Benefits Guidance: The Problem Nobody Is Talking About

*Originally published in the June edition of the Voluntary Benefits Voice.*

By Meg Collins & Ben Yomtoob

Benefits programs exist to protect the health, financial security, and well-being of workers. That is the foundational premise of everything this industry does. Yet the data tells a troubling story about how well we are actually delivering on that promise.

Businessolver's 2024 Benefits Insights Report -- drawn from data across 4.4 million employees -- finds that an average of 85% of employees struggle to understand their benefits [1]. LIMRA's 2025 BEAT Study confirms the gap: only 57% of employees say they understand their insurance benefits very or extremely well, meaning nearly 4 in 10 workers enter enrollment without a confident grasp of the coverage they are selecting [2]. This is clearly not a marginal failure. It represents a failure of the entire system. Employees default to last year's elections without reassessing their circumstances. They skip voluntary benefits they would genuinely value if they understood them. They find themselves financially exposed when a health crisis hits because they did not enroll in the supplemental coverage that would have protected them.

The industry's response has been to deploy technology -- a wave of guidance tools, AI-powered platforms, and decision support engines. But the proliferation of tools has not solved the problem. In many cases, it has made it harder to navigate.

Defining the Guidance Framework

Before evaluating the vendor landscape, it is worth establishing a framework for thinking about what benefits guidance actually does -- because the category names vendors use tell you very little about when and how a tool serves employees. As a starting point, via input from multiple stakeholders, we propose the following framework for Benefits Guidance.

We call this the Educate / Decide / Use framework (EDU). It effectively organizes the Benefits Guidance challenge into its three constituent events.

The value of this framework is not in ranking tools; it's in exposing gaps. For brokers advising employers and carriers evaluating distribution, that question has real consequences: which moments in the employee journey are covered and which are being left to chance? Beyond online or paper Benefits Guides, most employers have tools that cover only one moment -- typically Decide. The Educate and Use moments are the ones most often absent or outdated. For voluntary benefits carriers, the Use moment is the most critical for utilization and persistency -- and it is the most underserved.

A Market That Has Never Defined Itself Holistically

Walk into any benefits technology conference today and you will find no shortage of vendors claiming to solve the guidance challenge. Decision support tools. Benefits assistants. Recommendation engines. AI-powered guidance platforms. Care navigation solutions. The category names multiply faster than the independent evidence that any of them really add value.

This proliferation is a sign of genuine innovation, but the market has outpaced the infrastructure needed to evaluate it. The landscape spans at least six distinct categories: decision support tools, communication and education platforms, care navigation and advocacy, benefits administration platforms with embedded guidance, HCM/payroll platforms with benefits modules, and emerging specialized tools.

Most vendors now claim to serve at least two events of the Educate/Decide/Use framework. But most do not achieve this -- at least not equally well. Furthermore, not all tools operate from a neutral position. Some are carrier-connected, platform-embedded, or vendor-connected in ways that shape what employees see -- and those relationships are rarely disclosed proactively.

"AI-powered guidance" is used to describe tools with fundamentally different architectures -- rule-based calculators, algorithmic engines, and large language models -- and buyers often cannot tell the difference. The terminology problem is real.

AI Is in the Field. The Governance Is Not.

This is the most urgent dimension of the guidance problem -- and it contains a contradiction the industry has not yet fully reckoned with.

MetLife's 2026 Employee Benefit Trends Study found that 61% of employees are concerned about the ethical and safety risks of AI, including bias, misinformation, and lack of accountability. At the same time, 67% of employers say AI is creating new points of friction or mistrust between employees and management [3]. Yet a new WTW survey finds that while only 20% of employers actively use AI in benefits today, 72% plan to embed it within the next two years -- driven by the need to improve communication (68%) and deliver personalized employee support (57%). The ambition is clear. The readiness is not: 71% of benefit teams lack the internal AI resources needed, 70% cite data privacy and security as top barriers, and only 1% currently have a fully developed AI governance framework for benefits [4]. The tools are coming. The governance is not.

