Alongside the ample promise of speech recognition products like Alexa, Cortana, and Siri, users have always been sure to encounter and be frustrated by the technology’s limitations in accuracy. But these days, each incremental improvement to speech recognition technology is making a major difference. We are approaching 90-95% accuracy today (to give some perspective, in 2010 we were at 70% accuracy) – and yet the chasm between where we are today and that 99% accuracy figure is massive. Most people do not fully appreciate the magnitude of improvement 99% accuracy represents over today’s accuracy levels.
Besides accuracy, the other big issue with voice recognition is latency. The good news is that computational improvements along with massive investments in AI are bringing hope to solving the latency problem. Ultimately, voice recognition will go from niche usage to widespread adoption once the technology is able to combine pinpoint accuracy with speedy response times.
The current iteration of voice recognition, despite its many flaws, is already growing quickly…very quickly. Just consider:
- Though users complain about accuracy, Siri currently processes over 1 billion requests per week
- Microsoft says 25% of Bing searches on Windows 10 are driven by voice
- Baidu projects that at least 50% of all searches will be through images or speech by 2020.
Consider the rapid pace of innovation. One year ago, Alexa could handle 14 Skills, whereas today Alexa can handle almost 1000 Skills. It’s being reported that Alexa will soon be able to detect a person’s emotional state – imagine it detecting irritation in a person’s voice and offering an apology! Voice recognition is becoming more than just a novelty and is moving firmly into a much more functional realm. Some analysts now believe that the Echo platform, which showcases Alexa, could become one of the fastest growing items of 2016/2017. Some are even suggesting that Microsoft’s purchase of LinkedIn is heavily weighted to the future of Cortana. This reality has left Apple scrambling, as evidenced by the recently leaked plans to launch their own speaker system that’s integrated with Siri. In addition, at WWDC 2016 Apple shared details of Siri making her way to the desktop in the new macOS.
Regardless of one’s preferred platform, voice recognition technology is now firmly on the fast track for widespread adoption. How will you leverage this platform personally? How will it help your company? Where does it fit on your product road map? What impact will it have on your customers?