When I ask people what they think the Internet of Things (IoT) is all about, the vast majority will say “smart homes,” probably based on personal experience. If I say that it is also about industries making using of data from sensors, then most people’s immediate reaction is to think of manufacturing. Sensors have been used for a long time in manufacturing, and the concept of using data generated at the edge to monitor and run automated processes is well understood.
This perception, however, is underselling the IoT. In practice, it can be applied anywhere.
The use cases for industries with “things” to monitor are easy to identify.
Manufacturing is one of the most obvious. Connected sensors can be used to monitor and manage the health of manufacturing equipment, identify root causes of defects and improve quality.
Health care has equipment that generates digital information about how patients’ bodies are working (e.g., blood pressure) and what they look like (e.g., scans). There are numerous opportunities to monitor people’s health more closely and accurately and catch signs of disease early, or even avoid it altogether.
The insurance industry is using telematics to monitor driving behaviour and assess the risk posed by individual drivers. Telematics also helps with the claims process because information from before a crash can indicate who is at fault, and images of a damaged vehicle can be used to assess whether the car should be written off or repaired.
The IoT also, however, has potential in industries that, on the face of it, do not really have “things,” such as financial services. Banks and other financial providers are extremely interested in the IoT, focusing on “things” which do not belong to the banks themselves, but to customers: mobile phones and payment cards, for example. Banks can improve fraud detection by notifying customers each time their cards are used – in real time – and also checking that the customer is with the card at the time. That, clearly, is a huge service for customers: no more cloning and no more fraudulent transactions.
A change in business model
A fundamental shift in business model is being enabled by IoT analytics: a move from products to services. For example, Rolls-Royce is traditionally considered an engine manufacturer. The company made and sold engines, then sold services to maintain those engines. Now, however, rather than pay for maintenance, airlines can choose to pay an hourly rate for the time that the engine is propelling the aircraft. In other words, it can pay for what it actually wants: the plane in flight at particular times. Increasingly individuals, too, are choosing to pay for a service, rather than goods, such as access to a car-sharing service, rather than owning a car.
This shift, however, has challenges for the service providers. If you are providing a service that includes a physical asset, you do not want to have to spend time and resources inspecting that asset. Instead, you want it to run itself as much as possible. The IoT allows providers to remotely monitor and collect data on all the important aspects of each asset – how it is performing, how it is being used and environmental factors, for example – and therefore automate much of its management.
The data collected from the IoT is only really useful when you can derive useful intelligence from it, and preferably in an automated way. This automation, however, requires intelligence, and that means artificial intelligence (AI).
The importance of AI – and the problem
This is one of the biggest reasons why the IoT is really taking off now: AI algorithms are becoming more usable. There is, however, still a problem. Most AI algorithms need huge amounts of data and computing power. They therefore rely on powerful servers and central data storage.
In computing terms, we humans perform most of our computation and decision making at the edge (in our brain) and in the (pre-)moment, referring to other sources (internet, library, other people) where our own processing power and memory will not suffice. This is more or less the complete opposite of the current AI algorithms, which tend to perform most of their calculations far from the data source, in servers, drawing on stored data.
To enable timely decision making in the world of IoT, you need to be able to deploy some of the cleverness (predictive models and decisioning rules) at the edge, closer to the “things” that you are managing. Some businesses are already doing this, whilst many others are still trying to figure out how to organise and make sense of the deluge of data available to them. Those at the forefront of combining AI and IoT have a huge opportunity to steal a march on their competition.
In my personal view, this is the biggest change in business models since the dot-com boom. And, as in the 1990s, there will be some big winners, and there will also be those who don’t quite get it right, and fall by the wayside.