Hey guys, I am working on a library that targets a hands-free internet browsing experience. Right now my demo allows you to create a customized search on Zillow or Trulia using just your voice, but could be configured to order pizza's, search for movies, basically anything that might be faster and easier to do by voice than by clicking and filling in forms.
While this is kind of a cool novelty and a lot of fun to work on, I am wondering if this could be something a bit bigger. Specifically, could this help people with motor skill issues (like Parkinson's disease) to easily navigate the web ? Or just maybe an easier way to interact with mobile websites.
Here is the video demo:
I just built this and haven't even tested on a mobile device, so I'm not quite ready to post the actual web demo.
Thanks for any feedback you may have.
Your question makes it sound like you want us to validate your idea. It's not a very effective audience here to do that and not a very controlled way of asking. This self-selected audience tends to think more liberally about what's possible or worthwhile.
That said, my initial questions would be whether you are hoping to package this as a software library that apps could use to voice enable more complex functions, or are you making it a user-oriented product that they should be able to use with any complex search?
I think you mean to do the first.
In my limited experience, there are already a handful of digital assistants that offer a software library to voice enable many things (like Alexa, Google Assistant, Siri, etc.), and with a lot more common usage established than a new overlay voice enablement tool. Are developers going to choose a new tool or are they going to use one of the established digital assistants to voice enable their app?
There's a captive audience of people who own digital assistants, and yet a MINISCULE number of apps are choosing to enable voice activation. Ask yourself why.
Motor skill deficient users are a very small market. There are already accessibility tools specifically designed for those situations. Whether they're as capable as desired is a separate question.
I will be up-front that I am not a technology lover. And maybe I'm in the minority, but I ALWAYS give preference to using a PC over a mobile phone for internet activities. I am thoroughly unsatisfied with speech to text accuracy. I use voice functions when I absolutely must be hands-free (like hey, Google, gas station along route). In general I have no desire to have other people overhearing what I'm saying to my smartphone. And I find it incredibly annoying when other people do it.
Bad UI is not necessarily justification to enable voice commands. A mobile device is always considered a compromise. "Because we can" doesn't usually justify a business. There has to be real value.
Even so, no one was asking for automobiles when Ford created the first one. They thought "faster horse" or "leave earlier." The issue I see here is that you're assuming mobile apps or web sites will want to choose voice over visual UI to improve UX. My gut says that if they have stinky UI, they're not even thinking about how voice could improve UX, they just suck at design and aren't going to improve, period.
As for Trulia, I refuse to house hunt on my smartphone. If I'm not sitting at my desktop, I won't do the activity. There is nothing about my phone that makes that process appealing. I want way more than it will ever give. Surely there are other people who use their phone so much they'll probably end up with hand cancer and don't even own a desktop (except maybe at work). They may feel differently than I about what's good to do on a mobile device, but they may also blame their wallet for not buying a bigger phone than to blame a mobile web design and demand change from the designers.
I liken this to hearing some millenial tell his friend, "Gee, I wish there was a way to say more without having to type it all out." He was holding a dang phone in his hand, but it didn't even occur that a phone call would save all that typing. The name of the device is a mobile PHONE, not a mobile typewriter. We see what we want to see.
Joe - One factor will be the degree to which any speech impediments affect the speech recognition capability. Also, consideration for some kind of history retention that permits clean exit from an error state would be desirable. My experience is in voice biometrics and speech recognition, but not in that market, so I can't comment on that aspect, although I suspect the opportunity is much larger than web browsing.
Thanks Paul, for your response and your thoughtful answer. I would like to make a couple of clarifications:
1. Your first assumption is partly correct: I am looking for validation -- or invalidation. I'd hate to waste time on a bad idea.
2. "or are you making it a user-oriented product that they should be able to use with any complex search?". I'm not sure. I can think of many areas where it would easier to use voice over protracted interactions with a mobile app. Most digital assistants are quite good at getting you to a place, but then once you are there, they are no longer involved in the way you interface with a website.
3. " I ALWAYS give preference to using a PC over a mobile phone for internet activities" - so do I! But we are in the minority here.
The reason I prefer a PC over mobile is the very nature of mobile UI and the small viewports make it difficult to interact, I am constantly clicking on things by accident, and setting up options is time-consuming and difficult.
I have a pretty robust library here, I'm just not sure what to do with it. That is the main reason for the question. Could this library pivot into something...?
Great demo Joe. I agree with many of Paul's great thoughts. That said, the demo seems like a 'tip of the iceberg' type thing. NLP + tables suggests many cool use cases. I tend to think about events, travel, music - apps I have to use mobile. If you can extend the tech to work both directions (i.e. a table is used to generates the language or vice versa) then things get really interesting. I'm just not sure how specific or extensible your tech is compared to what's already out there - you'd know better.
I have an interesting way this could possibly integrated into a commercial app for the web. I’m on travel today, but tomorrow evening or Saturday afternoon, we could talk about it it briefly and maybe put our heads together. Ken 404-314-9559. After 6pm CST or Sat after 4 pm
Jim, thank you very much for your input.
Matt, thanks for your comment and your input (surprisingly, one cannot reply directly to comments on here). I would love to hear a bit more about a reverse table...maybe an example of how that would look? Seems really interesting and could possibly be a great inroad into machine learning...
Hi, Ken, let's connect on here and work out a time to talk. thanks!