Abstract: Visual features of a vehicle, a vehicle's orientation with respect to a transaction terminal, and an individual associated with the vehicle are derived from video captured of the terminal and a surrounding area of the terminal. Additional information associated with brute force attacks in a location associated with the terminal are obtained as additional features. The visual features and the additional features are provided as input to a trained Machine-Learning Model (MLM) and the MLM provides as output a confidence value representing a prediction as to whether the current vehicle and the current individual are potentially about to engage in a brute force attack on the terminal. When the confidence value exceeds a threshold value, enhanced feature detection is enabled on the video, external systems are notified, and an incident packet of information is assembled.

Abstract: Payment card slot tampering at a transaction terminal is detected in real time. One or more images from one or more cameras are analyzed. The camera(s) is/are focused overhead of the terminal on an area adjacent to and in front of a card slot for a card reader of the terminal. The images are analyzed for determining whether hands of arms for an operator of the terminal are performing actions that indicate the operator may be attempting to install a card skimmer/shimmer into the card slot and/or attempting to remove an existing skimmer/shimmer from the card slot. When card slot tampering event is detected at the terminal a variety of automated actions are processed, such as shutting down the card reader, shutting down the terminal, confiscating any payment card inserted into the card slot, notifying a financial institution, and/or notifying legal authorities.

Abstract: A consumer is identified engaged in an activity at a specific physical location. A context is formulated for the consumer based on the activity, the location, preferences associated with the consumer, and a history associated with the consumer. A natural language voice-based assistance session is initiated with the consumer. The session follows the consumer as the consumer travels and remains engaged in the activity. In an embodiment, the session is seamlessly maintained and is seamlessly transferred from a plurality of different devices as the consumer is engaged in the activity and travels. In an embodiment, the session is initiated without the consumer requesting the session.

Abstract: A user makes a request from a requesting device for access to a secure operation associated with a network-based service. A first biometric authentication is processed for the request and at least one second scope-based authentication is processed for the request based on an analysis of a physical environment for the requesting device. A determination is made based on at least the first biometric authentication and the scope-based authentication whether the secure operation can be: processed on behalf of the user by the network-based service, not processed at all, or processed only if requested from a specific medium/channel associated with a specific device, which may or may not be the requesting device.

Abstract: A consumer is identified at a transaction terminal. A natural-language voice chatbot is initiated for the transaction. The chatbot is configured with store information for a store associated with the terminal and, optionally, with preferences registered to the consumer. A natural language voice dialogue using, at least the store information, is initiated between the chatbot and the consumer while the consumer is at the terminal and during the transaction. The consumer may order store items from the store and pay for the items through the dialogue. In an embodiment, the items ordered are delivered to the terminal where the consumer is located.

Abstract: Visual features of a vehicle, a vehicle's orientation with respect to a transaction terminal, and an individual associated with the vehicle are derived from video captured of the terminal and a surrounding area of the terminal. Additional information associated with brute force attacks in a location associated with the terminal are obtained as additional features. The visual features and the additional features are provided as input to a trained Machine-Learning Model (MLM) and the MLM provides as output a confidence value representing a prediction as to whether the current vehicle and the current individual are potentially about to engage in a brute force attack on the terminal. When the confidence value exceeds a threshold value, enhanced feature detection is enabled on the video, external systems are notified, and an incident packet of information is assembled.

Abstract: Payment card slot tampering at a transaction terminal is detected in real time. One or more images from one or more cameras are analyzed. The camera(s) is/are focused overhead of the terminal on an area adjacent to and in front of a card slot for a card reader of the terminal. The images are analyzed for determining whether hands of arms for an operator of the terminal are performing actions that indicate the operator may be attempting to install a card skimmer/shimmer into the card slot and/or attempting to remove an existing skimmer/shimmer from the card slot. When card slot tampering event is detected at the terminal a variety of automated actions are processed, such as shutting down the card reader, shutting down the terminal, confiscating any payment card inserted into the card slot, notifying a financial institution, and/or notifying legal authorities.

