to as an artificial neural network (ANN). Any Al, machine learning and deep learning system may rely on an underlying model. The model may be tailored to a specific use case. Although Al is considered the broadest of term, it is common that any Al system includes some form of machine learning, with some systems further including deep learning.

[0074] Some examples of machine learning models may include, but are not limited to: decision trees, random forest regression, support-vector machines, K-means clustering, regression analysis, Gaussian processes, and the like.

[0075] Machine learning models (as well as deep learning models) may be categorised into classification or regression. Classification models may classify an input into one or more of a set of classifications, with the output being one of discrete classifications. Regression models may determine an output that may be a value or output across a continuous output range. A regression model may estimate a relationship between an input to an output.

[0076] ANNs may include a variety of structures, which are referred to as architectures. Different architectures are suitable for different use cases. Examples of ANN architectures may include convolutional neural networks or recurrent neural networks. Convolutional neural networks may be suitable for image-based data or multi-dimensional input data. Recurrent neural networks, such as long-short term memory networks, may be more suitable for time series applications.

[0077] An ANN may consist of interconnected units, commonly referred to as neurons, as they are inspired by and resemble neurons of the brain. The units may be made up of nodes and edges forming a connected network. The edges may connect nodes together. ANNs may be configured in the form of a layered structure with an input at the first layer and an output provided by the final layer. The layers between the first layer and final layer are hidden layers.

[0078] The input layer may include one or more nodes. An edge may extend from each node. Each edge may be connected to a node in a subsequent hidden or output layer. Each node may include more than one edge that connects the node to a plurality of other nodes in other layers. In some examples, an edge may feed back into a previous node in a preceding layer (a node not in subsequent layers but in a further layer), or to a different node in the same layer.

[0079] The output of a node may be computed by an activation function, which may be a linear or a nonlinear function of the sum of the inputs into each node in each layer. The output value of each node in the preceding layer is multiplied by a weighting value, which determines the strength of each nodes’ output value. Finally, the value that is determined at the node(s) of the final layer is the output of the ANN. For regression type ANNs, the output may contain only a single node with a value, or many nodes. For classification type ANNs, the output may include multiple nodes, where each node is an output of the probability of a classification type.

[0080] More complex ANNs are better suited to specific tasks. In addition to the weights and activation functions of a regular ANN, a convolutional neural network applies a filter (or a kernel) onto a two- dimensional data structure, which may reduce the number of edges between the hidden layers in the neural network. This may in turn reduce the number of weights within the neural network. A convolutional neural network may find application in image-based tasks, where image data may be structured as a two-dimensional data structure. A convolutional neural network may be extended into further dimensions by increasing the dimensions of the filter / kernel to match the number of dimensions of the input data.

[0081] Recurrent neural networks include a recurrent unit. This recurrent unit may maintain a hidden state over time, thereby providing a pseudo-memory capability. Such models may find application in time series or sequential operations, such as speech or text. Multiple recurrent units may be connected to each other, where the output of one unit at a first timestep may be used as an input into another recurrent unit at a second timestep. Examples of recurrent neural networks include, but are not limited to, long short-term memory networks, and gated recurrent units. Transformers are another form of deep learning architectures well suited for sequential based data. Transformers may utilise a self-attention mechanism instead of recurrence (such as in a recurrent neural network).

[0082] A trained Al model may be configured to run on a computing device, such as a Raspberry Pi ™, a NVIDIA Jetson Nano ™ developer kit, or a standard personal computer (PC) including a graphical processing unit (GPU). Example computing devices may be designed to perform specific computational tasks, which may include running multiple neural networks in parallel for applications including image classification, object detection, segmentation, and speech processing. In some cases, a trained Al model may be configured to run on a computing device in the form of a large computing system, such as a computing cluster (such as that found in a data centre). Training Al models may be a computationally intensive and time consuming. Models may thus be trained on a computing device provided by a large computing infrastructure or a cloud computing infrastructure that can be accessed over a network. These resources allow for dynamic computing resources to be dedicated to training a deep neural network, after which the trained model can be downloaded to run on a separate application.

