Humans have always demonstrated strong ability to analyze issues, and form expert opinions. And with internet at their fingertips, no topic is out of reach. The only problem is, most often than not, the opinions of various experts don’t match. These informed decisions involve fair amount of predictions and lots of intuitions. Now let’s face it, in the complex business world, with so much uncertainty, no one can tell whether a decision is going to be a right one or not. But at the same time, businesses cannot afford to gamble all the time. There has to be a scientific solution, and hence, businesses are adopting analytics.
Analytics is a systematic analysis of available information using statistical techniques to take “not so obvious” critical business decisions. Complex machine algorithms are designed to automate this process in large enterprises. This method has significant advantages.
– First of all, it attempts to standardize the way decisions are taken in an enterprise on a particular problem, so that it is not based on human discretion.
– Secondly, we can leverage the machines’ potential to process vast amounts of data or information, which is sometimes is necessary. Machines are also more consistently accurate than humans, and can process predefined algorithms quickly and consistently.
– Thirdly, all the agreements or disagreements can be discussed while creating the analytical model, and ones there is agreement on a specific model, then it is assumed that there is also an agreement on every decision taken by that model.
All this is easier said than done. Huge investments are required to ensure that the underlying data is accurate and complete. Analytical models (and the underlying data) needs to be constantly reviewed and updated to keep them relevant to the ever changing world. Also machines can many-times never understand the ‘human aspects’ of a problem.
Can these decisions be wrong? yes.. absolutely!. Although the machine algorithms can be improved overtime, they can never be accurate, as most of the real life complex situations are based on the probability of occurrence of an external event, which we have very little or no control on.
In reality, almost all the times, critical decisions are still taken by humans, and machines merely provide the supporting data to justify the rational for the decision… reducing the risk to some extent, and helping to get a consensus.
This has been the trend until now, but how the future may look is yet another discussion. New machine learning algorithms are being designed to make machines learn and think on their own. Machines of the future will have to take tons of decisions simply to operate themselves in an automated manner. With machine to machine communications on the rise, we see machines taking most of the decisions on their own, followed by actions without any intervention of humans!!