How Smart Lives Will Flow from Insight Streams
In the future, we’ll be surrounded with smart devices that anticipate our wants and needs by capturing and interpreting various streams of data in real time. This vision of so-called “insight streams” flowing across the Internet of Things will be built on emerging cognitive and machine learning technologies, and will prove to be a major disrupter to business plans.
Getting on the right side of this coming wave of technological disruption will take a lot of hard work, good timing, and luck. But organizations that start preparing themselves to adapt to new digital business models will have an advantage, according to Constellation Research founder R “Ray” Wang, who popularized the term “insight streams” with his 2015 book Disrupting Digital Business.
In the book, Wang wrote:
“One of the biggest opportunities for monetizing digital business will come from insight streams. These insights will come from both least likely sources and the most obvious. For example, least likely sources include the amount of power consumer, water used, visitors into the building, foot traffic on the sidewalk, and density of the parking lot. These sources may seem mundane and useless information to most of us, but large insight brokers will take that data to drive contextually relevant information.”
We’ll be surrounded by 80 billion sensors in various devices by the year 2020, according to Wang. These sensors will measure and transmit all varieties of data, including workforce performance data, customer satisfaction data, and product quality data. Companies that are successful in the new digital marketplace will be able to utilize these insight streams to create new services, or have some other advantageous impact, such as lowering costs, improving customer satisfaction, or driving demand.
These insight streams will be leveraged by creative companies that have the technological aptitude to hook it all together with the appropriate context. Consider the case of the smart car maker that connects the fuel gauge in the car to a mapping service that tracks the location of gas stations, as well as the cost of fuel. “I’d personally pay a few dollars a month for a contextual service that delivers the peace of mind of never running out of fuel on the road,” Wang writes.
One firm that’s looking to play a big role in brokering insight streams across the emerging IoT world is Neura. The company’s CEO and founder, Gilad Meiri, uses the term “Alternet” to describe how smart devices will automate mundane tasks and deliver massive personalization through the correct interpretation of contextual relevancy.
“The Alternet needs to understand me, understand my context, and tailor its behavior to my needs,” Meiri tells Datanami. “The more connected people are, the more pattern-driven, the more compute power that’s available, the better experience we can create.”
Since it was founded four years ago, Neura—which uses the Matlab statistical package from Mathworks, among other tools–has worked to develop an application that can apply context and relevancy to data collected from a variety of devices. The idea is to create an ecosystem of connected apps and devices that can improve the lives of people through machine learning-driven personalization and optimization.
For example, when one or more devices register the fact that a user has fallen asleep at his home, then a number of activities can be automated, such as a door knob locked or a lightbulb turned off. While each automated device could be programmed to take certain actions, the value to the user increases substantially when the devices work together in harmony.
The challenge is getting actionable information out of the devices, without requiring the user to issue a precise command or continually resetting his state. That’s where the machine learning algorithms come in.
Building algorithms that can accurate reflect reality is not as easy as it might sound. For example, the decision tree algorithm that’s used to detect whether a user is running or not took a lot of hard work. Neura’s algorithms don’t’ just look at data from the accelerometer in the user’s phone, but also look at the presence of other connections, such as a Bluetooth connection, to infer state.
“The understanding that someone is driving a car, for example, or commuting by bike riding is a much harder job than it sounds,” he says. “We break life down into 60 components. To take each and every one of these components and run them through this [machine learning] progression can easily take two years. It’s a long hard job of iterating and iteration. You get to a point where it’s commercially viable, and then you continue iterating.”
If the idea of machines knowing and broadcasting intimate details about our lives – like when we’re on vacation, where we sleep, and even who we’re sleeping with – makes you a little nervous, that’s a good sign.
One of Neura’s basic principles is that the user has full control over what insights to expose. If the user only wants his smart lock to know when he’s 20 minutes from arriving home – and not th
at he’s travelling on the other side of the world – then that’s all the lock will know.
While smart cars and smart homes are emerging, in Meiri’s mind the biggest opportunity revolves around smart health devices. “Smart home and smart car, while the upside is big, I don’t think the relevancy is immediate,” he says. “But if you look at smart health and wellness, it’s accelerating big time.”
For example, the rise of smart glucometer promises to help a surging population of diabetics better manage their condition. “The opportunity is vast,” Meiri says. “How can a glucometer serve you better by knowing you more? Can a glucometer increase compliance by reminding you to measure just before you start a run or just before you leave home if you haven’t measured yourself in the morning yet. Or can a glucometer realize that you just finished a run and you’re at risk of hypoglycemia and interject?”
The rapid spread of smart health devices, such as Fitbits, foretells success in the field. According to Meiri, in just a few years, more than 10% of the North Americans own some sort of smart health devices, while more than 20% have a smart health app on their phones. “It used to be a supply-driven phenomenon…but now the demand side is waking up,” he says.