Kelley Kage (pictured) does not frame Employers’ technology strategy as a race to deploy the latest tools. Instead, the insurer’s chief information officer positions the company’s AI journey as a measured effort to strengthen the decisions at the core of workers’ compensation.
“We were in a very good spot,” said Kage, senior vice-president and chief information officer at Employers. “We had started to mature through our AI transformation, and we were focused on embedding AI across multiple operational areas.”
That focus extended across underwriting, claims management, finance, IT and marketing. The objective, Kage said, was not innovation for its own sake. It was about removing friction from core workflows, improving consistency in risk assessment, and enabling the organization to scale without losing clarity or control.
Employers’ approach to innovation began with specificity. Rather than pursuing broad digital transformation programs, the company targeted concrete problems that slowed execution or introduced variability into decision-making.
“When we thought about transformation, we were targeting specific pain points,” Kage said. “That included reducing manual processes associated with our premium audits, improving consistency in our risk assessment process, and helping new employees get up to speed faster.”
As a specialist workers’ compensation carrier, Employers placed particular weight on underwriting and claims outcomes. “Those decisions were absolutely critical for us to get right,” she said. “That’s what mattered to our customers and to injured workers.”
In claims, the insurer combined data, automation and AI to streamline intake and triage while surfacing early risk indicators. The intent was not to displace experienced professionals, but to refocus their time.
“It was not about replacing adjusters,” Kage said. “It was about helping them concentrate on the judgment-heavy areas where they needed to make decisions and where they could add the most skill and value.”
The resulting benefits aligned with familiar industry priorities – cycle time reduction, severity control and improved outcomes.
Employers did not attempt to build everything internally. Strategic partnerships played a central role, but they were selected through a deliberately narrow framework.
“We used a simple framework to decide who we worked with,” Kage said. “First, we asked whether the partnership solved a real operational problem that we had today.”
From there, the insurer evaluated how cleanly a solution integrated into existing workflows and data environments, whether it could scale alongside the organization, and whether it avoided locking Employers into rigid or brittle dependencies.
Equally important was domain understanding. “We looked for partnerships with people and companies that understood insurance beyond a narrow technology problem,” she said. “We wanted partners who leaned in and helped us think about challenges from a risk and operational business perspective.”
That often meant co-designing solutions, piloting quickly and defining clear metrics for success. Employers was explicit about what it avoided. “We were not focused on pretty slides and broad recommendations,” Kage said. “We were getting into the weeds on very specific decisions and impacts.”
When AI was involved, explainability and governance were non-negotiable. “We also avoided big black-box solutions,” Kage explained. “Transparency, governance, and explainability of how these systems worked were critical, particularly around underwriting and claims.”
The most difficult part of technology integration, Kage argues, is rarely the technology itself.
“The biggest challenges in this area were change management, integration complexity, and data quality,” she said.
Underwriters, auditors and adjusters operate under regulatory scrutiny and time pressure. Any additional friction risks undermining adoption. “If we implemented something new and added friction at any point in that process, it felt like extra work,” Kage said. “They would not adopt those changes.”
Employers addressed this by embedding new capabilities directly into existing systems and designing with frontline users in mind. “We wanted people to see technology and AI as a partner and an assistant,” Kage said, “that helped them do their jobs better, improve quality, and move faster.”
Data presents a parallel challenge, particularly for a 113-year-old insurer. Rather than attempting a wholesale clean-up before innovating, the company advanced in stages. “We did not try to boil the ocean.....we moved in parallel, advancing our technology and improving data quality as we went.”
Trust, she adds, depended on explainability and human oversight. Teams need to understand not just what a system produces, but why it does so and how it supports better outcomes for customers.
For Kage, the most significant shift underway is organizational rather than technical. In 2026, Employers plans to equip its entire workforce with AI capabilities designed to improve day-to-day work.
“This is genuinely a cross-functional leadership effort,” she said. “Our entire executive team is leaning in together.”
Frontline leaders are trained alongside executives, reinforcing the message that AI is not an IT initiative owned in isolation. “At every level of the organization, we are saying: we are in this together,” Kage said.
Her role as CIO is to provide clarity and guardrails – aligning innovation with strategy, risk appetite and operational realities – while creating space for experimentation. That balance, she said, accelerates momentum and creates new internal partnerships.
“We are not treating it as a technology initiative,” she said. “It is about business decisions with technical dimensions that make them real.”
By embedding AI capabilities at the front lines and pairing them with strong governance, Employers signals that sustainable innovation comes from empowering people, not replacing them.
