We watched with great interest this past Monday as Dag Kittlaus, CEO and co-founder of Viv, demoed the highly-anticipated “next generation” smart voice assistant at TechCrunch Disrupt New York. We wrote about Viv over a year ago, speculating about the new assistant’s technology and implications.

Kittlaus’s live demonstration of Viv was impressive and went off without a hitch. There are at least two aspects of this new smart assistant, its underlying technology and integrated ecosystem, that brands and customer care managers should take note of. The first is a ground-breaking ability to generate code that can anticipate complex user queries. The second is an urgent need for brands to get into the “inner circle” of preferred providers that are integrated within Viv (and other intelligent assistant ecosystems).

In this post we take a look at the implications of Viv’s code-generating process for brands and customer care managers. We’ll save the discussion on the need to get into the “inner circle” for an upcoming post.

Kittlaus highlighted a significant new invention in his demo of Viv’s capabilities. Viv relies on the same speech recognition, natural language understanding, and user intent mapping technologies that drive all intelligent voice assistants. But the team at Viv has developed a process they call “dynamic program generation” that enables Viv to program itself based on what the user wants to accomplish.

Here’s an example of how Viv’s dynamic program generation could work. As a programmer building an intelligent assistant that relies on natural language input, you might configure your assistant so that it can accurately understand and respond to user inputs such as:

• What’s the weather going to be today?
• Make a reservation at Main Street Cafe for 8pm.

But what if the user only wants to eat at Main Street Cafe if the weather is nice and they can sit outside? How can you program your assistant to anticipate and correctly act on user inputs like this one?

• Make a reservation at Main Street Cafe for 8pm, but only if it’s not going to be raining when we’re there.

Based on what Kittlaus showed, Viv is able to “program itself” by dynamically generated the code required to process a more complex user intent such as the one above. Will customer care intelligent assistants have this same capability in the near future? One can imagine lots of examples where this type of automatically scalable intelligence could benefit customers. For example:

• Are there any concert tickets under $65 within 25 miles for any of my favorite bands?
• Is this a good lawn mower to buy based on the reviews? Is this the best price I can get, and can I return it easily if I don’t like it?
• If I book this vacation, will I still be within my quarterly budget, and if not, are them some better vacation deals that my family would enjoy?

In a recent post, we talked about how intelligent assistant technologies are entering new territory and tackling more difficult problems. Viv’s dynamic program generation might result in a more powerful digital assistant than any we’ve seen so far, raising customer expectations about their self-service experiences.

Categories: Conversational Intelligence, Intelligent Assistants, Articles

Tags: inteligent assistants, Viv, Dag Kittlaus