[0083] Figure 5 illustrates a general overview of a training and user of a machine learning model (514). The training involves a data preparation process (511 ) to prepare training data (512). A training process (513) receives the prepared training data (512) and iteratively updates a machine learning model until a predefined quality criteria and accuracy criteria are achieved. A machine learning model (514) is output at the end of the training process (513) to be used in the runtime process (522). When the model (514) is in use, an input (521 ) is received into the runtime process (522) that uses the model (514) to obtain an output (523) that may be used in a downstream process (524). The computation of the runtime process (522) is often referred to as ‘inference’.

[0084] The training process (513) may be computationally demanding and time consuming. To successfully train a machine learning model, very large datasets are often used which are stored on databases. The training process may be performed on a large computing cluster which may access the database to obtain the training data when required. Additionally, the trained machine learning model (514) may be stored on the database. The runtime process (522) may run on an end user device by downloading the machine learning model (514) over a network from a database, or the runtime process (522) may run on a large computing infrastructure such as a computing cluster. An example embodiment of interacting with the machine learning model (514) may include an end user device, such as a mobile device or a computer which may obtain or be the source of the input data (521 ), transmit the input data (521 ) over a network to a computing cluster to perform the runtime process (522).

[0085] Various types of machine learning model may be used as known in the art. The model may be a supervised learning model in which the machine learning model (514) may be trained using supervised leaning. Supervised learning requires a first set of input data with known output data, also referred to as training data (512). The data preparation process (51 1 ) may involve labelling the data to provide a known output for the raw incoming data, thereby generating the training data

[0086] (512). A second set of input data with known output data may be used to test the machine learning model during and after training, also referred to as validation data. The task of the training process

[0087] (513) is to minimize the difference (or error) between the output of the final layer of the machine learning model (514) and the known output data of the training data (512). The training procedure modifies the machine learning model (513) such that the difference is minimized.

[0088] Various components may be provided for implementing the method described above with reference to Figures 2A and 2B. Figure 6 is a block diagram which illustrates exemplary components which may be provided by a system for providing a digital transaction channel according to aspects of the present disclosure. The system includes a provisioning platform (102) which may be provided by one or more computing devices.

[0089] The provisioning platform (102) may include a processor (602) for executing the functions of components described below, which may be provided by hardware or by software units executing on the provisioning platform (102). The software units may be stored in a memory component (604) and instructions may be provided to the processor (602) to carry out the functionality of the described components. In some cases, for example in a cloud computing implementation, software units arranged to manage and / or process data on behalf of the provisioning platform (102) may be provided remotely.

[0090] The provisioning platform (102) may include the data preprocessing system (140) as described in Figure 1 for preprocessing consumer records. The provisioning platform (102) may also include a transaction modelling system (160) for modelling the propensity model and a real-time assessment system (170) for real-time assessment of a consumer activity.

[0091] The transaction modelling system (160) may include a training input data component (616) arranged to obtain training input data including training input data of trigger states of consumers, consumer profiles, consumers historical recharge data and / or action prompt messages received by consumers from third parties and training the propensity model. The transaction modelling system (160) may include a propensity model training component (606) arranged to train a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger. The transaction modelling system (160) may include a model applying component (618) arranged to apply the propensity model at regular intervals and / or trigger states and a model output obtaining component (620) arranged to obtaining model output as real time assessment of a consumer.

