Watch out, Siri, Google Now, and Alexa, because there’s a new virtual assistant in town and Viv wants to eat your lunch (after probably ordering it first, too). Dubbed “The Global Brain”, Viv is actually the handiwork of some of Siri’s creators prior to the technology being acquired by Apple, and as an astonishing demonstration today indicates, they’ve not been idle in the meantime.
Dag Kittlaus, CEO of Viv, took to the stage at TechCrunch Disrupt in New York today to show off not only some of what the assistant can do, but how it’s doing it in the background – not to mention how that differs from traditional AI technologies currently in play.
For a start, there’s advanced parsing of multi-factor requests: you can ask Viv a question about the weather on a certain day in the past, in a certain city, and it’ll automatically pull out the various elements of the query to get to the right answer. Follow-up questions are also supported without having to start again from scratch: Kittlaus asked Viv for help buying flowers, then refined the search with “what about tulips?” and the assistant remembered the intended recipient and other details.
It’s not the only thing Viv can do. Booking hotels and asking for transportation via Uber is supported as well; in the latter, the assistant recognized from the number of people needing to travel that it would have to book using a service that offers a van rather than a regular car.
Sending cash to a friend was as easy as telling Viv to “send Adam twenty bucks for dinner”: it automatically pulled up Venmo, figured out which Adam, and populated the description.
The really clever thing is what’s going on behind the scenes. Traditionally, a virtual assistant would need to have each service, feature, and ability hard-coded, but the Viv team has developed what it’s calling “dynamic program generation” or the ability to create custom services in real-time.
“When it understood the intent of the user, it generated this program,” Kittlaus explained. “This is software that’s writing itself.”
In around ten milliseconds Viv can not only pull out the basic building blocks of the request – using Nuance speech-to-text – but create a one-time program based on the core abilities of services like Uber, Venmo, and others.
For instance, there are already routines for parking, finding places to stay, handling recipes, productivity like email and calendar, and entertainment such as wine recommendations. Each can be woven together into an array of Viv abilities, not only on a phone but potentially baked into every connected device in the Internet of Things.
The goal, Kittlaus says, is to see “how easy can you make it to get things done by talking to things” in a conversational commerce environment. That’s the obvious path to monetization, too, with the chief exec predicting that “for developers this is going to be the next great marketplace” after app stores.
“Imagine when you’ve got hundreds of thousands of developers plugging in new services,” he suggests, “and you’ve got the efficiencies of conversational services.”
That will obviously require no small amount of developer buy-in, but if the flexibility of Viv is anything to go by – not to mention the current interest level in bots and AIs – then companies in all fields are likely to be curious whether it’s the best way to deliver their services in a personable way.