In order to use InfoPath to fill in a form, a designer must develop an InfoPath template first. According to Jean Paoli, one of its developers, a key architectural design decision was "to adhere to the XML paradigm of separating the data in a document from the formatting."[2] A patent filed in 2000 by Adriana Neagu and Jean Paoli describes the technology as "authoring XML using DHTML views and XSLT."[3]

All the data stored in InfoPath forms are stored in an XML format, which is referred to as the "data source". The form template must have one primary data source for submitting data and can have multiple secondary data sources for retrieving data into the form. Secondary data sources can be built into the form or they can be accessed through an external data connection to SharePoint or a Web service.

InfoPath provides several controls (e.g. textbox, radio Button, checkbox) to present data in the data source to end-users. For data tables and secondary data sources, "Repeating Table" and other repeating controls are introduced. Template parts and ActiveX controls can also be added as custom controls in the designer.

For each of these controls, actions (called "rules") can be bound in. Rules come in three types: formatting rules such as hiding or coloring a control, validation rules (e.g. allow only a nine-digit number), and action rules such as setting a field's value based on other fields. Rules can be triggered either by a user action such as clicking a button or by the evaluation of various conditions such as field values. For example, a conditional rule could be: "Set field 'Total' to 100 when field 'field1' is not blank".