The word ‘revolution’ is anathema to most business managers. It smells of disruption and risk – and no business wants unnecessary risk. However a revolution was eagerly embraced in the last decades of the previous century, when businesses around the globe embraced information technology to automate their business processes and transactional activity. This was the second business revolution – the first happened when machines prompted the industrial revolution in the nineteenth and twentieth century.
And so we are on the threshold of the third revolution. This is the automation of decision making, and its implications are far reaching. Virtually every aspect of business life involves decisions. Call centre operatives have to decide which offers and deals might be most attractive to customers. Procurement requires decisions on the most reliable suppliers offering the best value. Finance is required to establish which customers are worthy of credit. Many of these decisions are embedded into operational systems, and most of the applications used day in and day out have thousands of decisions embedded in them.
This third revolution has become possible because of the strides made in machine learning techniques. These are increasingly employed to create predictive models, addressing anything from the probability that a patient might be readmitted to hospital, to the likelihood that a customer might default on a payment. The second revolution mainly affected clerical workers. Automating business processes meant that head counts could be reduced while increasing work throughput. The third revolution will affect a broader category of people. Process automation did little to affect the productivity of underwriters for example as they assessed applications for health insurance, through to young drivers wishing to insure their overpowered cars. Predictive models are changing this, and the underwriting process has seen massive increases in productivity and efficacy as these models automate the underwriting process.
Resistance is futile. The best advice is to jump on the wave before being engulfed. Decision automation will eclipse process automation within a decade, and differentiate those with their heads stuck in the sand from those riding the wave, in a way we haven’t seen before.
Big data driven by rapidly falling data costs, the Internet of Things (IoT) producing huge quantities of low cost data, the rapid evolution of machine learning technologies, low cost data mining tools, and the pressing need to become ever more efficient and effective, create an unstoppable force. Better be there.