Job inequality and artificial intelligence in the enterprise

The future has fewer jobs, but those that remain will be better.

Much has been made of A.I.'s potential to destroy jobs by automating away routine and bureaucratic tasks. But when I think about what makes a job fulfilling, regardless of industry, a few things come to mind: consequential decisions, meaningful personal interactions and challenging problems to solve. It's not fun to do nothing or nothing important. So while we worry about A.I. reducing the job supply, it's also true that those that remain will be better. A.I. is a powerful engine of job inequality.

This is not theoretical; it's already happening.

The best example of this I've seen is at a major medical benefits management company here in the U.S. The company sits between payers and providers, determining whether procedures are medically justified and covered by a given insurance policy. To do so it employs a battalion of nurses and doctors that evaluate each preapproval claim (for a claim to be denied, an M.D. must review it). Some cases involve significant reasoning and follow-up with the provider for more details. Others are clear-cut.

This claims processing group is a major cost center for the company, so we worked with them to deploy A.I. technology that would automatically classify claims as either clear approvals or in need of review. For clear approvals, the system immediately sends the approval back to the provider without human intervention. For those claims in need of review, the system forwards them into the human-based process as before.

The business impact of this project is substantial, saving the company millions of dollars by allowing it to increase headcount more slowly as its claims volume grows. The project has prevented jobs from existing. But it has also made those that do exist more ideal. Instead of having to wade through a series of routine claims before arriving at a claim that requires some judgment and communication with the provider, now most claims have some problem to identify and solve.

Taken more broadly, A.I. is grounded in the day-to-day work of an enterprise's people. How can we allocate their time and attention to those tasks that are meaningful and difficult? It's a transition from pull to push: Instead of employees using enterprise applications to pull data via searches, reports or dashboards, applications enhanced by A.I. push tasks and alerts to the people that need to know, or who have relevant expertise.

So imagine a job where you are constantly fed interesting, challenging problems in subject areas where you're an expert. Or where you get to focus on interpersonal communication rather than routine paperwork. These fulfilling jobs will become more prevalent with A.I., and the need for a solution to the resulting inequality will become more pressing.

Copyright © 2016 IDG Communications, Inc.