The ability of the machines to accomplish tasks typical of human intelligence is broadly defined artificial intelligence. In this paper we concentrate on algorithms that are an AI subset. Machine learning algorithms are designed to learn relationships from data instead of researchers assuming explicit functional relationships between variables. Such established relationships can be used in different ways to improve data analysis based on claims. In this segment, some examples are given. The AI's ability to analyze broad multidimensional data with many variables is a considerable advantage over traditional statistical analyses. A statistician can be easily overwhelmed by many variables that can be used. By comparison, AI algorithms also detect which variables are important for the task and detect the optimal combination of such variables. The knowledge and experience of a wide range of medical professionals can be pooled together to identify complex patterns of certain illness, by the integration of claims data with other data sources such as patients ' test tests and EMRs. Such trends can be used to enhance early detection and diagnosis of underdiagnosed or rare diseases, to allow more precise assessments or to make tailored preventive care more proactive. In the case of underdiagnotic disorders, an example of a medical report data can be analyzed by physicians to decide which of the patients should have been treated and at what time (combined with laboratory values, consultant visits, and physician notes). This process can be used to train algorithms to detect other flagged instances and to identify early markers and indicators for the onset of future diseases. In recent research, for example, AI has been used to predict type 2 diabetes mellitus with claims. A large number of applications have been developed to help diagnosis, by predicting the onset of the diseases or by detecting drug-resistants disease strains (e.g., drug resistant tubercles) using AI with other types of data, such as a CT scan or Genome sequence in precision medicine. Information from hospital claims offers comprehensive information on patient outcomes, such as medications, vendors used and paid amounts and prescriptions performed with the healthcare system. The codes for billing are issued by health professionals such as the doctors and hospitals by business and public health programs for payment. Since the medical claims are routinely obtained for administrative purposes, for a substantial number of patients, health claims provide a relatively cheap source of information over long periods. However, given that our payment systems evolve and patients question their providers more, there will be a rapid increase in the amount of available claims data. The large sample dimensions can help to investigate rare conditions and the longitudinal dimension of the data enhances the observance and effects of therapy over time. Health benefits, however, usually provide only minimal clinical frequency, health status and other interest variables. This can be improved by adding claim data through ties with other data sources such as census data for local incomes or EMRs to collect more clinical data. Such links open the door to a number of research opportunities as related healthcare information can be available in a variety of unstructured formats including text, photographs, audio or video files