Artificial intelligence is reshaping the US employee benefits market faster than most employers were prepared for and not all of them are deploying it responsibly.
As the 2026/2027 open enrollment season gets underway, Houston, Texas based Robby White, US benefits line leader at Arthur J. Gallagher & Co., says the employer conversation has fundamentally shifted from whether to use AI to whether plan sponsors are building the governance, trust, and human oversight needed to make it work for the employees who depend on it.
Employer-sponsored subscriptions to AI tools surged from 16 percent in 2025 to 33 percent in 2026, the largest single-year increase recorded in the Society for Human Resource Management's (SHRM) 2026 Employee Benefits Survey. But only 20 percent of employers report actively operationalizing AI in benefits today, while 72 percent plan to do so within the next two years, according to the WTW 2026 AI Use in Health and Benefits Survey of 312 employers covering approximately 4.6 million US employees.
White says employers heading into the 2026/2027 season are trying to solve three interconnected problems simultaneously: rising healthcare and pharmacy costs, an overloaded HR function, and employees who do not understand or fully use the benefits already available to them.
"The conversation has moved beyond using AI simply to automate administrative work," he said. "Employees are weighing healthcare costs, prescription drug exposure, provider access, mental health resources, voluntary benefits and financial wellbeing support, often without feeling confident that they understand the tradeoffs. AI can help translate that complexity into more personalized and actionable guidance based on factors such as life stage, family circumstances, affordability concerns and prior utilization."
For employers, White says the opportunity is equally significant. AI can reduce the volume of routine questions HR teams must manage, strengthen communications, and provide better insight into utilization, care patterns, pharmacy trends, and employee engagement. He says the most meaningful shift for the 2026/2027 season is toward year-round support - moving AI out of the brief annual enrollment window and into an always-on model.
"Employees should have 24/7 access to the right resources and information at the moment they need them, with a clear path to human support," White said. "That includes benefit counselors who are available virtually or by phone when an employee wants help interpreting options, validating guidance or working through a more complex decision."
White draws a distinction that is easy to overlook in conversations dominated by technology: most employers do not have a benefits-access problem. They have an understanding and utilization problem.
"Many employers do not have a benefits-access problem; they have an understanding and utilization problem," he said. "Employers may offer valuable programs, but employees do not always know what is available, when to use it or how different resources fit together. A strong broker partner should help the employer identify those gaps, simplify the experience and evaluate whether an AI-enabled solution is increasing understanding and utilization rather than simply generating logins."
Benefits consultants, he says, have a specific and irreplaceable role here - gathering and analyzing data to identify barriers, evaluating vendors for potential conflicts, and running an in-depth third-party vendor (TPV) review process before any new tool enters the benefits ecosystem.
That review should cover data practices, security, operational readiness, integration capabilities, and potential conflicts of interest across the vendor landscape.
There is also a trust dimension that White says too many employers underestimate. "There's also an important distinction between deploying AI and earning employees' trust in it," White said. "Employers may be moving quickly to operationalize the technology, but employees will not automatically use an AI-enabled tool simply because it is available. Employees need to understand and trust how AI is being used in the benefits process."
He points to the deeply personal nature of healthcare and benefits decisions as the reason that trust must be built deliberately. Employees want to know what information is being used, who can access it, how recommendations are generated, and whether AI is advising them or making decisions on their behalf. They may also vary significantly in digital comfort, health literacy, and confidence navigating benefits systems.
For more on how AI-driven enrollment tools are reshaping the broker's role in guiding employer clients, IBB has tracked the market's evolution in its coverage of The Hartford's 2026 Future of Benefits Study.
White is careful to frame Gallagher's approach to AI in terms of outcomes, not features. Employers, he says, are not asking about AI because they want new technology. They are asking how they can control costs, reduce administrative burdens, and help employees get more value from the benefits they already provide.
"For HR leaders, the value is a clearer view of how the benefits portfolio is working as a whole," White said. "AI can support benefits communications, workforce analytics and post-enrollment analysis while helping employers identify gaps in employee understanding and evaluate how programs interact across a fragmented vendor ecosystem. Over time, that insight can help HR leaders make more informed decisions about program design, communications and investment."
For employees, he says the most meaningful change is that benefits decisions can become more personalized and less overwhelming. Rather than presenting every employee with the same catalog of plans and programs, AI-enabled guidance can surface options more relevant to an individual's circumstances - explaining tradeoffs involving cost, coverage, prescription drugs, provider access, and other available resources.
"AI can help employees compare plans, estimate costs, understand coverage differences and identify high-quality or lower-cost care alternatives," White said. "It can also connect them with resources such as mental health support, chronic-condition programs, voluntary benefits and financial wellbeing tools." The broader objective, he adds, is to improve how existing benefits are understood and used, reduce unnecessary utilization and waste, and direct employees toward higher-value resources.
White is direct about where employers are getting it wrong, and he identifies two failures he sees most often.
The first is adopting AI because it is marketed as AI - without determining whether it solves a meaningful problem. "A disconnected tool can add complexity to an already fragmented benefits ecosystem rather than simplifying it," White said. "Poor-quality or fragmented data can also weaken recommendations and quickly undermine employee trust." The second is waiting until enrollment begins to explain the technology to employees.
"Employers should test systems well in advance, clearly communicate their privacy practices, prepare HR teams and vendors with consistent messaging, and give employees a straightforward explanation of what AI will and will not do," he said.
Responsible deployment, in White's view, starts with strong data governance, defined privacy protections, transparency, and clear accountability - before any tool goes live.
Employers need to understand what information an AI-enabled tool uses, how that information is collected and stored, who is authorized to access it, and how recommendations are generated. Employees should have meaningful control over their own information, with confidence that it is being used only for its intended purpose.
He says consultant-led TPV reviews are essential to disciplined vendor oversight. Because benefits vendors may handle sensitive health, financial, and personal information, those reviews should cover security and compliance practices, data ownership and permitted uses, breach-response responsibilities, and how data will be returned or securely destroyed when a contract ends. The Department of Labor's (DOL) Employee Benefits Security Administration provides the governing fiduciary framework under which employer-sponsored plans evaluate vendor relationships and data stewardship obligations.
From a compliance and fiduciary standpoint, White says employers should treat AI as part of the broader benefits governance framework - not as a stand-alone technology purchase. "Tools should be evaluated for data quality, accuracy, integration, employee outcomes and their ability to support appropriate oversight," he said.
His message to brokers is unambiguous: the technology is only as effective as the governance and human expertise surrounding it. Employers who build that foundation - with the help of a broker partner willing to do the harder work of vendor oversight, change management, and trust-building - will be better positioned to demonstrate real value from their AI investments heading into 2027.
