How One Company Embraced Automation, Upskilled Employees, and Laid Off No One

Mention the words “automation,” “machine learning,” or “AI” to employees and their minds often turn to fear. They’re afraid that if companies automate part of their tasks, they’ll lose their jobs. They’re afraid their skills will become obsolete. 

But one Fortune 500 company — The Hartford, a financial services and insurance organization — recently automated part of its workers’ compensation business and the opposite happened. Employees received loads of upskilling, got a chance to do more complex work, and everyone kept their jobs. Instead of reducing staff, The Hartford increased it. 

“Because we were changing their jobs and upskilling them to handle more complex things,” says Melissa Pouliot, an assistant director of learning at The Hartford, “there wasn’t a ‘We’re taking this work away from you and not replacing it with something else,’ and therefore we have to lose personnel. We were replacing it with something else.” 

Let’s take a look at how The Hartford upskilled its employees during the automation process and allowed them all to keep their jobs. 

The Hartford aimed to automate the simplest tasks

As a leader in workers’ compensation and group disability insurance, The Hartford processes tens of thousands of new claims every year. About 80% of these workers’ compensation claims are what’s called “medical only,” which means that an injured worker only needs to be compensated for medications or medical care and not more complex issues such as lost wages or time off work. This might include, for example, a restaurant worker who cut themselves on a glass salt shaker, needed stitches, and was able to return to work the next day. 

In 2020, The Hartford had a team of 20 claim administrators responsible for handling “medical-only” claims, ensuring that transactions were completed and payments made. A review of the department that year, however, found that the team could become much more efficient if it automated its simpler claims. 

Automation meant that the claim administrators’ role would change. They’d be relieved of repetitive, administrative tasks, freeing them to handle more complex claims, which had previously been the responsibility of a higher-paid role. (And now, employees in that higher-paid role could manage claims that required an even higher degree of human touch.) 

Most of the claim administrators had dealt with complex claims in previous jobs and had many of the skills they’d need. But Melissa and Mark Wagner, The Hartford’s vice president of learning, quality, and leadership process, knew they’d still need considerable upskilling for the transition to be a success.

The company allayed fears by clearly communicating how roles would change

Mark and Melissa and the workers’ comp team leaders worked with The Hartford’s data scientists to determine how best to automate these simpler claims using machine learning. 

Throughout the process, they relied heavily on a book called Reinventing Jobs: A 4-Step Approach for Applying Automation to Work, by Ravin Jesuthasan and John Boudreau, which urges leaders to use a combination of automation and human workers to improve performance. In particular, they leaned into the authors’ guidance on deconstructing roles. From there, The Hartford’s data scientists were able to decide what could be automated and what could not. 

Team leaders helped break down roles to each particular task — and as they did, they paid close attention to what skills employees would need as their roles evolved. They then communicated this to employees, clearly and often. “It was fascinating to see how the leaders felt armed to communicate the change much better to people as a result of deconstructing roles,” Mark explains, “and not have that fear element attached to it.” 

By communicating often, leaders were able to remove a lot of the uncertainty that can accompany automation.

The company upskilled employees to handle more complex tasks 

The new technology was set up so that it would automatically detect “medical-only” claims, process, and pay them. Unless a customer contacted the department with a complaint or the status of the claim changed (such as a worker beginning to miss work due to an injury), administrators wouldn’t even see the claims.

That freed employees to handle claims that required a human touch — but to do the job well, they’d need additional skills. So Melissa and her team created an upskilling plan, which included identifying skills gaps. She realized that claim administrators would need additional skills for communicating with customers and interpreting data or insights. 

Because Melissa had already trained higher-level claims staff, she had much of the learning content she needed. She reused some of that existing content and created additional content. She and her team then offered online, instructor-led training, role-playing (in which they walked through claims scenarios), drills, and sessions on medical terminology and interpreting medical reports. 

Her team worked alongside Mark’s design teams and threw a lot of energy into updating the workers’ compensation performance support site — essentially a knowledge management system — because employees would now be handling claims in multiple states, each of which had different regulations. For example, they might need information for a case in which a construction worker was injured in the parking lot of a job site. In some states, the worker would be considered “on the job” only after they had clocked in and were on the floor of the worksite. In that scenario, the worker would not be eligible for compensation. But in other states, the parking lot would be covered — as would the worker. With a performance support site, claim administrators could access this information quickly. 

The Hartford’s team “turned the pipes on” slowly 

One of the smartest things that The Hartford did was roll out the new technology slowly, to work out kinks and give employees time to adjust. Between September of 2020 and June of 2021, they conducted a number of “test and learns.” 

“We didn’t turn the pipes on all at once,” Melissa says. “We tried a few states at a time, and then got feedback of what was working and what was not.” 

This gave time for employees to make sure the technology was working right and get up to speed with their own skills. It was an approach that helped keep employees at the center of the process. 

Final thoughts: Even as they created more efficiency, The Hartford added staff 

As a result of the automation, the company has been able to handle claims more efficiently than before. But most importantly, from a human perspective, employees were able to gain valuable new skills and keep their jobs. As the number of complex cases they manage has grown, they’ve even added to their numbers. Since 2020, the team has grown from 20 to 23 employees. 

“When we started, I had that feeling of ‘Gosh, this is scary, because it means jobs,’” Mark says. “But once we started getting into it and we saw that it enhances jobs to some extent, I realized, ‘Wow, this made things better.’”

Illustration by Visual Generation on iStock.