There may be a day when automation takes over human decision making in business.
The machines would replace managers who today rely on gut instinct, experience, relationships and pay-for-performance financial incentives to make decisions that sometimes lead to bad outcomes.
So says Gartner analyst Nigel Rayner, who has staked out a “maverick” position on this topic, meaning one not sanctioned as an official view of the analyst firm.
But hearing Rayner, and reading the latest economic headlines, begs the question: Would we be better off if machines made more business decisions?
Rayner didn’t dive into the politics — and he didn’t have to. Gartner executives already did that at its Symposium/ITxpo here with a no surprise warning of an impending second recession .
Instead, Rayner cited a U.K. study of 350 of the largest firms that found that pay for company executives rose by 700% since 2002 to this year, while the value of those firms went up by 21% and workers pay went up by 27%.
“There has to be something questionable about that picture,” Rayner said.
One problem is corporate pay-for-performance policies, said Rayner, noting that there’s a lot of evidence that paying people bonuses for knowledge work actually decreases their effectiveness.
“This pay-for-performance mentality,” said Rayner “doesn’t seem to be working right.”
What will arise to challenge or augment human decision making will be machines. Over the next 40 years, says Rayner, IT will be moving increasingly in the direction of automating human decision making.
“We humans are very bad at making decisions,” said Rayner.
Some of this change will be driven by need. Businesses are becoming overloaded with information, and the idea of a central data warehouse is being replaced by a “collective,” where information pours in from the outside into enterprises.
“Everywhere I look, predictive modeling, machine-based algorithmic systems and computer-based simulation outperform humans,” said Ryan.
That’s partly because humans are informed by our cognitive basis, said Rayner, “When we look forward we are too influenced by what’s happened in the past and our own perception of what we want to happen, rather than taking a rational view.”
Even in hiring, machines have an edge, said Rayner.
Talent management software uses a set of validated statistical assessments, developed by industrial psychologists, that result in much better results in terms of people who stay with the company longer, are more productive, and a much better fit for the job, said Rayner.
What Rayner sees IT developing not just dashboards, which provide a view into the past, but “intelligent business operations,” that treat businesses as a self-optimizing system operating in real-time, allowing the business leaders to focus on innovation in products and services.
The technology, in bits and parts, is already here. Companies are learning to handle Big Data, and emerging systems — evidence IBM’s Watson supercomputer — are increasingly capable of sophisticated analysis.
With the advent of a machine-based decision making, Rayner sees something of better future, with a more open and balanced economic system. The negative may be more job destruction and increased social divisions, he says.
Kevin Connor, an enterprise architect at Urban Science, which makes analytical systems, was among those who heard the talk and agreed with the overarching point.
Connor is already working in this predicative realm by automotive retailers who use his firm’s product to combine data about customer preferences in vehicles, with demographic data to help determine what buyers might be interested in.
But Connor said the problem he faces is in finding the right staffing resources.
Building systems that can make these kinds of decisions may require people with a mix of statistical, database and computer science skills.