[0092] The real-time assessment component (608) may be provided and arranged to provide real time assessment of a consumer at a given time and / or at a given trigger. The real-time assessment component (608) may include a consumer status component (610) arranged to input a consumer’s current status to the propensity model; a push component (612) arranged to provide a push of one or more action recommendations to the consumer based on the output of the propensity model; and a digital purchase component (614) arranged to, in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, execute, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

[0093] Figure 7 illustrates an example of a computing device (700) in which various aspects of the disclosure may be implemented. The computing device (700) may be embodied as any form of data processing device including a personal computing device (e.g. laptop or desktop computer), a server computer (which may be self-contained, physically distributed over a number of locations), a client computer, or a communication device, such as a mobile phone (e.g. cellular telephone), satellite phone, tablet computer, personal digital assistant or the like. Different embodiments of the computing device may dictate the inclusion or exclusion of various components or subsystems described below.

[0094] The computing device (700) may be suitable for storing and executing computer program code. The various participants and elements in the previously described system diagrams may use any suitable number of subsystems or components of the computing device (700) to facilitate the functions described herein. The computing device (700) may include subsystems or components interconnected via a communication infrastructure (705) (for example, a communications bus, a network, etc.). The computing device (700) may include one or more processors (710) and at least one memory component in the form of computer-readable media. The one or more processors (710) may include one or more of: CPUs, graphical processing units (GPUs), microprocessors, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs) and the like. In some configurations, a number of processors may be provided and may be arranged to carry out calculations simultaneously. In some implementations various subsystems or components of the computing device (700) may be distributed over a number of physical locations (e.g. in a distributed, cluster or cloud-based computing configuration) and appropriate software units may be arranged to manage and / or process data on behalf of remote devices.

[0095] The memory components may include system memory (715), which may include read only memory (ROM) and random access memory (RAM). A basic input / output system (BIOS) may be stored in ROM. System software may be stored in the system memory (715) including operating system software. The memory components may also include secondary memory (720). The secondary memory (720) may include a fixed disk (721 ), such as a hard disk drive, and, optionally, one or more storage interfaces (722) for interfacing with storage components (723), such as removable storage components (e.g. magnetic tape, optical disk, flash memory drive, external hard drive, removable memory chip, etc.), network attached storage components (e.g. NAS drives), remote storage components (e.g. cloud-based storage) or the like.

[0096] The computing device (700) may include an external communications interface (730) for operation of the computing device (700) in a networked environment enabling transfer of data between multiple computing devices (700) and / or the Internet. Data transferred via the external communications interface (730) may be in the form of signals, which may be electronic, electromagnetic, optical, radio, or other types of signal. The external communications interface (730) may enable communication of data between the computing device (700) and other computing devices including servers and external storage facilities. Web services may be accessible by and / or from the computing device (700) via the communications interface (730).

[0097] The external communications interface (730) may be configured for connection to wireless communication channels (e.g., a cellular telephone network, wireless local area network (e.g. using Wi-Fi™), satellite-phone network, Satellite Internet Network, etc.) and may include an associated wireless transfer element, such as an antenna and associated circuitry. The external communications interface (730) may include a subscriber identity module (SIM) in the form of an integrated circuit that stores an international mobile subscriber identity and the related key used to identify and authenticate a subscriber using the computing device (700). One or more subscriber identity modules may be removable from or embedded in the computing device (700).

[0098] The computer-readable media in the form of the various memory components may provide storage of computer-executable instructions, data structures, program modules, software units and other data. A computer program product may be provided by a computer-readable medium having stored computer-readable program code executable by the central processor (710). A computer program product may be provided by a non-transient or non-transitory computer- readable medium, or may be provided via a signal or other transient or transitory means via the communications interface (730).

[0099] Interconnection via the communication infrastructure (705) allows the one or more processors (710) to communicate with each subsystem or component and to control the execution of instructions from the memory components, as well as the exchange of information between subsystems or components. Peripherals (such as printers, scanners, cameras, or the like) and input / output (I / O) devices (such as a mouse, touchpad, keyboard, microphone, touch-sensitive display, input buttons, speakers and the like) may couple to or be integrally formed with the computing device (700) either directly or via an I / O controller (735). One or more displays (745) (which may be touch-sensitive displays) may be coupled to or integrally formed with the computing device (700) via a display or video adapter (740).

