Monday, 29 May 2017

Twilio Signal learnings and comparisons. Options for India

I attended the Twilio Signal conference in SFO this week and this post is about the new offerings announced and how we compare with the feature list in India. What is available and what is not and how to achieve similar features.

TL;DR: If its voice, we have got you covered in India :)

First up, let me congratulate the Twilio team on a superb conference. Signal is a must attend conference for all cloud telephony enthusiasts. They have done an awesome job of keeping the conference interactive through hackpacks and also the silent event headphones were cool.

My general feeling was that this time around, Twilio was much more serious about contact centers. So its good that we at Ozonetel came to this realization about 4 years ago. We knew that the serious money was with call centers and thats why we launched Cloudagent separating it from the Kookoo platform.

Jeff Lawson, CEO of Twilio, shared their thought process with the below classification.

It was really satisfying to see this as we are also in pretty much in sync with this thought process. Our product portfolio mirrors this:

System: Kookoo API 
Department: Cloudagent.io 

Lets now look at the announcements and alternatives for businesses in India if they want similar features.
1. Twilio Speech Recognition: This was a long time in coming. So finally it was announced now. And as expected, the speech recognition heavy lifting is done by Google speech. Though speech recognition on phones sounds cool, given that Google speech cloud is now out of beta, this was available to all developers. Even to Twilio developers. Twilio developers and in India Kookoo developers can achieve something similar using the <record> verb and Google speech API.(Example). In fact, a student had built a cool hack almost 5 years back using the phone and Google speech.(Kookoo Siri)

2. Twilio Understand: This is one step above Twilio speech. Even this is not revolutionary, but more of an increment. Kookoo developers can use the output from Google speech and feed that to wit.ai or api.ai or any other AI/NLP engine out there. Twilio speech and Twilio understand do provide Twilio developers an easier way to get started on AI and machine learning over telephones and allows Twilio to get on the AI bandwagon. But, as explained above, both are not really revolutionary and most of the heavy work is actually done by Google.

3. Twilio Proxy: This one is a very good addition to the Twilio arsenal. One of the most successful use cases for phone systems in recent times has been number masking along with cloud call centers. Startups like Uber, Ola, Zomato, Swiggy and others have been using number masking to protect their consumer's privacy from drivers and delivery people. Twilio proxy will now allow everyone to deploy such features. In India, number masking has been used since the last 5 years and we, along with other cloud telephony providers already provide well rounded products for number masking. Startups can easily deploy such solutions already in India using very minimal coding.

Twilio also allowed a bunch of new features for their messaging products and video products. I did not listen to them in detail as I was more interested in the voice space.

One other product which was really cool was Twilio functions. This allows Twilio developers to deploy their Twilio app without dealing with servers.

So to sum up, Twilio announced a bunch of new features which will move the needle for Twilio developers. They can now build telephony solutions using speech recognition and NLP. They can also deploy easily to the cloud(server less) using Twilio functions.

And if you are in India, we have got you covered :)