Subscribe to emnos news!
, created by Jesus Garcia
The rise of the machines - Fact or fiction in data analytics?
With movies such as Avengers – Age of Ultron and the ubiquitous Terminator franchise, Hollywood continues to kindle the debate on man versus machine focusing on everything that can go wrong when machines go rogue! Entertaining fantasies, no doubt but there is also no doubt that “intelligent” machines will have a role in the future of commercial data analytics.
It is generally agreed that data is growing at an increasingly faster rate and the ability to handle it needs to increase at the same pace. But where there is no destiny, there is no road - or in other words, what is the point in building the capability if we do not clearly know what is to be achieved with the data. The bottom line is that time and money spent bringing all data together are wasted when only a percentage of the available information is valuable or utilised.
Bringing the technology closer to business core activities will help you know what you can do - and then to do it without filtering through masses of unnecessary information. The data analysis process needs to move on from providing hypotheses to delivering solid, business oriented decisions. Only then is the data completely at the service of the organisation and can develop into a proactive, fully optimised advisor.
From fiction to reality
In the retail and FMCG sectors our first steps towards machine learning began with distilling the insights which impact merchandising, pricing and promotional choices on a daily basis. Next our smart software began to provide recommendations and predictions. But what if the machine could go one step further, become even more hands-on and start making real-time decisions which positively impact the store and/or the assortment - and all without human intervention?
There is already easy-to-use, yet extremely smart software available which helps people do a better job with less effort – automating routine decisions is simply the next step forward. This idea definitely falls within the bounds of reality and, to an extent, is already part of our lives. Online recommendations from Google, Amazon and Netflix are good examples of machine learning applications we take for granted.
And what about self-driving cars – a concept which has definitely moved from Hollywood fiction to reality. Remember KITT, the artificially intelligent and self-aware car which fought crime alongside a very young David Hasselhoff in the 1980s Knight Rider TV series? At the time, it seemed like pure science fiction but a mere 30 years later, much of the technology is now being developed.
Many daily business decisions are recurring and could relatively easily be automated by linking data analysis to business processes. The machine would, of course, learn through iteration and independently adapt when exposed to additional data. Using previous computations, it would develop the ability to make reliable, repeatable decisions. Data can bring intelligence to business processes and save both time and money by taking care of the repetitive and low value tasks which unnecessarily occupy a lot of people in the organisation.
Today we use data mining to uncover patterns and knowledge to support decisions. Tomorrow the machine learning algorithms will reproduce the patterns, automatically apply them to other areas of the business, and in turn, decide and act. And we should face this not with fear, but with an intelligent understanding of how it can improve commercial processes.