A system aggregating data reflective of modifiable health behaviors from disparate sources into a unified format receives an accumulated health behavior data for a patient from a behavior database. The accumulated health behavior data is from a plurality of different sources in a plurality of disparate formats. The system filters the accumulated data reflective of health behaviors to select data elements that portray only modifiable clinical and non-clinical patient behaviors. It then uses these data to calculate a plurality of health behavior indices, each focused on a category of health behavior, The unified format is different from the disparate formats of the accumulated health behavior data. The system analyzes the health indices to create a single measure that reflects the aggregate of a patient's favorable and unfavorable health behaviors. The single measure and the health behavior indices from which it is derived provide the patient with access to previously unavailable information in the unified format at a single access point. Review of the health behavior data used to derive the health behavior indices allows the patient to gain insight about specific behaviors that contribute to the value of each index category and the resultant single measure. From this insight, a patient is empowered to improve their unfavorable and maintain their favorable health behaviors, which can result in increases in index category scores and in the single measure that reflects a patient's overall health behaviors.

A policy evaluation method includes, when acquiring a first policy that is a policy including a conditional branch and a node connected by a directed edge, referring to a storage that stores a plurality of second policies and an evaluation value of each of the plurality of second policies, the evaluation value indicating a number of targets to which a plurality of targets is assigned to a node by the policy, calculating a similarity measure between the first policy and each of the plurality of second policies, and outputting an evaluation value of a similar policy that is a policy including the conditional branch and the node having the similarity measure equal to or greater than a threshold among the plurality of second policies, by a processor.

A data processing system implements a system for training and fine-tuning a plan recommendation model that recommends one or more insurance plans for a user based the cost of the insurance plans and the needs of the user. The plan recommendation model is trained and/or fine-tuned using training data that has been labeled by a human actuary to improve the accuracy of the model. The plan recommendation model is also fine tuned based on feedback received on recommendations made by the model to further improve the performance of the model.

An information processing apparatus includes an acquisition unit that acquires condition information indicating a desired condition of the subject for insurance, and an extraction unit that extracts a related description related to the desired condition from a document describing an insurance product that is a candidate recommended to the subject, using an extraction model with machine learning in such a way as to output a portion related to data in the document using a set of the document and the data as an input. Decision making in a case where the subject selects an insurance product can be supported.

An agentic artificial intelligence system processes insurance claims, medical claims, financial transactions, and sales leads by receiving and preprocessing claimant, patient, transaction, and prospect data to standardize formats, remove sensitive identifiers, and enrich records. It uses machine learning, deep learning, natural language processing, and computer vision to analyze both structured and unstructured data, identify errors, inconsistencies, or fraudulent patterns, verify eligibility and compliance, and assign relevant codes based on historical and contextual information. The system calculates expected payouts or reimbursements, assesses transaction feasibility, and generates risk scores while adapting its predictions to market conditions, contractual factors, or clinical guidelines. A multi-agent framework coordinates specialized agents for eligibility verification, coding, pricing, fraud detection, and sales outreach, supporting multi-channel communication, lead prioritization, and natural language generation of outreach messages and decision-making explanations. Continuous learning is achieved via retraining, feedback loops, federated learning, and blockchain-based recordkeeping, ensuring secure, transparent, and compliant operations across multiple domains.