Abstract: The systems and methods disclosed herein relate to a persistent memory-based architecture for avatar-based applications that maintains user-specific data across multiple interaction sessions. A briefing data that includes interaction data between a user and an artificial intelligence agent and/or user time-series data is accessed from a persistent memory. In response to receiving additional interaction data, portions of the additional interaction data are linked to the briefing data structure. A user-specific input is constructed by combining the briefing data structure, the portions of the additional interaction data, a predicted state output and/or a dataset retrieved from one or more data repositories. The user-specific output is transmitted to input layers of the artificial intelligence agent to generate a user-specific output. The briefing data structure is updated to include a representation of the user-specific output.

Abstract: Disclosed herein are systems, methods, and computer-readable media for generating and transmitting communication signals based on update signals that are detected using artificial intelligence (AI) agents. One or more AI agents can monitor a plurality of sources to detect update signals associated with available operations targeting one or more entities. In response to detecting an update signal, the AI agents can identify operations based on relevance and adherence to policy rulesets. The AI agents can generate a multimodal communication artifact that causes presentation of the operations, describes factors supporting the operations, and identifies a degree of adherence of the operations to the policy rulesets. The AI agents can automatically capture indications of the operations and adherence to the policy rulesets, allowing for provable compliance of the AI agents with the policy rulesets.

Abstract: Systems, methods, and devices for generating constrained outputs for large language models are disclosed. The system receives an input query relating to a request from a user via a user interface associated with an entity. The system retrieves knowledge elements from a proprietary knowledge corpus, including frameworks, methodologies, or constraints specific to the entity. For each knowledge element, the system inputs the query and knowledge element into a large language model to generate a preliminary output. The system applies user-specific rules to generate a constrained output by parsing the preliminary output, evaluating response components against user-specific constraints as logical predicates, and modifying components that violate rules. The system generates citations linking outputs to source identifiers, aggregates the constrained outputs and citations to generate a composite response, and transmits it to the user interface.

Abstract: Disclosed herein are systems, methods, and computer-readable media for generating and transmitting event artifacts based on events that are detected using artificial intelligence (AI) agents. One or more AI agents can monitor a plurality of sources to detect events associated with available interactions with one or more public entities. In response to detecting an occurrence of a first event, the one or more AI agents can assign a first priority metric to the first event. The first priority metric can be based on particular properties of a primary entity. In response to the first priority metric falling within a first predetermined metric range, the one or more AI agents can generate a first event artifact based on the first event and the properties of the primary entity. The first event artifact can be transmitted to the primary entity via one or more engagement channels of a multimodal interface.

Abstract: The systems and methods disclosed herein provide an architecture for deploying artificial intelligence (AI) avatars that generate audio and video responses to user interactions across various infrastructure environments. The systems and methods disclosed herein establish connections between client applications and a first infrastructure layer that invokes multiple agents trained on domain-specific knowledge bases to retrieve associated data records. The retrieved data and user input are routed to a second infrastructure layer hosting a generative AI model that identifies relevant data fields and generates avatar responses as audio or video streams. The systems and methods disclosed herein are enabled to evaluate the user inputs and/or avatar responses by applying one or more validation criteria. The systems and methods disclosed herein adjust the user inputs and/or avatar responses based on the evaluation.