[0100] The computing device (700) may include a geographical location element (755) which is arranged to determine the geographical location of the computing device (700). The geographical location element (755) may for example be implemented by way of a global positioning system (GPS), or similar, receiver module. In some implementations the geographical location element (755) may implement an indoor positioning system, using for example communication channels such as cellular telephone or Wi-Fi™ networks and / or beacons (e.g. Bluetooth™ Low Energy (BLE) beacons, iBeacons™, etc.) to determine or approximate the geographical location of the computing device (700). In some implementations, the geographical location element (755) may implement inertial navigation to track and determine the geographical location of the communication device using an initial set point and inertial measurement data.

[0101] The foregoing description has been presented for the purpose of illustration; it is not intended to be exhaustive or to limit the technology to the precise forms disclosed. Persons skilled in the relevant art can appreciate that many modifications and variations are possible in light of the above disclosure.

[0102] Any of the steps, operations, components or processes described herein may be performed or implemented with one or more hardware or software units, alone or in combination with other devices. Components or devices configured or arranged to perform described functions or operations may be so arranged or configured through computer-implemented instructions which implement or carry out the described functions, algorithms, or methods. The computer- implemented instructions may be provided by hardware or software units. In one embodiment, a software unit is implemented with a computer program product comprising a non-transient or non- transitory computer-readable medium containing computer program code, which can be executed by a processor for performing any or all of the steps, operations, or processes described. Software units or functions described in this application may be implemented as computer program code using any suitable computer language such as, for example, Java™, C++, or Perl™ using, for example, conventional or object-oriented techniques. The computer program code may be stored as a series of instructions, or commands on a non-transitory computer-readable medium, such as a random access memory (RAM), a read-only memory (ROM), a magnetic medium such as a hard-drive, or an optical medium such as a CD-ROM. Any such computer-readable medium may also reside on or within a single computational apparatus, and may be present on or within different computational apparatuses within a system or network. Flowchart illustrations and block diagrams of methods, systems, and computer program products according to embodiments are used herein. Each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, may provide functions which may be implemented by computer readable program instructions. In some alternative implementations, the functions identified by the blocks may take place in a different order to that shown in the flowchart illustrations.

[0103] Some portions of this description describe the examples in terms of algorithms and symbolic representations of operations on information. These algorithmic descriptions and representations, such as accompanying flow diagrams, are commonly used by those skilled in the data processing arts to convey the substance of their work effectively to others skilled in the art. These operations, while described functionally, computationally, or logically, are understood to be implemented by computer programs or equivalent electrical circuits, microcode, or the like. The described operations may be embodied in software, firmware, hardware, or any combinations thereof.

[0104] The language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the present disclosure be limited not by this detailed description, but rather by any claims that issue on an application based hereon. Accordingly, the present disclosure is intended to be illustrative, but not limiting, of the scope of any accompanying claims.

[0105] Finally, throughout the specification and any accompanying claims, unless the context requires otherwise, the word ‘comprise’ or variations such as ‘comprises’ or ‘comprising’ will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.

Claims

CLAIMS:

1. A computer-implemented method for providing a proactive digital transaction channel, comprising: extracting data from raw telecommunication event detail records of a group of consumers; building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; providing real time assessment of a consumer at a given time and / or at a given trigger including: inputting a consumer’s current status to the propensity model; providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

2. The method of claim 1 , including using the action recommendations and selected action recommendation to update the propensity model for the consumer.

3. The method of claim 1 , wherein providing real time assessment of a consumer includes inputting historical metadata of the consumer to the propensity model.

4. The method of claim 3, wherein the consumer’s current status relates to an account status and historical metadata relates to historical telecommunication activity for the consumer including one or more telecommunication events and, for each telecommunication event, an account status at the time of the telecommunication event.

5. The method of claim 1 , wherein building metadata includes obtaining information relating to action prompt messages received by consumers and related to events.

