Predictive Apps are the New Relationship Between User and Smartphone
Mobile phones, or smartphones, are no longer accessories but extensions of us, as users.
This post was originally published at the RankMyApp’s Blog
The development of technology in the mobile handset industry is increasingly innovative. Nowadays, the cell phone is no longer just a means of communication with voice or text messages. These functions remain, of course, but with the possibility of a mobile device providing so many other purposes, with computational capacity and sensors, obviously, that promise was positive.
Mobile phones, or smartphones, are no longer accessories but extensions of us, as users. They are personal assistants who remember the appointments, wake us up in the morning, store our contacts, photos, and music. As if that were not important enough, some sensors coupled to phones make it possible to get the pulse, cholesterol level, temperature, heart rate, location. With a database full of information, it’s easy to find out what we want, where we’re going, and what we’ll do.
Within the infinite possibilities, a new concept of the app is making the users’ lives easier, the Predictive Apps. Unlike a computer, for example, which needs an active operator to command the functions and click on programs for them to run, the Predictive Apps are proactive and automatically anticipate the user’s need.
These actions are based on machine learning software that, through the users’ personal data, captures their needs and develops its own functions. With intuitive interfaces, the model of communication with this app flows naturally, as demonstrated by Google Now, Siri and Cortana.
As a prominent example, Google Now identifies when the user is about to take action and offers help or even, based on inferences, makes decisions by itself and only ask you for approval. Your database is made up of email account information, calendar and surveys done on the internet. That way, the app knows where you live, work, and what roads you usually take, so it always gives you the best traffic information. It can also analyze what you eat (using the camera) and with previous information about your diet, alert the level of calories you are ingesting. With your location, if you are walking, for example, it also calculates the speed of your step and the data in your calendar, identifies your pace and tells you whether you will be on time or late for the appointment.
A new generation of apps
Predictive Apps are the future of technology within our reach, they indicate that there is room for innovation and that the potentiality of these sensors needs to be explored. There are still many current apps that act as desktop versions for smartphones and tablets, which don’t use the potential of the sophisticated predictive algorithms.
This technology layer of contextual apps already exists, such as NLP (Natural Language Processing) systems, which are already well evolved and improving every day. In 2011 came the IBM’s Watson and the Apple’s Siri, and since then the capacity of computers has been actually visible. Over the years, the evolution of interfaces in natural language is evident and the big bet is that they will evolve much more, to the point that the use of NLP as a natural communication interface will be something common.
With the acceptance of the first generation of Predictive Apps, the new question for engineers at Google and elsewhere is to look for ways to get data about their users. Bill Ferrell, Founder, and CEO of Osito, a company that offers an app with similar functions, said its engineers are trying to learn more from people’s past traces of predictions about future activities, all to refine information later and offer an increasingly more accurate service.
Almost now, Google Now has recently started showing a weather forecast for places where the app thinks you might be going. Besides, it starts notifying about houses that are on sale in your neighborhood, in case you have searched something on the Internet that may suggest that you are looking for a new home.
Did you find it too far fetched? Some machine learning experts go even further. Grokr, an iPhone predictive app, has discovered a way to differentiate users’ gender, ethnicity and age, all with a high degree of accuracy: “This can help us predict the places you’d rather go,” says Srivats Sampath , CEO of Grokr, who wants to use this information to address the recommendations the app offers for music events and restaurants, for example.
Technological, but not human
The way these apps are presented is a striking factor. Different than you can imagine, there is no idealization in giving personality to the app, to bring it closer to the real. They benefit by means of data mining techniques, of course, but they are not virtual butlers. In 2010, Apple launched Siri, a voice-operated app that has become known for rightly mimicking a virtual assistant with almost human Intelligence, as well as giving funny responses to users (read more about it here). “An assistant is probably the worst use case because you create the expectation that it will reach the human level,” says Mike Volpi, a partner at Index Ventures, a company that invested in Donna, an iPhone predictive app.
In the case of Google Now, Bit.ly Data Scientist Hilary Mason believes that some of the information the app provides is unnecessary. She says that, for example, the app does not need to tell bus timetables every time the user walks past a bus stop: “He’s not exactly focused on what matters to me,” he confesses. Either way, Mason recognizes that the Predictive App represents a milestone in computing: “It’s important because it’s the first time Google has taken everything it knows about us to make a product that makes our lives better.”
What do you think of this close relationship between user and smartphone?
Predictive Apps are a technological breakthrough — that’s undeniable — but the intrusion into a routine helps or disturbs you?
Twitter: @ju_rankmyapp | Linkedin: julianassuncao | RankMyApp’s Blog
