As part of its operating philosophy, Siemens, the German conglomerate operating in more than 200 countries and regions, applies a yearly target range of 3% to 5% of productivity growth to each of its subsidiaries, to go along with certain cost objectives. Not unexpectedly, Siemens Brazil has had a tough time hitting those targets over the last three years or so.

A wholly owned subsidiary that takes in about $1.5 billion a year in sales, the São Paulo-based company has faced  “a very particular situation here because of the economics and political crisis in Brazil, mixed with a very high inflation,” Siemens Brazil’s finance chief, Martin Kerkhoff, tells CFO.

Although Brazil has shown signs lately of  emerging from the biggest economic morass the country has ever known, it’s still in the grips of the consequences of the aftermath of a massive government corruption scandal. The nation’s inflation rate hit a 12-year high of 10.67% in 2015, and its unemployment rate hit a five-year high of 13.7% just this March.

All these factors have made growth a hard thing to come by for any company in the country, including Siemens Brazil. The subsidiary sells such things as turbines and other power-generation equipment; factory automation systems; and health-care X-ray and MRI machines to large Brazilian organizations. Peril to its customers, of course, imperils the company.  

Lacking substantial growth, Siemens Brazil’s targeted productivity gains had to stem from some other source. “If we don’t achieve productivity through growth, we have to work on the other side of the equation, which is basically adjusting the cost levels,” Kerkhoff says.

One way the subsidiary has cut its costs has been to lay off 15% to 20% of its employees — with the cuts stressing administrative rather than blue collar workers — over the last two years, according to the CFO.

As a German, Kerkhoff refrains from commenting on Brazil’s political problems. Asked how he and other executives can assess political risk in the country, he replies that the “political crisis here is such that nobody knows where the country will be one year from now.”

Nevertheless, a company spokesperson said, Siemens has no intention of exiting Brazil, where it’s been operating for more than a century. Although the subsidiary has been affected by the crisis, “we keep trusting in Brazil’s political system,” the spokeswoman said, and will continue to invest in the country.

But amid such turmoil, Kerkhoff says the only way his company can manage political risk is to implement what it refers to as its “master plan project.”

The project consists of  “addressing our main internal administrative processes, like purchase-to-pay and order-to cash, and working on automating these processes to continue adding higher levels of productivity,” he adds.

A key goal is to eliminate steps in each of the processes, says Kerkhoff, who reports to the finance chief of the organization’s American operations who, in turn, reports to the global parent company’s CFO in Germany. And at the center of these efforts is something he calls robotic process automation (RPA).

“I’m not a techie. But, to put it in my words, it’s just simple algorithms that are applied to routine work. Basically it’s a rule-based automation of routine work that is currently done by human beings,” he says, attempting to define RPA.

But it’s robots that actually do the work. “These RPA tools can read structured digital input coming from a pdf file or an email from a workflow [document] or a database from SAP, and then interact with humans,” the CFO says.

Provided by Blue Prism, a U.K. software company specializing in RPA, the system can run a program, extract information from a tool, copy information from one tool to the other, autonomously send or receive emails, and process information received within emails, he notes. Launched in 2016, RPA was the start of an effort to digitalize the company.

Kerkhoff is quick to caution that RPA isn’t nearly as artificial as rocket science. “You have to be careful. This is not [artificial intelligence]. It’s not an IBM Watson, for example. This is just a simple tool that can work on routine work and do what it’s been told to do beforehand,” he says. “It replaces human beings on a very transactional basis.”

Still, wasn’t this a heavy cost to assume in the teeth of a national financial crisis?

“Actually, I thought the same,” Kerkhoff recalls. But when his team showed him a video of how RPA worked, his first reaction was that the firm needed to invest in the technology — “which I was actually willing to do, by the way,” he says.

“But in the end the investment is very low because Siemens Brazil pays moderate fees on the basis of how much the company uses the RPA,” he says. In the last few months, the company created an algorithm enabling the robots to work in any specific area that it defines.

The programming work is done by an internal team. “And these guys are real techies, a team of young guys coming from a variety of different areas in our company, people with finance and technical backgrounds,” he says, noting that the company has retrained the team to be able to program the robots.

“It is not that difficult to re-skill somebody to program these robots. It might sound very high level and difficult to program them. But in fact it’s just a basic way of designing a workflow for the robot, just like creating, for example, an Internet page,” according to the CFO. “This is something that an interested person can learn in two to three months.”

The retraining is being done by a core unit of five IT experts, two of them recently hired. One member of the team, formerly an industrial engineer with Siemens, has considerable knowledge of factory automation.

Kerkhoff gives the following example of what the robots are doing for his company.

“I want to send a transformer with a weight of two tons from our factory to a customer in the southern part of Brazil. This information is then extracted by the robot. The robot analyzes this situation. It realizes that we need to send a transformer to a certain region in the country. Then it analyzes other demands for the same day and for the same region. And then it automatically sends out an email to the sales departments of a preselected group of freight shippers,” he says.

“Our suppliers have one day to send an answer to this request for proposal coming from the robot confirming or not if they are willing to take on this task. The robot automatically compares the incoming answers of these suppliers and automatically kicks the cheapest supplier through. Once the robot has chosen the cheapest supplier, it creates the purchase order in our SAP system and sends it to the cheapest supplier,” according to Kerkhoff.

“And once the service is done, the robot also compares the incoming invoice with the purchase order created in the system. So a process that previously was highly manual, involving many phone calls back and forth among many people,” the CFO says, “is now done by one virtual robot that replaces 15 people, more or less.”