Voice usability is hard. Here are some suggestions for basic improvements.
As the former content lead for voice transactions at Google, I know a bit about the limits and constraints of building voice UI. Users are skittish about privacy, especially in light of GDPR. Systems to support interactions are new, proprietary, and still evolving. There’s a ton of work to do to make these devices truly useful in our lives. At present, they feel more like toys for the tech curious, not tools we’d readily integrate into our day-to-day.
My observation is that teams are going overboard in favor of protecting privacy. As a result, usability is taking a hit. Here are a few suggestions for voice interaction improvements — it’s what I’d push for if I were in charge.
Let’s take a look at Reminders. If I set a reminder, Google won’t even read it out loud because we have multiple users for our device. It’s erring on the side of protecting me in case my reminder is something like, “Remind me to wash the sex toys.” Guess what, my reminders will NEVER be that private, as I’m sure any average person’s would never be. Instead, Google could ask when I set a reminder if it’s OK to read it out loud at the specified time. Or, even better, it could ask if it’s always OK to read my reminders out loud. Today if I set a reminder, at the specified time I get, “I have a reminder for Bobbie.” I don’t even know what to say to that other than, “OK, yeaaahhhhh, so what is it?” It’s annoying to have to ask what the reminder actually is.
If it’s a family reminder, like take out the trash cans on Tuesday nights, I don’t want Google to treat that as a reminder just for me. I want Google to just say it out loud, because the trash needs to go out to the curb and I don’t care who does it. If my kid has ballet and soccer practice on Thursdays, I want to set a reminder 30 minutes ahead so we can get ready and out the door in time. “Don’t forget, Grace has soccer practice and ballet in 30 minutes.” Thanks, Google!
I have a sweet, wonderful, great uncle who suffered a stroke and lost 75% of his vision. I knew the Google Home could help him with lots of tasks like making phone calls, ordering groceries, texting his wife or even reading things out loud for him, like recipes. So I gave him a Google Home as a gift. That gift is sitting on his counter completely unused because I couldn’t find an in-home tech support service that knew anything about setting up the device. Talk about a barrier to entry! If you have to hire tech support to get started you’re in a usability swamp. It would be so fantastic if the setup was completely guided by voice. Even if a sighted-person had to help poke a few buttons in the Google Home app, that would be better than the current difficulties around connecting devices, granting permissions, or setting up a Google Voice number to make calls via VOIP. If it were up to me, I’d make as much of this work transparent to the user as possible to increase adoption.
I’d love to be able to tell Google, “I fed the dog.” If my husband gets up after I’ve left the house, it would be awesome if he could ask, “Hey Google, did anyone feed the dog?” Google would answer, “Yes, Bobbie fed the dog at 8:23 this morning.”
And don’t even get me started on how impossible it is to create and maintain a useful shopping or to-do list. Currently, to-do lists are not supported. My husband and I maintain a shared house-related weekend list with things like:
- Cut up the cardboard for recycling
- Ask the gardener to trim the driveway hedge
- Take extra clothes to Goodwill
- Sweep out the garage
It just can’t be that hard to record a list, and then allow for items to be removed when they’re done. Lists could be named. “OK Google, are there any to-do lists?” “Yes, the House list has 4 items on it. Should I read those to you?”
IFTTT or Who-What-When commands
I’ve stopped using my Google Home for anything except trivia quizzes (“Hey Google, can guinea pigs eat avocado?”), turning my Nest thermostat up and down, and cooking timers. The simple reason is that the thing is pretty dumb. It could be soooooo much smarter if users could set up simple if this > then that commands (IFTTT) or who-what-when commands.
Here are a few examples:
- Hey Google, if the phone rings when I’m listening to music, turn the volume down to 2.
- Hey Google, take a message for Christina and read it to her when she talks to you. (Who: Christina (the dog sitter). What: Instructions for dog care. When: when she identifies herself.)
- Hey Google, if I tell you we’re going out of town, set the thermostat to 55 degrees and pause all reminders.
Like anything, these types of interactions are always harder to build than they sound at the surface. That said, I guarantee a company like Google has the brilliant designers and engineers to pull this stuff off.
One caveat: I’ve been away from Google for 6+ months — some of these solutions might be in the works? I hope they are!
Thanks for reading. 🙂