Click here to return to the Bright Angles website

Friday, 24 May 2013

Predictive intelligence - the next generation of mobile apps

The new generation of mobile apps no longer wait for users to access them, or ask for help. Based on a new approach to software, these apps actively learn about their users and try to serve up timely information when it's needed. These predictive apps, based on machine learning, have the potential to make our mobiles more indispensable than ever before. Let's see some examples:

Search based predictive apps

Google has all the dope on you, and Google Now is an attempt to use this information to the user's advantage. Google Now, a build on Google Search, will automatically pull up a boarding pass from your email, the Google Map from home to office with a traffic prediction, and nearby tourist attractions when you are on holiday. It is an attempt to take search to the next level, by making things available before you search for them. 

iOS apps like Grokr and Osito are in a similar space, and learn intricate details about users lives including sensitive information like ethnicity, gender and age. This helps them to serve up better recommendations about events, music, or restaurants that might be relevant to the user. 

It is obvious that search-based predictive apps are an advertising goldmine, representing the ability to closely profile and serve customised content as well as ads.

mHealth apps

The mHealth sector is attracting increased attention from entrepreneurs and investors, who acknowledge its vast potential to reach lower income, small town and rural populations who are often under-served by the health system. Health apps also have the potential to manage diseases, bringing up alerts for patients and healthcare providers, anticipating need for hospitalisation, or providing information to prevent it. Asthmapolis,  the mobile app for asthma management has reduced patient hospitalisation by 70% and also cut the duration of hospital stay. This business model targets those who pay/ provide health care - for example, health insurance companies. Even consumers would prefer fewer hospital visits, so it is likely that preventive health will become one of the integral mobile services. 

The future of mHealth could be a device that monitors your heart rate regularly and calls for an ambulance when it suspects that you are about to have a heart attack. Meanwhile, it relays your information to the cloud, so it can be downloaded and analysed by doctors even before you have arrived at the hospital.

Social Intelligence

Artificial intelligence (machines which can think as humans do) have been the holy grail of computing since the days of Alan Turing. Apple's personal assistant Siri, represents one of the few hits in a field which has largely been about misses. Siri is not based on the most sophisticated AI  - it's based on a simple approach of understanding keywords and matching them to appropriate responses - but this is done with cleverness, humor and charm and that makes all the difference. Siri is a hit because she is fun to talk to. A lot of predictive intelligence is involved in building the data bank and pulling out the relevant responses. Will we ever create a machine which matches the human ability to use language? Probably not, but meanwhile, we can definitely create applications which more accurately recognise voice commands and provide response. It's not the easiest area to work in, but there lies the challenge, and the fun.

The Internet of things

Would you like to know  when the pot of rice on the stove is about to burn? What if you're out shopping and the fridge sends you a reminder that you need to buy tomatoes? The internet of things could make this a reality, in as little as five years from now. Already, 10 billion devices are 'wirelessly' connected to each other - including cars, TVs, gaming stations, kitchen equipment and more. As the figure grows, we will have more devices that talk to us and to each other, in a bid to save us time and money by anticipating what we need to do next.

Sources : MIT Technology ReviewVenture Beat