Abstract: An item return transaction for an item is identified at a transaction terminal. The item is authenticated using multiple factors at least some of which are independent of item code identification for the item. A fraud score is calculated based on the multiple factors, item code identification, and data that is specific to a customer associated with the transaction, specific to the item, specific to a store associated with the transaction terminal, and specific to a retailer associated with the store. The fraud score and a customer-return grade for the customer are processed to determine whether the transaction can complete at the terminal without assistance or whether the transaction is to be held in abeyance for an audit (onsite audit or remote network-based audit). In an embodiment, the t terminal is a Self-Service Terminal (SST) and the transaction is a self-item return transaction conducted by a customer at the terminal.

Abstract: Payment card slot tampering at a transaction terminal is detected in real time. One or more images from one or more cameras are analyzed. The camera(s) is/are focused overhead of the terminal on an area adjacent to and in front of a card slot for a card reader of the terminal. The images are analyzed for determining whether hands of arms for an operator of the terminal are performing actions that indicate the operator may be attempting to install a card skimmer/shimmer into the card slot and/or attempting to remove an existing skimmer/shimmer from the card slot. When card slot tampering event is detected at the terminal a variety of automated actions are processed, such as shutting down the card reader, shutting down the terminal, confiscating any payment card inserted into the card slot, notifying a financial institution, and/or notifying legal authorities.

Abstract: Visual features of a vehicle, a vehicle's orientation with respect to a transaction terminal, and an individual associated with the vehicle are derived from video captured of the terminal and a surrounding area of the terminal. Additional information associated with brute force attacks in a location associated with the terminal are obtained as additional features. The visual features and the additional features are provided as input to a trained Machine-Learning Model (MLM) and the MLM provides as output a confidence value representing a prediction as to whether the current vehicle and the current individual are potentially about to engage in a brute force attack on the terminal. When the confidence value exceeds a threshold value, enhanced feature detection is enabled on the video, external systems are notified, and an incident packet of information is assembled.

Abstract: Cash trapping at an Automated Teller Machine (ATM) is detected in real time. One or more images from one or more cameras are analyzed. The camera(s) is/are focused overhead of the ATM or on a cash slot of a dispense module for the ATM. The images are analyzed for determining one or more of whether the dispenser module is authentic, whether the cash slot opened or did not open when it should have opened for a cash withdraw, whether visual features of the dispense module have changed over a configurable period of time, and whether hands, gestures, and actions of a person present at the ATM indicate that cash trapping is taking place at the ATM. When cash trapping is detected at the ATM a variety of automated actions are processed, such as shutting down the dispense module, shutting down the ATM, notifying a financial institution, and/or notifying legal authorities.

Abstract: A natural-language voice chatbot engages a consumer in a voice dialogue. The chatbot is customized for engaging the specific consumer based on features and characteristics of that specific consumer’s speech and a lexicon associated with terms, words, and commands for item ordering. The consumer can perform voice queries for specific items and/or specific establishments for placing a pre-staged order with the chatbot. Once the consumer selects options with a specific establishment, a pre-staged order is provided to the corresponding establishment on behalf of the user. Location data for a consumer-operated device is monitored and when it is determined that the consumer will arrive at the establishment within a time period required by the establishment to prepare the pre-staged order, a message is sent to the establishment to begin preparing the pre-staged order.

Abstract: A natural-language voice chatbot engages a consumer in a voice dialogue. The chatbot is customized for engaging the specific consumer based on features and characteristics of that specific consumer's speech and a lexicon associated with terms, words, and commands for item ordering. The consumer can perform voice queries for specific items and/or specific establishments for placing a pre-staged order with the chatbot. Once the consumer selects options with a specific establishment, a pre-staged order is provided to the corresponding establishment on behalf of the user. Location data for a consumer-operated device is monitored and when it is determined that the consumer will arrive at the establishment within a time period required by the establishment to prepare the pre-staged order, a message is sent to the establishment to begin preparing the pre-staged order.

