Tackling finance transformation at Kelly Services

When Troy Anderson assumed the CFO chair at $4.3 billion global staffing company Kelly Services in fall 2024, he was given a hefty mandate: Transform a 400-person finance operation with teams in multiple countries from a highly fragmented function to one with integrated technology and processes.

And as if that weren’t a heavy enough lift, Anderson had to do so against the backdrop of changes to the CEO and board, while preparing for a new ERP system launch.

Leading through change is a thread that’s run throughout much of Anderson’s career. At $4.5 billion business process services company Conduent, a Xerox spinoff, for instance, he implemented a major finance transformation project. As SVP and CFO for the public sector industry at Xerox, he had to make the difficult choice to cut an unprofitable segment.

Anderson spoke with CFO Brew at the National Association of Corporate Directors’ annual summit to talk about what it takes to successfully transform.

A fragmented function. Anderson inherited a finance function with many overlapping pieces and unclear reporting lines. Redundant work was a problem. “We had common work distributed across many different teams,” he recalled. “We had finance people doing accounting work, and we had accounting people doing finance work.” The international team, he said, reported into finance, accounting, and operations. Plus, teams from previous acquisitions “had not been integrated,” so the company had “multiple accounting teams, multiple back office teams.”

Anderson set about realigninged “the organizational model: How do we work, where do we work? What processes can we change?” Technological upgrades were key to the process. For instance, the billing and payroll functions had “bad data coming in, which creates manual workarounds,” Anderson said. The company recently adopted a single human capital management platform. “Now all of our payroll for our employees is done on one platform, and that allowed us to bring together all the payroll teams” and adopt “a core set of processes,” he said. “We’re still bringing those teams together.”

Building trust. At the same time Anderson was transforming the finance function, Kelly was seeing a shakeup in its long-established leadership. Chris Layden, who spent nearly 20 years at ManpowerGroup, became CEO in September 2025, replacing a 22-year company veteran, Peter Quigley.

“Kelly has a very close culture,” Anderson said. “There are a number of very tenured people in the organization, and they’ve built relationships over many, many years.”

So it took time for Anderson to build trust, he said. He focused on transparency, and acknowledging the stress people were under, he said. It helps to “communicate that very clearly: ‘Hey, we hear you. We understand there’s a lot of pressure,’” he said. He also encouraged personnel “to identify the challenges” they faced “versus me just coming in saying ‘This is wrong, that’s wrong,’” he said.

And Anderson also found it helpful to emphasize to staff how the transformation would improve their jobs: by resolving pain points they had brought up, by making their jobs easier and allowing them to work fewer hours, or by giving them more opportunities for career progression “because we’ll now have more defined roles and structures and better portability across the organization,” he said.

Readiness precedes AI. Another key decision Anderson made was not to implement AI too early. “There were a lot more foundational elements of the transformation that we needed to effect first,” he said, such as fixing “fragmented processes” and improving data quality.

He also didn’t want staff to have another change to adapt to. “If we led with AI, I felt we were going to lose,” he said. “There was already so much change management to bring to the organization…Layering AI on top of that would have been difficult.”

About six months into the transformation, when they reached a point where staff were engaged and understood why the changes were being made, Anderson thought, “Okay, there’s a role now for AI to start playing here.”

Leadership identified early adopters of AI whose projects they wanted to scale. “We built a framework to bring the organization along with that as well,” Anderson said. They held team meetings and lunch-and-learns, and encouraged staff to experiment with AI.

Now, he said, the goal is to expand individuals’ AI successes to teams, and teams’ wins to functions. For instance, “If one finance team is using it to do forecasting, how do we get all finance teams to leverage that?”

Though the company hasn’t solved all its issues with data and seeing how AI might benefit the whole organization holistically, “we’re much further along,” Anderson said. Early on, some staff told him they didn’t have time to focus on AI. “And now they’re the same people who say ‘I can’t do my job without it,’” he observed.