6. The method of claim 3, wherein historical metadata of a consumer includes information relating to action prompt messages received by the consumer and the consumer’sreaction to the action prompt messages, wherein the consumer’s reaction includes a telecommunication action executed in response to and relating to the action prompt message.

7. The method of claim 1 , including preprocessing of large-scale data records including filtering and transforming the data records to standardized formats and storing the transformed data records into an appropriate data store that supports a large number of records.

8. The method of claim 1 , including obtaining training input data including training input data of trigger states of consumers, consumer profiles, consumers historical recharge data and / or action prompt messages received by consumers from a mobile network operator and / or third parties and training the propensity model.

9. The method of claim 1 , including applying the propensity model at regular intervals and / or trigger states and obtaining model output as real time assessment of a consumer.

10. The method of claim 9, wherein a trigger includes the an account status of a consumer account meeting a first predefined criterion; the account status of the consumer account meeting a second predefined criterion which is different from the first predefined criterion; or an action execution failure notification relating to failure to execute a telecommunication action against the consumer account.11 . The method of claim 1 , including providing a push of one or more action recommendations to the consumer as push notifications.

12. The method of claim 1 1 , wherein the one or more action recommendation is an airtime or a mobile bundle purchase and the method includes executing the telecommunication action relating to the selected action recommendation with a non-airtime digital payment method.

13. A system for providing a proactive digital transaction channel including a memory for storing computer-readable program code and a processor for executing the computer- readable program code, the system comprising: a data extracting component for extracting data from raw telecommunication event detail records of a group of consumers; a metadata building component for building metadata relating to telecommunication activity of consumers from the extracted data for use as modelparameter inputs; a propensity model training component for training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; an assessment component for providing real time assessment of a consumer at a given time and / or at a given trigger including: a consumer status component for inputting a consumer’s current status to the propensity model; a push component for providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and a digital purchase component for, in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

14. The system of claim 13, including an action prompt obtaining component for obtaining information relating to action prompt messages received by consumers and related to events.

15. The system of claim 13, including a data preprocessing component for preprocessing of large-scale data records including a filtering component for filtering the data records and a transforming component for transforming the data records to standardized formats and storing the transformed data records into an appropriate data store that supports a large number of records.

16. The system of claim 13, including a training input data component for obtaining training input data including training input data of trigger states of consumers, consumer profiles, consumers historical recharge data and / or action prompt messages received by consumers from an MNO and / or third parties and training the propensity model.

17. The system of claim 13, including a model applying component for applying the propensity model at regular intervals and / or trigger states and a model output obtaining component for obtaining model output as real time assessment of a consumer.

18. A system for providing a proactive digital transaction channel, the system comprising: a non-transitory computer-readable storage medium; and one or more processors coupledto the non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium comprises program instructions that, when executed on the one or more processors, cause the system to perform operations comprising: extracting data from raw telecommunication event detail records of a group of consumers; building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; providing real time assessment of a consumer at a given time and / or at a given trigger including: inputting a consumer’s current status to the propensity model; providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and, in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, against the consumer’s account, a telecommunication action relating to the selected action recommendation.

19. A computer program product for providing a proactive digital transaction channel comprising a computer-readable medium having stored computer-readable program code for performing the steps of: extracting data from raw telecommunication event detail records of a group of consumers; building metadata relating to telecommunication activity of consumers from the extracted data for use as model parameter inputs; training a propensity model to output one or more action recommendations for a consumer in a given time period and / or at a given trigger; providing real time assessment of a consumer at a given time and / or at a given trigger including: inputting a consumer’s current status to the propensity model; providing a push of one or more action recommendations to the consumer based on the output of the propensity model, including transmitting an action prompt message including the one or more action recommendations; and in response to receiving a consumer selection of an action recommendation of the one or more action recommendations, executing, againstthe consumer’s account, a telecommunication action relating to the selected action recommendation.

Images