Abstract: Disclosed herein are system, method, and computer program product embodiments for generating corrective contents. For example, the method includes identifying noise content in responses of a large language model (LLM), determining a theme for corrective contents to counter the noise contents, determining a target genre distribution for the corrective contents, generating, using an artificial intelligence (AI) model, the corrective contents based on the target genre distribution and the theme, and transmitting the corrective contents via one or more channels.

Abstract: System, method, and computer program product embodiments control routing within an agentic ecosystem using agentic reinforcement learning. A first computer processor receives log entries of agentic activity from an agent activity log data store writable by a second computer processor associated with at least one of a number of AI agents of an agentic ecosystem. The first computer processor generates agent transition probabilities based on the received log entries. The agent transition probabilities influence probabilistic routing decisions of AI agents of the agentic ecosystem. The first computer processor transmits the agent transition probabilities to a plurality of the AI agents of the agentic ecosystem. These steps can be repeated iteratively to improve learning.

Abstract: Disclosed herein are system, method, and computer program product embodiments for generating and customizing micro agents. An embodiment comprises receiving, from a user device, a request to customize template agents. The embodiment further comprises selecting a set of template agents, receiving a selection of at least at template agent, personalizing the a least the template agent to obtain a micro agent, receiving from the user device a request to complete one or more tasks associated with a service provider system, transmitting using the micro agent, a command to the service provider system to perform the one or more tasks, and transmitting, to the user device, a notification indicating that the one or more tasks are completed.

Abstract: Disclosed herein are system, method, and computer program product aspects for resolving a transaction claim associated with a merchant system. An example implementation receives, from a user device, a transaction claim associated with a merchant system. The implementation performs an information search within a database storing disclosure information between the user device and the merchant system. The implementation retrieves, from the database, specific disclosure information associated with the transaction claim. The implementation generates a prompt that directs an LLM to generate one or more actions associated with the transaction claim. The implementation determines whether at least one of the one or more actions applies to the user device. When the at least one of the one or more actions applies to the user device, the implementation determines whether it applies to the merchant system. The implementation resolves the transaction claim based on the at least one of the one or more actions.

Abstract: Various embodiments are described relating to systems and methods of managing failed contactless transactions. In one examples, a system comprises a payment device that is configured to initiate a transaction with a first wireless-enabled point-of-sale (POS) device and generate a payment credential for the transaction. The payment credential is transmitted to the wireless-enabled POS device. A transaction error is generated based at least in part on an error code received from the wireless-enabled POS device or an expiration of a time out period. The transaction error comprises an identifier for the wireless-enabled POS device and store the transaction error in the memory.

Abstract: Disclosed herein are system, method, and computer program product embodiments for increasing computer, network, and data security via user generated tokens including zero-knowledge proofs within an agentic architecture. A client device may utilize an autonomous agent to communicate with an autonomous agent at a merchant device to identify whether the merchant device is selling a desired product. A client device may generate a token including the product, a merchant identifier, purchase amount, timestamp, and zero-knowledge proof. The client device may sign the token using an encryption key. The agent at the client device may transmit the signed token to the agent at the merchant device identified via the merchant identifier. The client device may receive a result indicating whether the token and zero-knowledge proof were verified by the merchant device.

Abstract: Disclosed are various embodiments for implementing a interaction counter to enable multi-device transactions. A computing device can receive an input from a user interface, the input indicating a number of interactions associated with a transaction session. The computing device can generate an interaction counter for the transaction session based at least in part on the number of interactions. Next, the computing device can send a request for a first set of data corresponding to a first interaction and receive the first set of data in response to the request. Then the computing device can increment the interaction counter based at least in part on the first data for the first interaction.

Abstract: System, method, and computer program product embodiments control routing of user data within an agentic ecosystem. User preference data, including routing information indicative of at least one preferred, un-preferred, or prohibited route of data flow within the agentic ecosystem, is received and a token is generated based on the user preference data. The token includes the routing information, and is cryptographically signed. An artificial intelligence (AI) agent of the agentic ecosystem is selected based on the routing information and a request message that includes the token and a payload describing a request for a task to be completed by the AI agent is transmitted to the AI agent. A reply message containing a confirmation that the AI agent successfully completed the task or an error message describing a reason why the AI agent did not successfully complete the task is received from the AI agent.

Abstract: Various systems and methods are described for generating indicators for facilitating near field communication (NFC) interactions. In some examples, a system is configured to identify intent to initiate a transaction with a point-of-sale (POS) device and initiate an image stream using a camera to capture a view of the POS device. A POS device type in the image stream is identified based at least in part on a visual object detected in the image stream. An antenna location for the POS device is determined based at least in part on the POS device type and an inventory of POS terminals. The system is configured to augment the image stream to include a visual terminal indicator for a terminal antenna field based at least in part on the antenna location of the POS device.

Abstract: Various approaches for an automated selection of a payment instrument based at least part on a token linked to other accounts. In one example, a system, comprising is configured to identify a token payload for authorization of a transaction in which the token payload comprising a token of a user profile. Account identifiers are retrieved based least in part on the user profile. Account identifiers have individual rules that can be applied to the transaction data. A risk assessment score associated with the transaction is determined. The system is configured to select an account identifier among the account identifiers for the transaction based least in part on the risk assessment score and a rule applied to the transaction data. The system is further configured to determine whether to authorize the transaction using the account identifier selected for the transaction.

