AI First

“An important shift from a mobile first world to an AI first world”.

With those words, Google’s CEO Sundar Pichai summarized the Google I/O 2017 keynote. His summary wasn’t chosen randomly: almost all of Google’s announcements were related to artificial intelligence (AI) in one way or the other. And it’s not just Google. Microsoft, Amazon, Facebook and Apple have all made similar announcements during their keynotes.

This is an abbreviated version of an article by Thomas Mons in Shift, In The Pocket’s annual report on the 10 next big trends in digital. Download Shift and read the full version.

The battle for your living room

Google has its Google Home with Google Assistant, Amazon its Echo with Alexa and Apple its HomePod with Siri. Microsoft currently relies on HP and Harman Kardon to productize its Cortana technology as a standalone speaker. They are all able to help you in the same way a personal assistant would: you just need to ask.

These smart speakers allow third parties to develop new services, much like apps on traditional operating systems. So developers can exploit a completely new digital channel with a minimal amount of effort. The hard problems are solved by the platform (understanding intent and context). All a developer needs to do is build a new front-end.

Companies that have invested in their APIs over the past years, are now able to move quickly and offer their users compelling cross-channel experiences.

Companies that have invested in their APIs over the past years, can now move quickly to exploit this completely new digital channel.

A good example of such an experience is the Netflix integration on Google Home. Just saying “Ok Google, Play Narcos on the living room TV”, will open Netflix on the TV in your living room and resume the season where you’ve left off, regardless on which device that was.

You could also ask your assistant to order food, get an Uber, turn up the thermostat or remind you to get flowers for your mother’s birthday. It’s not hard to see how these smart speakers could quickly become an essential part of our everyday lives.

Understanding a user’s context and assisting her or him with relevant suggestions, wherever, could cause a disruption in multiple markets.

So why is this battle so important for the big tech companies? With these virtual assistants, they can build rich user profiles and understand context, which is invaluable information in a lot of use cases. The stakes are high: the ability to understand a user’s context and assist her or him with relevant suggestions, wherever they are, could cause a disruption in multiple markets, such as advertising, to name but one.

An explosion of ready-made services

Virtual assistants are a prime example of how AI will change digital products in the coming years, but it’s far from the only one. In the past year, we’ve seen an explosion of ready-made services. They open up advanced technology in a way any developer can use it.

Ready-made services open up advanced technology in a way any developer can use it.

Previously expensive or unavailable solutions in the fields of natural language processing and computer vision can now be used readily to improve a digital product, reinvent an experience or enable further automation.

The possibilities of today’s ready-made services are growing by the day.

• You can detect people, objects, text, landmarks and settings in images, for instance with Google Cloud Vision API.
• Services like Microsoft Face API or Amazon Rekognition help you find your friends’ faces within your photo library or predict people’s ages and genders based on a photo.
• Other examples are luis.ai or wit.ai, which can detect intents and entities based on what a user types or says, in multiple languages.
• Tracking faces, text or barcodes in video is made possible by Apple Vision framework and its equivalents.
• …

Machine learning at your fingertips

But things don’t stop there. Google, Microsoft and Amazon are actively supporting the ability to create, train and evaluate machine learning models on their cloud infrastructure. This alleviates the process of training and deploying machine learning models in production.

All you really need, is a decent set of data that can be used to train your model, and a clear idea of the unknown that you want to predict. Once you have that, start with something simple that works reasonably well and keep improving on that.

Ready for the shift to AI first?

Thanks to these recent developments, it’s a relatively small effort to include AI technologies in your digital products. Or even build entirely new products on top of them. Opportunities are abundant.

It’s a relatively small effort to include AI technologies in your digital products.

Imagine how you could use AI to amaze your users, increase product success, enhance operational efficiency or maybe even disrupt your market. But make no mistake: while AI might be new today, it will be considered normal tomorrow. Users will expect it. And at the rate our ICT world is evolving, a new shift is never far away.

Originally published by Thomas Mons, our director of engineering, at inthepocket.mobi.