A Voice Layer on a Web World (and a shift in focus).

Just over two years ago in jumping the curve I wrote how we were shifting to artificial intelligence from traditional web application development.

Since then we have explored simple and complex NLU services. First we created a rules based, Facebook Messenger chatbot. Rules based chatbots have a roll but have to be programmed in an “if this then that” manner; they adapt by themselves. The AI behind a rules based chatbot is limited to natural language understanding.

Next, we combined our web applications development talents with AI to prototype an airline natural language chatbot. We first built an IBM Watson chatbot (now called Watson Assistant) and an associated NLU (natural language understanding) model.

The power of APIs.

Once  that worked we connected it to a new airline API that British Airways was implementing. The finished prototype let users get BA flight booking information just by asking a chatbot in plain English.

While chatbots were all the rage in 2016 and 17 they have been superseded by the VoiceFirst movement. It’s a natural evolution since voice doesn’t require an intervening device (the keyboard) to use, and voice assistants are ubiquitous now.

Most recently we created an Alexa skill version of Air Charter Amy. The question then became how do we monetize this? In fact, this was a dominant theme with many of the entrepreneurs I met at at Voice Summit 2019 last week.

The popular media tends to look at Alexa skills and Google Assistant actions as standalone applications, similar to a smartphone app. It’s a narrow vision of what voice is.

Voice is a user interface layer to NLU, which powers a voice-enabled use case. While that use case today is often a single skill or action, it does not have to be.

In fact for smaller companies, it is quite likely that deploying a standalone app is an inefficient use of capital. Deploying voice assistants on Alexa and Google is complicated, requires a voice search optimization effort, and requires ongoing updates to the bots. By analogy, it would be like an Inn or B&B developing and operating its own booking system instead of signing-on with Booking.com and Air B&B.

Voice can also be a portal to a marketplace whose participants share a common product. Accommodations, restaurants, airline flights, etc.

This is the approach we are taking at Planeviz.

Developing a voice-enabled platform is a full time endeavor. Come follow our progress at Planeviz