Viv revealed — the rise of the Assistant

--

OMFG!

They pitch it as:

“Radically simplifying the world by providing an intelligent interface to everything”

Viv revealed

Dag Kittlaus demoed Viv for the first time today (Monday May 9th, 2016) and finally revealed just what makes Viv different from previous ChatBot AIs.

For one — Viv is much more than just one Virtual Assistant. Its a bot marketplace, providing a single point of entry for an entire industry of ChatBots, AIs, Agents — call them what you will.

Viv does this by offering a radically new way to “code” Virtual Assistants. They call it “dynamic program generation.” Instead of coding every single aspect of a computer program to implement a ChatBot — Viv enables coders to work with a new way to program — defining what the intent, goals and milestones of a ChatBot — and allowing those programmers to describe and model that intent.

Since Viv will reside on smartphones, inside automobiles, kiosks and a wide range of electonic and IoT devices — Viv ends up looking like an AI platform. When developers create Viv Virtual Assistants they will have to take into account the “context” of who the user is, where they are and on what device or context they’re accessing Viv.

Viv enables end-users to interact via a conversational UI (whether through messaging of voice recognition) and put together sophisticated queries — which match what natural language is today. This means that those short, blunt Siri queries — which barely work — will become a thing of the past.

Viv’s sentences can be made up of multiple parts, recall previous time and location information and remember who the end-user is — as well!

Our industry has never seen this kind of end-user personalization, or had hands-on access to a platform — that others can get develop and market their own products or services.

The paradigm shift that Viv enables means that they’ll be no more on-boarding. Once an end-user purchases a Viv enables device, all of the global settings and preferences that the end-user has trained Viv about — will be downloaded immediately into the newly purchased device.

End-users will be able to walk up to Viv kiosks, get into Viv cars, work in a Viv enabled workplace — and get all of the benefits of being at home and accessing one’s work, media content or the of web services they’re members of.

eg. goodbye Apps

Viv experiences do not require:

  • finding the App, downloading it
  • learning the App, come up the learning curve
  • trying to move info or context from one App and apply it to your next task or planning endeavor

Viv integrates all on-line tasks into one AI experience.

The evolution of the world of Apps towards a world of APIs has commenced and its driven by Viv.

How does Viv work?

Dag Kittlaus opened up the kimono just a bit at totday’s demo by talking about and showing the Viv Developer Center. Once a query has been spoken into the Viv engine, it is broken down into this “dynamic program generator” form.

Here’s a sentence that reads: “Was it raining in Seattle three Thursdays ago?”

Another sentence demoed was: “Will it be warmer than 70 degrees near the Golden Gate bridge after 5pm the day after tomorrow?”

You can see from the “sentence editor” examples (above) that Viv takes a phrase (or sentence) and breaks it up into a flow tree for language processing.

Presumably developers can create and add their own:

  • nouns, verbs, and other kinds of language constructs
  • tasks or actions — “coded up” in Viv
  • use case scenarios, UX procedures, context adjustment and exception handling

This makes Viv a mainstream AI programming environment in a world where this kind of functionality has been limited to Prolog, Lisp and other forms of esoteric AI languages.

Dag then quickly showed the kind of rich Knowledge Base and range of libraries built into Viv.

Here are two screen grabs for “weather” and “wine.”

Notice the inter-connected aspects of the “Knowledge Base.”

Here’s what a Viv knowledge Base about “wine” looks like.

In conclusion Viv is going to change the world. Now lets see how far they get with distribution.

--

--