Abstract: Disclosed are various embodiments for recommending digital credentials for a user to present at a terminal based at least in part on an informed hypothesis of the context surrounding the terminal and user. A user can store multiple types of credentials (e.g., payment account data, coupons, tickets, identification, etc.) in a digital wallet. A trained recommendation model can be executed to make an informed hypothesis of what context may be sought by a terminal and what credential should be presented at the terminal by analyzing data associated with the given context. A credential payload associated with the recommended credential can be generated and preloaded by the wallet so that the user can present the recommended credential to the terminal without having to select which credential to present.

Abstract: Disclosed are various embodiments for a single tap payment transaction. First, a user of a computing device can be authenticated. Then, in response to authentication of the user, the user can be prompted to select a payment credential. Next, the payment credential and a digital identity document associated with the user can be provided to a payment terminal in data communication with the computing device through an NFC radio. Subsequently, a receipt for a transaction can be received from the payment terminal, the receipt indicating that payment for the transaction was made with the payment credential and that an identity of the user was verified with the digital identity document.

Abstract: Disclosed are various embodiments for automatic enrollment in a merchant rewards program. A point-of-sale (POS) terminal can establish a near field communication (NFC) session with a client device to communicate with the NFC enabled client device. The PoS terminal can then receive a payment instruction for a transaction from the client device, the payment instruction comprising a unique user identifier for a user of the client device. Next, the PoS terminal can determine, based at least in part on the unique user identifier, that a user of the client device is not enrolled in a merchant rewards program. Subsequently, the PoS terminal can send to the client device a request to enroll in the merchant rewards program. Then, the PoS terminal can receive, from the client device, an enrollment authorization. Finally, the PoS terminal can enroll a user of the client device in the merchant rewards program.

Abstract: Disclosed are various embodiments for an NFC-based self-checkout system. Embodiments can include a computing device comprising a near-field communication (NFC) reader. The computing device can obtain, via the NFC reader, a stock keeping unit (SKU) identifier for an item from an item NFC tag attached to the item. It can also obtain, via the NFC reader, a shopping cart identifier and a terminal identifier from a cart NFC tag attached to the shopping cart. It can then receive a user input indicating that a purchase of the item is to be completed. Subsequently, the computing device can establish a secure connection with a payment terminal service associated with the terminal identifier and send the SKU for the item, a payment instrument, and the shopping cart identifier to the payment terminal service.

Abstract: Disclosed are various embodiments for securely and efficiently transmitting various types of digital content (e.g., images, audio, video, etc.) with added assurance regarding the integrity and similarity to the original form of the content. Content to be transmitted can be encoded into a latent encoded representation of the content using a trained AI encoder, thereby reducing the size or compressing the content. A digital signature can also be applied to the encoded representation. The sender can send the original content and the signed encoded representation to a receiver. The receiver can encode the received content using the AI encoder and compare the encoded representations to determine similarities. In addition, the receiver side can validate the digital signature of the received encoded representation. The comparison of the encoded representations and the validation of the digital signature can both be used to validate the integrity of the received content.

Abstract: Disclosed are various embodiments for an encrypted context-based prompt system for interactions between multiple entities. In one embodiment, a system comprising a computing device configured to receive a request to activate a prompt session. The request comprises a template identifier representing a type of interaction between a first device and a second device. A decrypted term is determined by decrypting an encrypted term associated with the request. A large language model prompt is generated based at least in part on the template identifier and the decrypted term. A proposed term is generated by executing a large language model application based at least in part on the large language model prompt. An encrypted proposed term is transmitted to the first client device or the second client device.

Abstract: Disclosed herein are system, method, and computer program product embodiments for using generative AI to support an upsell engine. A point-of-sale (POS) system may combine transaction data, merchant data, and consumer data into a shared embedding space comprising a numerical matrix. A machine learning model in connection with the POS system may generate an offer based on the shared embedding of the transaction data, consumer data, and merchant inventory data. The offer may be unique to the transaction. The POS system may prevent merchant data from being transmitted to a banking server, and may prevent consumer data from being transmitted to a merchant server. The POS system may output the offer.

Abstract: Disclosed are various embodiments for implementing a decentralized messaging system that allows users to exchange messages via an inbox maintained through a distributed ledger. A messaging inbox can be created for a user in a distributed ledger using an inbox address that includes a string of alphanumeric numbers without an identifying domain name or identifying information about the user. Furthermore, a user may define content preferences associated with the inbox that can be used by the distributed ledger to filter content received in the inbox to minimize unwanted or undesired content provided to the user.

Abstract: Disclosed are various embodiments for long term context-based small language model aided authentication. Various embodiments of the present disclosure can receive a contextual instruction associated with a credential. The contextual instruction can represent a category of input that is accepted to obtain a user credential. Various embodiments can train a small language model to generate an authentication small language model such that the authentication small language model returns the credential in response to determining that an input is within the category of input of the contextual instruction. Various embodiments can then receive a client input and prompt the authentication small language model to make a determination regarding the client input is within the category of input of the contextual instruction.

Abstract: Disclosed are various embodiments for dynamic enforcement of large language model prompt templates and prompt template categorization. In one example, a system comprise a computing device that is configured to identify a prompt that has been submitted by a client device for a large language model (LLM) service and determine that the prompt fails to match an existing prompt template. The prompt and an unidentified prompt are determined to meet a similarity threshold based at least in part on a common prompt component shared between the prompt and the unidentified prompt. A prompt template is generated for the LLM service based at least in part on the prompt and the unidentified prompt meeting the similarity threshold, the prompt template comprising the common prompt component and a placeholder.