The company behind Viv, a powerful form of AI built by Siri’s creators which is able to learn from the world to improve upon its capabilities, has just closed on $12.5 million in Series B funding. Multiple sources close to the matter confirm the round, which was oversubscribed and values the company at north of nine figures.
While Iconiq may not be a household name, a Forbes investigation into its client list revealed people like Facebook’s Mark Zuckerberg, Dustin Moskovitz and Sheryl Sandberg, Twitter’s Jack Dorsey, LinkedIn’s Reid Hoffman and other big names were on its roster.
In addition to Iconiq, Li Ka-shing’s Horizons Ventures and Pritzker Group VC also participated along with several private individuals. This new round follows the company’s $10 million Series A from Horizons, bringing the total funding to date to $22.5 million.
Viv Labs declined to comment on the investment.
We understand that Viv Labs was not in need of new capital, but was rather attracted to the possibilities that working with Iconiq Capital provided. It was a round that was more “opportunistic” in nature, and was executed to accelerate the vision for the Viv product, which is meant to not only continue Siri’s original vision, but to actually surpass it in a number of areas.
Viv’s co-founders, Dag Kittlaus, Adam Cheyer and Chris Brigham, had previously envisioned Siri as an AI interface that would become the gateway to the entire Internet, parsing and understanding people’s queries which were spoken using natural language.
When Siri first launched its product, it supported 45 services, but ultimately the team wanted to expand it with the help of third parties to access the tens of thousands of APIs available on the Internet today.
That didn’t come to pass, because Apple ended up acquiring Siri instead for $200 million back in 2010. The AI revolution the team once sought was left unfinished, and Siri became a device-centric product – one that largely connects users to Apple’s services and other iOS features. Siri can only do what it’s been programmed to do, and when it doesn’t know an answer, it kicks you out to the web.
Of course, Apple should be credited for seeing the opportunity to bring an AI system like Siri to the masses, by packaging it up and marketing it so people could understand its value. Siri investor Gary Morgenthaler, a partner at Morgenthaler Ventures, who also invested personally in Viv Labs’ new round, agrees.
“Now 500 million people globally have access to Siri,” he says. “More than 200 million people use it monthly, and more than 100 million people use it every day. By my count, that’s the fastest uptake of any technology in history – faster than DVD, faster than smartphones – it’s just amazing,” Morgenthaler adds.
But Siri today is limited. While she’s able to perform simpler tasks, like checking your calendar or interacting with apps like OpenTable, she struggles to piece information together. She can’t answer questions that she hasn’t already been programmed to understand.
Viv is different. It can parse natural language and complex queries, linking different third-party sources of information together in order to answer the query at hand. And it does so quickly, and in a way that will make it an ideal user interface for the coming Internet of Things — that is, the networked, everyday objects that we’ll interact with using voice commands.
A Wired article about Viv and its creators described the system as one that will be “taught by the world, know more than it was taught and it will learn something new everyday.”
Morgenthaler, who says he’s seen Viv in action, calls it “impressive.”
“It does what it claims to do,” he says. The part that still needs to be put into action, however, is the most crucial: Viv needs to be programmed by the world in order to really come to life.
While to some extent, Viv is the next iteration of Siri in terms of this vision of connecting people to a world of knowledge that’s accessed via voice commands, in many ways it’s very different. It’s potentially much more powerful than other intelligent assistants accessed by voice, including not only Sir, but also Google Now, Microsoft’s Cortana or Amazon’s Alexa.
Unlike Siri, the system is not static. Viv will have memory.
“It will understand its users in the aggregate, with respect to their language, their behavior, and their intent,” explains Morgenthaler. But it will also understand you and your own behavior and preferences, he says. “It will adjust its weighting and probabilities so it gets things right more often. So it will learn from its experiences in that regard,” he says.
In Wired’s profile, Viv was described as being valuable to the service economy, ordering an Uber for you because you told the system “I’m drunk,” for example, or making all the arrangements for your Match.com date including the car, the reservations and even flowers.
Another option could be booking flights for business travelers, who speak multi-part queries like “I want a short flight to San Francisco with a return three days later via Dallas.” Viv would show you your options and you’d tell it to book the ticket – which it would proceed to do for you, already knowing things like your seat and meal preferences as well as your frequent flyer number.
Also unlike Siri today, Viv will be open to third-party developers. And it will be significantly easier for developers to add new functionality to Viv, as compared to Siri in the past. This openness will allow Viv to add new domains of knowledge to its “global brain” more quickly.
Having learned from their experiences with Apple, the Viv Labs team is not looking to sell its AI to a single company but instead is pursuing a business model where Viv will be made available to anyone with the goal of becoming a ubiquitous technology. In the future, if the team succeeds, a Viv icon may be found on Internet-connected devices, informing you of the device’s AI capabilities.
For that reason, the investment by Iconiq makes sense, given its clients run some of the largest Internet companies today.
We understand that Viv will launch a beta of its software sometime this year, which will be the first step towards having it “programmed by the world.”
Morgenthaler says there’s no question that the team can deliver – after all, they took Siri from the whiteboard to a “world-changing technology” in just 28 months, he notes. The questions instead for Viv Labs are around scalability and its ability to bring in developers. It needs to deliver on all these big promises to users, and generate sufficient interest from the wider developer community. It also needs to find a distribution path and partners who will help bring it to market — again, things that Iconiq can help with.
But Viv Labs is not alone in pursing its goal. Google bought AI startup DeepMind for over half a billion, has since gone on to aqui-hire more AI teams and, as Wired noted, has also hired AI legends Geoffrey Hinton and Ray Kurzweil to join its company.
Viv may not deliver on its full vision right out of the gate, but its core engine has been built at this point and it works. Plus, the timing for AI’s next step feels right.
“The idea of embedding a microphone and Internet access is plummeting in price,” says Morgenthaler. “If access to global intelligence and the ability to recognize you, recognize your speech, understand what you said, and provide you services in an authenticated way – if that is available, that’s really transformative.”