Into that gap steps the Shadow Agent. Employees are not waiting for their employer to provide a better tool -- they are already using one, just not one managed by their employer. Recent data shows that 52% of U.S. adults now use large language models like ChatGPT, and nearly two-thirds of Americans have used AI for financial advice [5]. In the benefits context, that means employees are copying their Summary of Benefits and Coverage into ChatGPT, asking Claude to explain their HDHP, and using Gemini to compare premiums. The Shadow Agent fills the void the official guidance stack left open -- and it does so with absolute confidence and zero accountability.

The problem is not that employees are using AI. It is that they are using AI that has no knowledge of their specific plan design, no understanding of voluntary benefit eligibility rules, and no ownership for the guidance it provides. A public model trained on the internet -- not your plan documents -- will confidently answer a question about fertility coverage, HSA eligibility, or supplemental life insurance with information that sounds authoritative and may be completely wrong. The employee acts on it. The consequences are real.

The sanctioned tools being deployed in the market carry their own risks. A rule-based calculator may not handle the underwriting nuances that make voluntary benefits complex. An algorithmic engine trained primarily on core medical data may systematically underweight voluntary products. A generative AI system may produce confident-sounding guidance that is factually incorrect about specific plan designs. But the Shadow Agent carries all of those risks -- just without the oversight.

The era of "benefits guidance is a nice-to-have" is over. In a world where the alternative to a company-sponsored tool is a consumer tool not designed for the job, the absence of a governed guidance strategy is itself a decision -- and not a defensible one.

The Baseline the Market Has Not Yet Set

Across our work developing an evaluation framework for this market, a consistent picture has emerged of what separates genuinely effective guidance from tools that are guidance in name only. Effective guidance requires:

• Year-round presence: Not just at open enrollment, but at every guidance moment -- life events, claims questions, mid-year decisions.
• Voluntary benefits depth: Coverage of the full benefits portfolio, not just core medical, dental, and vision.
• Transparency of methodology: The ability to explain why a recommendation was made, and disclosure of any commercial relationships that may shape what employees see.
• Validated outcomes: Demonstrated evidence -- independently reviewed -- proof that those employees who used the tool made better decisions than those who did not.

Rather than being aspirational standards, these measures should be a baseline. And right now, the market does not have a shared framework for measuring whether any given tool meets them.

Building the Infrastructure the Market Needs

The benefits guidance market has the tools it needs. What it lacks is the infrastructure to evaluate them -- shared terminology, independent evaluation frameworks, and validation methodologies that give buyers, brokers, and carriers a consistent basis for distinguishing between tools that work and tools that are well-marketed.

Building that infrastructure is the work of the Benefits Guidance Consortium, an independent coalition of vendors, brokers, employers, carriers, and investors working to establish a shared standard for benefits guidance grounded in three principles: Clarity, Transparency, and Validation.

Later this year, the Consortium will publish its inaugural State of the Market white paper -- the first independent, methodology-driven evaluation of the benefits guidance landscape. It will establish a shared taxonomy and assess tools across the dimensions outlined in this article.

This article is the first in a series. Future installments will examine individual categories in depth and report on the Consortium's findings as the white paper takes shape. The tools recommending -- or failing to recommend -- your products are already in the field. The standard for how they should perform is still being written. The brokers and carriers who engage in shaping that standard now will be the ones best positioned for what comes next.

Learn more at benefitsguidanceconsortium.org.

---

*About the Authors: Meg Collins and Ben Yomtoob co-founded the Benefits Guidance Consortium to define and promote the Benefits Guidance category. Ben is Founder / Lead Consultant at BuckleyRoberts and a former CEO of Workterra. Meg is Founder and CEO of RevGem Consulting and a former Chief Growth Officer at Ideon.*

Sources

[1] Businessolver | 2024 Benefits Insights Report -- businessolver.com

[2] LIMRA | 2025 BEAT Study: Benefits and Employee Attitude Tracker -- limra.com, June 2025

[3] MetLife | 2026 Employee Benefit Trends Study -- metlife.com, March 2026

[4] WTW | 2026 AI Use in Health and Benefits Survey -- wtwco.com, May 19, 2026

[5] Elon University, Imagining the Digital Future Center | "Close Encounters of the AI Kind" -- imaginingthedigitalfuture.org, March 12, 2025; Intuit Credit Karma | Generative AI and Financial Advice Survey -- creditkarma.com, October 2025