Sensory is a brand whose name may not be immediately recognizable, but whose tech surely is.
“There are any number of devices that, for example, if you’ve said, ‘Hello Alexa,’ you’re actually talking to Sensory technology,” Sensory’s chairman and CEO, Todd Mozer, told Karen Webster in a recent conversation, noting that Sensory’s close collaboration with Amazon “on a variety of projects” is also public knowledge.
Sensory gets around. It’s worked with some of the biggest brands in the world – mobile phone OEMs and operators, FIs, digital start-ups, toy and apparel manufacturers, to name but a few – with its three main product offerings: Truly Hands Free, which powers all kinds of voice-activated technology; Truly Natural, its natural human language embedded engine; and Truly Secure, biometrics technology that fuses facial recognition and voice recognition into a layered method of authentication.
In fact, Sensory gets around so well that sometimes when it’s pitching its services, it finds that it is competing with itself when potential clients come back with “other” firms they are considering.
“We work all types of institutions in all kinds of ways, because we have so many partners that have the ability to package our stuff and sell it,” said Mozer. “So that means sometimes they think they are comparing different solutions when they are in fact comparing Sensory to Sensory – we sometimes end up in a shootout with ourselves.”
The good news, of course, is that it is hard for Sensory to lose that draw – but the better news is what that situation says about the world today. For the last two decades plus, the company has been building technology ahead of its time – but the recent explosion of interest from both consumers and institutions in Sensory’s suite of services means that in some ways, the future is now. Earlier this fall, the company powered its partnership with LG on their smartphones, and forged a partnership with Fujitsu to provide facial biometric authentication for Japan’s Mizuho Bank’s mobile banking app.
The Biometric Puzzle
Consumers like mobile banking, as scores of data points will attest. Features like facial recognition are particularly popular, because when done right, they are fast and frictionless, as customers are already looking at their phones.
But, Mozer noted, consumers only want that convenience if security comes baked in. No one wants a situation where someone can pick up their phone, gain access and empty their accounts.
It’s a challenge, Mozer said, because the reality is that any system can be spoofed or hacked.
“I would never claim we can’t be spoofed – every technology in the world can be spoofed with enough work,” he said. “The question is, can you layer enough different approaches on top of each other that implementing a successful spoof is too hard to be worth anyone’s time?”
Good products build against spoofing – the latest update on Sensory’s deep learning anti-spoofing model was built of the literal terabytes of data it got from a free app it released called AppLock.
“What we saw when we released it was that a lot of people interested in our technology were specifically interested in downloading it to try and spoof it,” said Mozer. “That was both competitors and just regular users.”
The good news is it gave them a vast trove of data on spoofing techniques that they could then use to train their deep learning anti-spoof model, meaning it is a system that is hard to fake out.
Moreover, Mozer noted, the real secret in the sauce is creating enough layers of biometric authentication that can passively weave into a consumer’s experience, sufficiently enough to dissuade an attacker onto an easier target.
“There can be a temptation to make it harder and harder to spoof by making the person operating it do more unique actions. But it becomes intrusive, which kind of defeats the purpose of authentication. We like to layer unobtrusive biometrics on top of each other so they are working in tandem, but outside the user’s notice mostly.”
And, he noted, Sensory’s technology is built to work independently of brand or platform – all it needs is a smartphone with a camera and a speaker.
“That’s one of the reasons brands work with us,” Mozer told Webster, noting that interoperability in all of these functions is important, as the areas in which they are relevant is expanding.
The Changing World Of Automotive Voice
While Mozer had no specific comment as to whether or not Sensory is a part of Alexa’s integration with Garmin, he did note that Amazon’s notable strength has been not just its ability to ignite voice on its eCommerce home turf, but to move the game out to places like the car.
Particularly the car, Mozer noted – because as an expert in natural language processing and hands-free AI assisted tech, he said that automobiles so far have not exactly been an area where voice technology has flourished, despite the fact that consumers are enthusiastic about it. “It is consistently the most highly criticized feature in a car, because it just never works well. The reality is the app in your phone works a lot better.”
The problem thus far, Mozer explained, is that car design and product cycles have historically been so long that they just weren’t compatible with quickly evolving technology. As a result, the “voice activation” feature in a car still needs to be started with a button, often doesn’t understand the commands it is given, follows commands it isn’t given and, on the whole, creates a frustrating experience for the user.
“They are getting a lot smarter about it at the car manufacturers, and realizing that it might be more about hooking up with mobile phones and letting them take over the speech recognition side of things,” said Moser. “Because the reality is, in terms of voice recognition, they still have a long way to go.”
The ecosystem around voice, smart AI, deep learning and biometrics is old hat for Sensory – but it’s new to the public consciousness, and quickly evolving in the spotlight. That means, Mozer said, that the questions that need to be asked – and the answers that need to be sought – are also changing.
For example, consider the debate about whether security credentials are better stored on a device or up in the cloud. Sensory’s technology saves the information to the device.
“The question, though, isn’t really about the security of the cloud versus the security of the device,” Mozer pointed out. “In this specific case, the question is incentive. Cracking a cloud database is a lot of work that yields a huge trove of data, [so] there is a good incentive. Hacking a phone is a lot of work for a small amount of data.”
But, he noted, the bigger-picture views are that the future of security is going to be about layers of features – some of which live in the cloud and properly should, some of which live on devices and things like data tokens, which move between.
“It is going to be all of these things in a lot of combinations that will make it work well, and help it not intrude on the customer experience.”
Because customers want their experiences to be easier, and what voice navigation and biometric authentication provide is just that: The chance at a smoother experience that is no less secure than what they already have, even though it is much easier to operate.
“At this point, it’s just a matter of time,” Mozer noted.
And we’re inclined to believe his prediction – he did see all of this coming 20 years before almost anyone else.