Systems and methods are described for efficiently meeting the needs of property buyers, sellers, and agents. Attributes of the user and of each property are combined and processed by a machine learning model. The output represents a prediction that a user will choose to view the property. Each property is ranked and displayed to the user. When the user selects a property, the interaction is used to update the machine learning model. In some cases, a natural language processing system identifies property and user attributes based on a natural language query. In some cases, a chatbot may respond to a natural language query in real time. The system may also be utilized to provide matching services for buyers or sellers to agents, or third party service providers, as well as matching buyers and sellers directly to each other, or user to other users for transaction, interaction and other needs.