All smart thermostats share a common goal: save money while staying comfortable. An Internet connection allows them to adjust the indoor temperature based on outdoor conditions–some even factor in humidity. But the best way for a home to adapt to its residents is still up for debate. Your schedule may dictate which method is best for you.

If you have a regular routine …

The Nest Learning Thermostat ($249) uses algorithms to learn from your patterns. The device monitors when you change the temperature and builds a schedule. Eventually, Nest takes over, so when you wake up or arrive home, the house will already be at your ideal climate. Recent studies have found that using Nest can save up to $145 annually.

If you come and go a lot …

The Honeywell Lyric Thermostat ($249) shifts its energy-saving mode based on whether there are people in the house. Family members give Lyric access to their whereabouts via a smartphone app. When they pass a geofence, a preset distance that can range from 500 feet to seven miles, the system toggles on or off. This could cut annual costs by up to $220.

In The Works: Artificial Intelligence

Machine-learning algorithms in products like the Nest thermostat are only the beginning of a self-automating home. Soon, apps might recommend tasks based on patterns–“I see you turn on the lights at the same time every day, shall I schedule this task?” But the ultimate goal is autonomous artificial intelligence, where you’d buy a product and your home would figure out the rest. For now, SmartThings founder Alex Hawkinson says, “We’re probably years from saying ‘I’m gonna go live my life and let the tech figure itself out.’ ”

This article was originally published in the May 2015 issue of Popular Science as part of our Smart Homes feature. Click here to read more about how to make your home intelligent.