It’s pretty clear that smart machines—computers and robots that can digest information, make recommendations and decisions, and take informed actions—are going to be a significant factor in the workplace of the future. In some areas, like insurance underwriting, credit decisions, and financial trading, they’re already in wide use. In other areas such as medical diagnosis and treatment, document analysis in commercial litigation, and digital marketing, they’re taking hold rapidly.
But the prospect of a lonely intelligent machine in a lights-out office is not very likely. Smart machines take over individual tasks rather than entire jobs. We humans will be working alongside them for the foreseeable future. Leaders, then, need to prepare their employees and organizations to collaborate effectively with cognitive technologies.
The first task is to consider different processes within your business and assess the fit of cognitive technologies within them. These tools can answer customer and supplier questions, perform digital or physical tasks, and make data-intensive decisions in real time. But if delicate or subtle communications are needed, you’ll want to consider a human. In insurance underwriting, for example, most policy decision-making is now done by machines, but many insurance firms still use underwriters for communicating policy decisions to agents and customers.
Once you’ve identified some areas of opportunity for smart machines, it’s important to focus on the specific technologies that can perform those tasks. There are many different cognitive technologies to choose among—IBM’s Watson alone now has almost 30 different cognitive programs, and there are many other vendors with more. Each has different capabilities. In answering customer questions, for example, if the answer is probabilistic—where there isn’t a single right answer to a question–that requires a different technology than a deterministic solution.
Perhaps the most important leadership issue is preparing your employees for roles in which they augment smart machines, and vice-versa. There will be new jobs involving implementation and oversight of these technologies—getting them installed, monitoring their daily performance, and improving them over time. Employees with some aptitude need to be groomed for such roles.
Those who are not inclined or qualified to work closely with cognitive technologies need to be told what the options are. If, for example, a financial services firm is moving to “robo-advisors” for investment advice, is there a “behavioral finance” role that can augment it? A machine can identify an optimal investment portfolio for a given customer, but it’s likely that only a human can persuade that customer not to sell when markets are crashing.
There will also be senior management roles that need to be defined. Some companies have already identified Directors of Automation, for example. And if key decisions are being made by smart machines, senior leaders need to focus on how their logic and algorithms relate to the broader world. Is something changing in regulation, markets, or customer environments that would necessitate changing the way machines make decisions? Without this role, automated decisions can rapidly go astray. Cognitive technologies offer massive potential benefits, but there is no faster way to lose money than with a poorly-performing machine making automated decisions.