With the introduction of Siri on the iPhone 4S in 2011, voice recognition, not exactly a new innovation, suddenly took center stage with consumers. But before that, and especially since then, there has been a rush of developments to meet the demand for ever better, more responsive technology across smartphones, tablets, computers, phone services and whatever else will come next. Beyond Verbal, a startup based out of Israel, claims that it has developed a way to take computer-based voice recognition one step further, by creating software that is able to detect not just the words, but the emotional nuances of a voice to decipher how a person speaking is feeling. Today, it’s announcing its first round of funding, $2.8 million led by the newly-launched Genesis Angels, to roll out its patented technology commercially.
As part of the investment, Kenges Rakishev, investor and co-founder of Genesis Angels, is joining the board.
Beyond Verbal offers its technology as a API-style cloud-based licensed service that can be integrated into bigger projects: the sky’s the limit for what these services might be, but you can imagine emotion detection used in all kinds of scenarios, from customer services to games, dating services (maybe to help people figure out if someone is really interested in them) and, yes, personal assistants like Siri.
Interestingly, while we tend to think of voice recognition as a customer-focused service, one potential application can actually be to help customer services representatives improve their own performance. “We can even tell if an agent is losing his temper. We can tell if you are losing your ‘sales perfect’ intonation,” said CEO Yuval Mor in an interview. Indeed, the premise of Beyond Verbal’s service is that a lot of the voice recognition services on the market today only give part of the story:
“It’s not just what people say but how they say it,” he says.
Beyond Verbal also provides analytics and diagnostics that helps customers track different emotions picked up through their applications. Beyond Verbal has yet to announce any customers, although the first is likely to come in the next few weeks, according to CEO Yuval Mor.
While the idea of voice recognition brings to mind companies like Nuance and True Knowledge (whose Evi we reported last month appears to have been acquired quietly by Amazon), which are able to parse natural language into actionable commands on a handset, Beyond Verbal is also touching on another trend in this area: the idea that these software-based services will become ever more human-like over time.
This is something also being developed by Affectiva, a startup spun out of the MIT Media Lab that takes facial recognition one step further by being able to detect emotions.
In the case of Beyond Verbal, Mor says that the system has been created with algorithms that can detect changes in vocal range that indicate things like anger, or anxiety, or happiness, or satisfaction, and cover nuances in mood, attitude, and decision-making characteristics. “To understand these three things is to understand emotions,” is how Mor puts it.
Given that many voice recognition systems — from Google Glass to IVR-based customer services — don’t actually work particularly well, I’m guessing that dissatisfaction may be one of the more highly calibrated of the emotions that Beyond Verbal can track.
Beyond Verbal, of course, is not the first to claim that it is able to read emotions in a voice recognition service. But it claims that it’s providing something different from what is on the market today. Nice, for example, also says that it can track emotions, but it uses speech analytics and word choice and volume to gauge levels of emotion. But without including intonation, Tor says such a solution is “very limited.”
Beyond Verbal is the first funding announcement to come out of the $100 million Genesis Angels fund, which was launched specifically to focus on early-stage investments in artificial intelligence and robotics.