Siri and Alexa are intelligent agents because they use sensors, such as microphones and other inputs, to perceive a request and they draw on their collective experience and knowledge via supercomputers and data banks all over the world to make a decision. The amount of information these intelligent agents are exposed to grows and develops as they get to know more about their users and gather new data as it is presented in the world.
By definition, an intelligent agent is anything that can make decisions about how to react or respond based on how it perceives its environment and its experiences. Siri is able to retrieve information about the weather, scores, or world capitals (or anything you ask) based on its perception of its environment.
It does so through the use of sensors and delivers the end result through actuators. If you compare sensors and actuators to the human world, sensors would be things like ears or eyes we use to observe or perceive the world around us. Actuators are items like our voice that we can use to respond to something we've heard, for example. With intelligent agents, sensors are things like cameras and microphones that can receive, and actuators include speakers and voice files that perform tasks such as transmit information or put other devices into action.
Think about the world of self-driving cars. These are types of intelligent agents. They make decisions about their environment (the highways and roads they travel) based on sensors such as GPS, cameras, and computer applications in order to maneuver the car through steering and braking (actuators) to help passengers arrive at their destination safely.
