Wit.ai: Facebook's Secret Weapon

Wit.ai: Facebook's Secret Weapon

Discover how Facebook is democratizing Natural Language Processing with Wit.ai, making it easy for developers to create intelligent bots and decision-making systems in minutes.

Gerardo Ortega

Facebook has clear objectives, which is why Mark Zuckerberg himself presented at the last F8 (the annual event where they present their innovations) their plans for a 10-year period, of which we’re already almost halfway through. The first five years focus on the services and products we already know: Facebook, Messenger, Instagram, WhatsApp… but the next 5 years include three interesting categories: Connectivity, Artificial Intelligence, and Virtual Reality.

Of the three categories that will carry the most weight in Facebook’s strategy for the next five years, the one that has caught my attention the most is Artificial Intelligence, because unlike the other two, it’s already generating profits and is an integral part of the products they currently have on the market. So all they need to do is launch many of their internal products as additional services. And of course, it was to be expected that a company like Facebook wouldn’t sit idle, and has begun acquiring companies and hiring talent related to this area. Although many of these moves that the social media giant has made have gone unnoticed. And today I want to talk specifically about one: Wit.ai

What is Wit.ai?

wit-ai-fb

Wit.ai is a service (acquired by FB last year) that aims to bring Natural Language Processing (NLP) to the masses (well, without exaggerating, to most developers who don’t know much about artificial intelligence or simply don’t have time to develop their own NLP platform), allowing you to create practical applications in a matter of minutes (yes, I’m not kidding), for example:

Bot Development

Creating bots for different platforms (Slack, FB Messenger, Telegram, etc.) is now trendy. With Wit.ai you can create a bot that is autonomous and learns over time. It has an interface specially designed to simplify the creation process, and several SDKs to integrate it with your favorite language, not to mention they have excellent documentation for their API.

Decision-Making Systems

Yes, many times you encounter cases where given an input (a tweet, a voice command, a message, etc.) you must make X or Y decision to execute a certain action. Depending on the complexity of what you need, Wit.ai can also help, allowing you to analyze inputs, extract the data you need, and make decisions by executing predetermined actions.

Voice Recognition

Wit.ai was designed to respond to the growing demand for interpreting and responding to voice commands. This, in the face of the growing trend of having everything interconnected (Internet of Things), is very useful. For example, using Wit.ai you can automate actions in your home, such as turning lights on or off, using only your voice and programming the interaction with the lighting system.

Automation

This is another very trendy topic. After all, we live in the era of the “Internet of Things”: everything is connected, there are sensors everywhere, and brilliant things like Arduino make things easier for us. However, the elements we interact with are not necessarily “intelligent”, but Wit.ai can take certain data, interpret it, and make decisions based on that.

Those are some of the applications this service has. I know, it seems like a paid advertisement, but we have to call good things by their name. And now, let’s look beyond. The point I really want to highlight here is that Artificial Intelligence will increasingly take over new areas, and organizations that don’t take advantage of these technologies will be left one step behind the competition. In a few years, doing without AI will be the equivalent of not having a website today.

This has been just a taste. Next week I’ll be publishing a tutorial and an example of creating a bot for Facebook Messenger using Wit.ai.

Gerardo Ortega

About Gerardo Ortega

Software craftsman with a focus on scaling, polyglot programmer, coffee enthusiast, and lifelong learner. Passionate about machine learning, data science, and building great products.