3 Comments

Summary:

Now that touch input has revolutionized the mobile device market, what’s next? Voice interaction and natural language processing (NLP) surely have to be high on the list. But it’s not always easy for app developers to enable voice features. Enter OneTok, a cloud-based NLP solution.

Siri commercial

In the past five years, touchscreens have transformed the way we use mobile devices and voice input could well be the next transitional technology. Adding natural language understanding (NLU) and voice recognition isn’t something that every app developer can easily do, however. That spells opportunity for a company like OneTok, which offers a voice-interpretation platform to make it easy for devs.

OneTok — one of our Launchpad finalists at this week’s Mobilize event — launched in 2011, and made its platform available for mobile app developers on Wednesday. The company pairs native support for iOS, Android, and BlackBerry 10 with libraries for third-party developer frameworks such as Appcelerator Titanium and PhoneGap. Although developers enable voice input with OneTok at the client level, the service is actually cloud based: Audio processing the NLU analysis runs on servers with the results quickly shot back to a mobile device.

OneTok’s CEO, Ben Lilienthal noted how voice recognition solutions such as Apple’s Siri have raised awareness of NLU with consumers, but it’s is a limited solution ripe for opportunity.

“With the new Siri, users can open apps with their voices, but it stops there. Siri gets users to developers’ doorsteps, but without OneTok as the voice enabler, they can’t come in. OneTok enables them to talk and actually interact with and use their apps through simple voice commands.”

This type of solution makes sense, provided app developers can predict and enable voice phrases that users will most want to use. The service — which is subscription based — offers a helpful Top 10 package of the most likely used phrases, such as “Post to Facebook”, “Share via email or Twitter”, or “Zoom in and out”. But to enable truly useful voice interpretation to apps, solutions will have to grow with the complexity of those apps.

  1. I recall my old Nokia E61 was able to place calls based on voice recognition of contact names. Pretty old technology, now all hail to Apple’s Siri, and to OneTok – which brings us to the previous level as long as the voice recognition has to be done in the cloud.

    Share
    1. Natural language processing and voice recognition are different technologies.
      It’s one thing to be able to translate an utterance to a text string then compare it to a set of values, but discerning what the speaker actually means and being able to account for any combinations of words that could mean this is where services like Siri and Onetok are breaking new ground.

      Share
      1. Matt, this is exactly correct. OneTok does much more than simply provide a text file of the spoken audio. We interpret intent and tell the app what to do.

        Share

Comments have been disabled for this post