Abstract: Cash trapping at an Automated Teller Machine (ATM) is detected in real time. One or more images from one or more cameras are analyzed. The camera(s) is/are focused overhead of the ATM or on a cash slot of a dispense module for the ATM. The images are analyzed for determining one or more of whether the dispenser module is authentic, whether the cash slot opened or did not open when it should have opened for a cash withdraw, whether visual features of the dispense module have changed over a configurable period of time, and whether hands, gestures, and actions of a person present at the ATM indicate that cash trapping is taking place at the ATM. When cash trapping is detected at the ATM a variety of automated actions are processed, such as shutting down the dispense module, shutting down the ATM, notifying a financial institution, and/or notifying legal authorities.

Abstract: A mapping of a physical environment and objects within the physical environment are generated. The objects are tracked within the physical environment and any transaction data associated with transaction correlated with the objects. The objects, relationships between each object, and any transaction data are evaluated to assign a risk score or an assessment needs score. Each score is mapped to one or more unique visual features and a heat map is generated that depicts the objects within the physical environment along with their visual features. The heat map is continuously updated in real time as scores change for the objects and the visual features change. The heat map is rendered within an interface of a device operated or worn by an attendant to assist the attendant in quickly ascertaining any risk of loss, any risk of threat, and/or any customer in need of assistance within the physical environment.

Abstract: Image analysis is performed on a user at a transaction terminal. Based on behaviors, expressions, and activities of the user, fraud or potential fraud is flagged. When fraud is flagged, the transaction terminal stops processing an active transaction on behalf of the user and alerts are sent. When potential fraud is flagged, images/video associated with the active transaction are sent for review and the active transaction may be suspended or permitted to proceed at the transaction terminal. In an embodiment, a same user conducting multiple transactions with different accounts at a same transaction terminal or multiple different transaction terminals within a configured period of time is automatically identified as fraud based on a fraud rule.

Abstract: A monitoring project is defined interactively through an interface. Resources needed for the project are obtained, configured, arranged, and verified within a target area. Video and audio feeds are captured and analyzed during a project period within the target area and customer interactions and intentions are derived from detected customer behaviors within the feeds. The intentions and behaviors are indexed with aggregated metrics within a data store. The interface provides custom queries, reports, interactive graphics, and real-time notifications from the data store that depict the custom aggregated metrics for the intentions and behaviors of the monitoring project.

Abstract: An item return for an item is identified and a workflow is determined based on first factors defined in customized rules. Resources associated with the workflow may be staged on devices associated with the item return and the workflow is initiated. When the item is detected as being present at a drop-off location for the item, item and customer information are gathered in accordance with the customized rules associated with the workflow. Security checks are processed using second factors defined in the customized rules. A decision is made based on the security checks as to whether the item return can continue processing with the workflow or as to whether an item return exception is to be raised for evaluation and/or inspection of the item and item return information collected before the item return exception was raised.

Abstract: A monitoring project is defined interactively through an interface. Resources needed for the project are obtained, configured, arranged, and verified within a target area. Video and audio feeds are captured and analyzed during a project period within the target area and customer interactions and intentions are derived from detected customer behaviors within the feeds. The intentions and behaviors are indexed with aggregated metrics within a data store. The interface provides custom queries, reports, interactive graphics, and real-time notifications from the data store that depict the custom aggregated metrics for the intentions and behaviors of the monitoring project.

Abstract: Image analysis is performed on a user at a transaction terminal. Based on behaviors, expressions, and activities of the user, fraud or potential fraud is flagged. When fraud is flagged, the transaction terminal stops processing an active transaction on behalf of the user and alerts are sent. When potential fraud is flagged, images/video associated with the active transaction are sent for review and the active transaction may be suspended or permitted to proceed at the transaction terminal. In an embodiment, a same user conducting multiple transactions with different accounts at a same transaction terminal or multiple different transaction terminals within a configured period of time is automatically identified as fraud based on a fraud rule.