Self-service terminal (SST) predictive error analysis
A cash tender module (CTM) of a self-service terminal (SST) records transaction media usage and error or warning codes for media peripheral devices of the SST in one or more log files. The log files are processed to discover patterns and relationships between the error/warning codes. Predicted error/warning codes are generated for the peripherals based on the patterns and relationships. Service records are automatically and proactively generated based on the predicted error/warning codes. In an embodiment, the service records include links to knowledge articles that provide step-by-step actions to resolve the error/warning codes. In an embodiment, real-time media counts by denomination are render into a dashboard interface for monitoring in real time the media in the SST.
1 . A method, comprising:
receiving telemetry data for media peripherals of a terminal, wherein the media peripherals accept and dispense currency;
predicting an error code for a certain media peripheral using the telemetry data by providing the telemetry data as input to a machine-learning model to identify sequences of error codes specific to media peripheral failures and receiving the predicted error code as output from the machine-learning model, wherein the predicted error code includes a predicted time period within which the error code is expected to occur; and
determining whether to automatically schedule a service call to resolve the error code before the error code occurs on the terminal;
wherein determining further includes:
identifying an existing service call scheduled for a store associated with the terminal and piggybacking the service call with the existing service call scheduled based on a severity of the error code, wherein the severity of the error code is a hard fault or a soft fault, the hard fault indicating that the terminal is unable to perform a currency transaction and the soft fault indicating that the terminal is able to perform the currency transaction.
2 . The method of claim 1 , wherein receiving further includes receiving the telemetry data in log files provided by an agent executing on the terminal.
3 . The method of claim 1 , wherein receiving further includes receiving the telemetry data in log files provided by a cash tender module executing on the terminal.
4 . The method of claim 1 , wherein determining further includes immediately scheduling the service call when the severity of the error code is the hard fault.
5 . The method of claim 1 , wherein predicting further includes linking the error code to a knowledge article that provides for resolution actions to resolve the error code associated with the certain media peripheral.
6 . The method of claim 5 , wherein determining further includes embedding a link to the knowledge article in a service record associated with the existing service call that is scheduled.
7 . The method of claim 1 , wherein determining further includes interacting with a maintenance and support system to generate a service record for the service call.
8 . The method of claim 1 further comprising:
maintaining real-time currency counts by denomination for the media peripherals based on the telemetry data and based on previous telemetry data received for the media peripherals; and
rendering the real-time currency counts by denomination in a dashboard interface for a store associated with the terminal.
9 . The method of claim 8 further comprising, color coding the real-time currency counts based on an optimal currency count for each denomination within the dashboard interface.
10 . The method of claim 1 further comprising, providing the method as a cloud service that interacts with a server associated with the terminal.
11 . A method, comprising:
training a machine learning model (model) on telemetry data specific to currency peripheral operations and service records associated with media peripherals of a terminal to predict error codes that are likely to occur but have not yet occurred on the media peripherals by training the model to identify patterns and relationships between error codes and sequences of the error codes using telemetry data and based on service records, wherein the patterns identify when media peripherals will be unable to accept or dispense currency;
providing current telemetry data as input to the model;
receiving a predicted error code for a certain media peripheral as output from the model, wherein the predicted error code includes a future time period during which the predicted error code is expected to occur; and
scheduling a service call to address the predicted error code on the certain media peripheral with a maintenance and support system;
wherein scheduling further includes:
piggybacking the service call that is scheduled with an existing service call already scheduled for a store where the terminal is located within the maintenance and support system when the existing service call is scheduled to occur before the future time period.
12 . The method of claim 11 further comprising:
updating real-time currency counts by currency denomination based on the current telemetry data; and
rendering the real-time currency counts by currency denomination for the terminal within a dashboard interface.
13 . The method of claim 11 further comprising:
receiving updated telemetry data for the media peripherals;
identifying a current error code in the updated telemetry data linked to a hard fault of a specific media peripheral; and
scheduling an immediate service call to address the hard fault on the specific media peripheral with the maintenance and support system.
14 . The method of claim 11 , wherein providing further includes receiving the current telemetry data in real time or near real time from the terminal.
15 . The method of claim 11 , wherein scheduling further includes linking the predicted error code to a knowledge article that provides resolution steps to resolve the predicted error code on the certain media peripheral and providing a link to the knowledge article in a service record associated with the service call that is scheduled.
16 . A system, comprising:
at least one server comprising a processor and a non-transitory computer-readable storage medium;
the non-transitory computer-readable storage medium comprises executable instructions; and
the executable instructions when executed on the processor cause the processor to perform operations comprising:
using telemetry data associated with media peripherals of a terminal to predict an error code that has not yet occurred on a certain media peripheral of a terminal by training a machine-learning model on the telemetry data and service records specific to media peripheral failures to predict error codes and sequences of error codes that will render a particular media peripheral unable to accept or dispense currency, and providing the telemetry data as input to the machine-learning model;
linking the error code to a knowledge article that provides descriptive steps for resolving the error code on the certain media peripheral;
generating a service record that includes the error code and a predicted time frame when the error code will occur, and scheduling a service call to a certain media terminal to resolve the error code in advance of the error code being reported from the terminal;
providing a link to the knowledge article within the service record; and
determining whether the error code is associated with a hard fault that prevents currency transactions or a soft fault that does not prevent currency transactions, when the soft fault is identified determining when to optimally schedule the service call, and when the hard fault is identified determining to immediately schedule the service call.
17 . The system of claim 16 , wherein the terminal is a self-service terminal.