Is Voice Recognition Ready for Mass Markets?
I’m probably the last person anyone would have thought to buy and use an Amazon Echo or Google Home. I’m pretty disinterested in having the latest phone or what’s new in the world of electronics (ironic, I know). I still use disposable cameras, and it took me three years at Mindtribe to try out an iPad. (Spoiler: I ended up carrying paper books with me and using it only to watch movies when I was traveling.)
I am, however, really interested in efficiency and streamlining. I installed a Nest about a year ago and have to say; it’s been empowering from both an “I can install this and not die in the process” perspective, and a “wow this technology actually helps me” perspective. Since then, there have been small but notable situations I’ve found myself in where I think to myself, “there should be a better way to do this by now.”
I’m also a researcher and list-maker. By now, I’ve read handfuls of reviews and comparisons about home-automation products from smart locks to smart lights. I have lists written down with features I want from each product and which products work for my needs. I’ve asked my co-workers, I’ve played with the products we have scattered around the office, and I pulled the trigger… I bought an Amazon Echo Dot and some Philips Hue Smart lights to add to my ever growing smart house.
I’ve gone head-first into finding out what voice-recognition and home-automation could do for me.
I’m a trained baker and amateur cook. I spend a lot of my time in the kitchen, so a voice-controlled functionality is important for me because:
- I’m lazy
- I have stuff on or in my hands often
- I’m curious and like asking questions about information I don’t know
- Set multiple timers by voice
- Play music via Pandora
- Integrates with Nest
- Decent search ability (I wasn’t sure how I defined decent when I started this)
- Be able to name timers so I can remember timer 1 is for the dough rising on my counter and timer 2 is for the bagel bites I have in the oven
- Create lists/notes
- Integrates with Phillips Hue Lights (this might fall above the MVP line, I haven’t played with these enough yet)
- Integrates with Google Calendar to tell me what my day is like
- Integrates with other smart products TBD
- Connects with multiples of the same system around the house (potentially I’d expand to multiple rooms and it’d be cool if they all knew what was going on)
- No Bluetooth needed. I refuse to use it because my phone is a hundred years old and it never connects or always drops….that’s a blogpost for another day.
- Easy to set-up. My patience is limited.
- Easy to figure out. IFTTT (If This Then That) is awesome but again, my patience is limited, and the requests I have aren’t crazy.
- Smaller is better. Counter space is precious.
- Aesthetics are important since I’ll be looking at this thing regularly.
All of the above led me to purchase an Amazon Echo Dot. It’s pretty cheap to start with, it’s small, and the aesthetics aren’t terrible. It hits all of my requirements above the MVP line and most below it. It also came out on top in the many reviews I read.
Google Home does the same so why didn’t I purchase that? Other than price, it was because of how they respond when you ask them to make animal sounds. That’s right, when you ask Alexa what sound a dog makes she uses her voice to bark, and when you ask Google, it has recorded animal sounds. I found Alexa making animal sounds funny — there you have it.
Here’s what I found after using Alexa for about two weeks:
WHAT WENT RIGHT AND WHAT WENT WRONG
The setup was easy… to start with. I connected her to my Nest pretty quickly and found how to enable “skills” apps within Alexa to increase her functionality.
I got her to recognize I’m in San Francisco and configured Pandora. I setup the timer sounds and volumes — really everything I had for my MVP line.
I started asking her to do things. “Alexa, set a timer for 50 minutes.” She repeated back to me, “Okay, I’ve set a timer for 15 minutes.” Oh. It’s going to be like this huh? I tried again, and she set the right timer, then I canceled the other one. That was cool.
I tried moving down my user needs list by asking, “Alexa, set a timer for mac and cheese for 50 minutes”.
She didn’t understand my request.
Okay, maybe naming the timers is too advanced. Now I had to think about where my MVP line actually was.
It also sparked more thoughts about what else I’d like her to do, like, in the absence of a display, give me time warnings so I don’t constantly have to ask her how much time is left. I didn’t realize how impactful a display is.
“Alexa, what’s the weather today?” She replied, but went into much more detail than I had the patience for. She also repeated the surf warning that was in effect, every time I asked.
Maybe I can configure IFTTT to use WeatherUnderground, but I haven’t tried yet.
“Alexa, what should I do today?” She didn’t know how to answer that.
I asked her specifics, like “Alexa, what time does the Women’s March in San Francisco start?” She also didn’t know how to answer that question. I think direct questions garner better results. Things like should implies opinions, which my Alexa does not have…yet.
“Alexa, what does a dog sound like?” She didn’t know.
“Alexa, what noise does a dog make?” She still didn’t know.
“Alexa, what sound does a dog make?” It worked! She barked! Score! I was on my way to figuring out the important things in life.
This is when I realized how particular she is about how you ask for something.
It took many attempts (too many days and ways) for me to set my Nest all because I added “thermostat” to the request. I kept saying, “Alexa set my hallway thermostat to 68.” She kept pushing back. It really wasn’t until I sat down out of frustration to look up the instructions before I said, “Alexa set the hallway to 68” and it worked. From me yelling while I was in the bathtub. That was a magical moment.
The phrasing though, is the biggest issue for me. It takes work to get simple things done because you have to figure out, and remember, the specific phrasing to make things work.
I set up a couple of smart lights to see how things would work. There have been issues, but more with Philips and my internet than with Alexa.
The first light I got working was like in Sleeping Beauty when the fairy godmothers fight over which color Aurora’s dress should be. Blue! Pink! Blue! Pink! On! Off! 50%! Off! On! That was really fun.
I tried to see how much context Alexa understood by asking follow up questions to information she’d give me using “it” or “they”. From a technology perspective, it’s incredibly hard to do (so my coworkers tell me) and from my experience it never worked.
I scrolled through Alexa’s skills. And wow, it’s a huge library. There are thousands of skills to choose from, but most of them have very few with reviews. I did see some cool ones, but the phrasing used to enable the skills is, again, very specific, and can be awkward.
I tried using a bartending skill, but honestly after enabling it and failing to remember which sentence to use, I haven’t really thought about it since and certainly haven’t used it.
I realized, shortly after playing around with Alexa, that my MVP line moved, and things that were in the nice to have bucket became more important to me. I also wasn’t impressed with the search functionality; although it wasn’t on my list when I bought it, I realized it limited my interactions with Alexa, and thus how much I could rely on her.
As with most technology for early adopters, there’s too much room to fumble to make a good user experience unless you’re willing to put in the hours. It’s frustrating because she feels so close to what I’m looking for, but also so far away. With an easier integration between the supported apps and Alexa, I can imagine using it regularly.
I also thought more deeply about what I wanted to be connected, since it seems like the technology is most useful around home automation. I already have a ChromeCast, and Google Home integrates with all of my other smart devices and hits most everything for my user needs.