by Eckart Zollner, head of business development at the Jasco Group
The Internet of Things (IoT) is one of the latest technology buzzwords, promising to deliver many benefits by connecting devices with sensors that transmit data via the Internet to a platform where this data is stored. The real value in IoT, however, is the ability to analyse the data to glean valuable insights into a number of areas of business.
Much of IoT’s current use and benefits are centred around single-purpose requirements, such as identifying specific situations or anomalies and triggering a response, either automated or requiring human intervention. Data sets from a single source with predefined parameters are set to trigger at pre-defined events, where after people can get involved to resolve any issues or make adjustments. There is still a lot of human reliance in this process, from defining parameters to analysing data to responding to triggers.
Artificial Intelligence (AI) and machine learning can make a passive process far more interactive and less reliant on human interaction. A system that uses AI and is constantly learning can record these anomalies and automatically makes predictions and recommendations based on historical data, auto-adjusting pre-configures and effecting the change itself, or pinpointing to a tee exactly what must be done and what the predicted outcome will be if not attended to. Although a person may still need to physically intervene to correct the anomaly, they are now equipped with precise knowledge on what to do, where to do it, why, and even how.
AI allows a system to change its behaviour in accordance with the data received and how it is used. It is constantly evolving and learning based on new data received, making automatic changes to the system and adapting to constantly improve itself. It does not require people to be cognisant of incoming data and change the responses accordingly.
It essentially brings a further level of automation reliability and removes the possibility of human error, in terms of acting on specific data input. For instance, a smart, IoT-enabled heart rate monitor that is linked to a medicine dispenser, which dispenses the appropriate dosage, based on incoming data, can leverage AI and machine learning to automatically identify a problem. It can then act upon learned knowledge based on historical data to adjust the dosage without needing a doctor’s input, while alerting the doctor of impending problems and suggesting possible solutions going forward.
Data can be used in real time, as it is collected, to teach the system how to respond with a minimum of pre-programing, if any (in some cases). AI allows the IoT system to respond that much quicker, and with far fewer errors than in a passive IoT scenario.
One of the largest barriers to a fully automated, AI-enabled IoT solution is that it relies heavily on real-time data traffic. The network infrastructure that supports these solutions needs to be fast, reliable and always-on. Because the level of automation that AI enables occurs in real time, a small delay can cause problems and can result in a backlog on unprocessed, unregistered data.
Take the scenario of self-driving vehicles as an example. Sensor data allows for vehicles to drive without requiring an actual human driver. AI, on the back end, takes all the data from all the self-driving vehicles to provide information on best routes, minimising traffic congestion and avoiding possible accidents. If a particular piece of data is delayed in reaching the system, it can erroneously direct a vehicle the wrong way, causing rather than avoiding traffic, or even an accident.
Another barrier is the concern around the impact on jobs. Organisations are reluctant to embrace a technology which highlights the removal of human involvement as a benefit. However, AI brings with it its own requirement for specialised skills and effectively encourages organisations to re-skill rather than replace their workforce. Businesses can take mundane, time consuming, process-driven tasks from staff and, instead, retrain their staff in new areas of business. While this requires an entire shift in mindset, it also means that South African businesses can invest in driving South African skills towards a digital world, providing their staff with the necessary training to operate in a digital economy.
Enabling intelligent thinking
Organisations can derive a lot of business value by enabling their IoT solutions through AI and machine learning. However, they need to ensure that they have the right stability and speed built into their network. They also need to shift their mindset away from thinking of IoT as a process, rather than as a business enabler.