According to certain industry analysts and software vendors, we are now midway between a stage 10 years ago when few applications used machine learning, and a stage 10 years into the future when apparently, most applications will function with it.
The Gartner “Hype Cycle” shows machine learning due to become mainstream in software in about four or five years’ time. In that case, business continuity, like any other area of business activity, is likely to be affected. The time to start thinking about it may well be now. But what is machine learning, and how might it influence BC planning and management?
Simply put, machine learning is the capability of systems to construct models from data, without the intervention of human beings. Examples of machine learning today include self-driving vehicles and fraud detection, both of which reflect aspects of business continuity. Machine learning can be done in two ways.
First, supervised machine learning, in which programmers “train” the software application with specific models and example (“training”) datasets, to enable the application to identify such models in new datasets afterwards. Second, unsupervised machine learning, in which the software application derives its own models from datasets without any prior examples.
Machine learning contributes to business continuity through this identification or creation of models. It can then be linked to decision-making software able to ensure business operations are adjusted as needed to avoid interruption. In a practical sense however, the advantages of machine learning may still be difficult to obtain.
Unlike business intelligence, which is now available on user-friendly, affordable platforms, machine learning still requires specialist skills and longer term investment to achieve useful results.
Yet in a few years, there may be inexpensive solutions available with similar ease of use to today’s business intelligence apps. So, watch and wait, and be prepared to seize opportunities for better, more cost-effective BC as they arise.