Ask most any automation expert or CFO who’s implementing automation that you like, and you’ll get the same story: no, automation isn’t going to take finance jobs away. Instead, finance workers will have time freed up for higher-level, strategic, value-add work.

But does that actually pass the smell test?

Lee Coulter

One expert, Lee Coulter, speaking at the recent CFO-hosted CFO Live conference in New York, sprinkled some doubt on the dreamy forward-looking view of finance labor.

“There are near-term use cases where we’re going to see significant levels of automation being used in finance,” said Coulter, CEO of transformAI and chair of an IEEE working group on standards in intelligent process automation. In fact, he claimed, “70% of what finance does today can be automated. Some jobs are just going to go away.”

He continued, “The risk management function will be dramatically reduced, because compliance will be managed by automation. The fraud department will be meaningfully non-existent in the future. Credit analysts will no longer exist.” And, “If you know somebody in underwriting, send them back to school.”

These forecasts represent a five-year view, Coulter specified.

Other roles, such as many in accounts payable, accounts receivable, and payroll, may survive massive automation deployment, at least in the immediate future.

“It may take over 9% of their job, or 3%, or 6%,” said Anjan Roy, CFO of global finance for General Electric, who participated in a panel on robotic process automation that preceded Coulter’s session.

But such workers will still need to be upskilled to take on new value-added duties enabled by their freed-up time.

In fact, retraining likely will emerge as an increasing theme. PricewaterhouseCoopers, for instance, recently announced plans to invest $3 billion to $4 billion in employee training over the next three to four years. Other big employers, like Amazon, have announced similar commitments.

Such companies “have a lot of people whose jobs are going to go away” because of automation, in Coulter’s view.

He suggested that companies should be clear with employees about their plans for automation. “I encourage you to be very explicit about your intentions. And add supporting facts that people can point to that say, ‘They’re serious about this.’ If you don’t, it’s going to be really hard to get people to show up for meetings and be active participants in the work.”

Meanwhile, Coulter exploded some misconceptions about intelligent automation.

AI is easy. “No, it’s actually really hard,” he said. “I recommend that you start with simple task automation so that you begin to learn more about the data dependencies for machine learning and augmented intelligence.”

Even RPA, which is much less complex than AI, usually requires a measured approach. “I’ve seen organizations go in and deploy a thousand bots in a year. It’s generally disastrous. Humans are not ready to change at that pace. It’s very risky to let this go viral.”

We can wait. “No, you can’t wait. It’s good that you’re here because automation has been part of almost every conversation I’ve had or listened to at the conference. Be on the front end of this wave or be materially disrupted. And I don’t mean that as a parable.”

Someone else will own this. “No, they won’t. Intelligent automation is a business-led project. Don’t let IT try to own it — that’s not a formula for success. Would you hire IT to input payroll to general ledger entries or do cash applications? No. This is digital labor.”

It’s a hyped fad that’s going to fade away. “No. This is very, very real,” Coulter said, solemnly.