One of the biggest rules I try to follow when adding anything to my smart home is simple:
“What problem am I trying to solve?”
That rule actually came from the problem that first introduced me to Aqara.
Years ago, while potty training our oldest, he wanted to use the bathroom on his own. The problem was that he couldn’t reach the light switch and was afraid of the dark. A step stool wouldn’t safely fit behind the door, and, as kids do, he usually waited until the absolute last possible moment to make a run for it.
I needed the light to come on without him having to do anything.
My solution was a motion sensor, an automation, and a smart switch — and that search for a solution led me to Aqara.
Years later, that same automation is still running. When someone walks into the bathroom, the light comes on. Nobody thinks about it. And on the rare occasion the automation fails? You hit the switch. The bathroom still works like any other bathroom.
My oldest is 10 now and doesn’t exactly need the help anymore, so I’ve suggested removing the automation.
His response: “Why?” 
That might be the best evidence I have that it works. To him, it isn’t smart-home technology. It’s just how the bathroom works.
Over the years, that little automation has become the model for how I think about the rest of my smart home.
For me, it comes down to three questions:
What problem am I trying to solve?
Smart-home technology is fun, and I certainly have things in my house that exist just because I thought they were cool. But when I’m building an automation or changing how something in the house works, I want to know what I’m actually making better.
Is the solution accessible?
A good smart home should make the house easier to use, not require everyone to learn how the smart home works.
If someone walks into a room and wants the lights on, the action should be intuitive. Maybe the normal switch still works. Maybe a sensor handles it automatically. Maybe there’s a well-placed physical button.
Whatever the solution is, extended family, guests, kids, or babysitters shouldn’t need to read My Smart Home for Dummies to figure out which app to open or what voice command to use.
What happens when it fails?
Fails? What do I mean, fails? Certainly my smart home NEVER has an issue, right?
Networks go down. Automations occasionally don’t trigger. Hubs reboot. Kids unplug things to charge tablets… again…
Local control and protocols like Zigbee, Thread, and Matter can reduce unnecessary dependencies, but every technology eventually has a bad day.
So I try to design things so that failure is an inconvenience, not a broken house.
Automations, sensors, and voice assistants should add ways to interact with the home, not unnecessarily take familiar ones away. Presence detection can turn on a light, but the switch can still work. Voice can add another control, but nobody should have to remember the magic phrase. Sometimes a simple wireless button is the best interface of all.
The goal isn’t a smart home that never fails. That’s probably impossible.
The goal is one that fails gracefully.
Ultimately, all three questions are really trying to accomplish the same thing: make the technology disappear into the background.
If people have to learn how to use your smart home, you’ve automated the technology.
If they can forget the technology exists, you’ve automated the home.