Making Health and Wellness Benefits Perform: What Modern Card Infrastructure Changes

Most finance leaders can say what their benefits program costs in premiums and contributions. Fewer can say what it gives back.

Total employer compensation costs for private industry workers averaged $46.60 per hour worked in March 2026, according to the U.S. Bureau of Labor Statistics. Benefits, including health coverage, retirement contributions, paid leave and similar programs, accounted for 30.1% of that figure, or $14.01 per hour worked. That means nearly one-third of what an organization spends on its workforce moves through benefit programs instead of paychecks. Whether that money does what it’s meant to do depends less on the size of the budget than on the infrastructure that moves it.

Benefits programs, whether employer-sponsored HSAs, government-funded assistance or wellness stipends, exist to support the people who use them. Whether they do comes down to three things the payment system either offers or doesn’t: real-time controls, automatic eligibility checks and a clear view of the money for the participant. When those are missing, it shows up in the numbers. NFP’s 2025 U.S. Benefits Trend Report notes that less than one-third of employees fully use their supplemental benefits, and 13% forget they have them at all. When all three are in place, the same budget reaches far more of the people it was meant for.

1. Where Benefits Programs Win Value Back

Every benefit dollar moves through several steps before it does its job: enrollment, allocation, spend, an eligibility check and, finally, reconciliation. At each step, a program either holds on to value or gives it away. Modern systems hold on to it at every step, and most of that gain never shows up on the premium line.

Figure 1. The premium line shows only part of what a benefits program controls, and only part of what it can win back.

Compliance certainty: enforcement at the point of purchase, not in the audit

Most benefit programs, such as HSAs, FSAs, wellness stipends and government-funded assistance, come with eligibility rules attached: This kind of purchase qualifies; this one doesn’t. What matters is when the rule gets applied. Applied at the point of purchase, it’s invisible to everyone and costs nothing to enforce. Applied later, in an audit, it arrives after the money is gone.

Spend controls tied to merchant category codes close that gap automatically, right at the point of purchase. An eligible purchase clears instantly. An ineligible one doesn’t clear at all. The value is easy to size. Take an organization with $5 million in annual HSA and FSA spend and 2% flagged in a post-transaction review. That’s roughly $100,000 a year at risk, plus the staff hours spent investigating, clawing back or writing off each flagged transaction. Real-time controls make all of it a non-issue.

Participant utilization: money that reaches the people it was set aside for

A program only pays off if the people it serves understand it. Castlight Health’s 2026 Employee Benefits Report found that when a new need comes up, 36% of employees go straight to their own solution, often paying out of pocket, instead of checking what their employer already offers. The same report shows what changes that: More than 80% of employees said they’d be more likely to stay with an employer that offers a flexible wellness allowance. It’s clarity and flexibility, not just generosity, that turn a benefits budget into loyalty.

Showing balances in real time is the most direct fix: A participant who can see the money is far more likely to use it. The prize is big. For an employer setting aside $12 million to $15 million a year, a 15% to 25% unused rate is $1.8 million to $3.75 million already paid for and waiting to be spent. Every point of use recovered is pay that reaches an employee instead of coming back as a forfeited balance.

Operational capacity: hours returned to the benefits team

The scarcest resource on most benefits teams isn’t budget. It’s time. Hours that could go toward tracking use and improving the program go to manual reconciliation and after-the-fact review instead. Automating that work hands the hours back.

Take a benefits team running 5,000 active accounts. If 5% need manual attention (an eligibility dispute, a reimbursement request or a reconciliation exception) at 15 minutes of loaded staff time each, that’s roughly 62.5 hours a month. On a real-time, automated platform, most of those cases never happen, and those 62.5 hours go back to work only the benefits team can do.

Figure 2. This is what a real-time, automated platform gives back every year.

Program agility: one platform, every benefit type

Run HSAs, FSAs, LSAs and government-funded benefits on one platform, and each new benefit type becomes a setting to change rather than a system to build. The alternative is a separate cash register for every product category in one store: more hardware to maintain, more places for something to break and no single view of the business. It also compounds. Each program added to shared infrastructure launches faster than the one before it, which is what lets a benefits strategy grow at the pace the business needs.

2. The Six Capabilities Behind a High-Performing Program

Six specific capabilities decide how well a benefits program performs. Together they change what it can promise: a real-time answer for the participant, backed by real enforcement rather than a faster version of the same paperwork.

Category-level spend controls

Spend controls tied to merchant category codes (and to specific merchants, where needed) limit funds to eligible purchases automatically, at the point of purchase. Organizations stay compliant without leaning on audits or manual checks after the fact.

Support for multiple benefit types on one platform

HSA, FSA, LSA and government-funded benefits can all run on one system instead of a separate one for each. Organizations can manage several programs at once and roll out new ones faster, without rebuilding from scratch every time.

Real-time authorization and balance management

Transactions get approved or declined in real time based on eligibility rules and available balances, and balances update instantly after each one. That means fewer errors, fewer disputes and fewer support calls for administrators and participants alike.

Centralized administration and oversight

Central reporting, monitoring and admin tools show administrators exactly how benefits are being used and how the program is doing, replacing the spreadsheets and manual tracking that older programs tend to collect. That’s what lets the numbers in the framework above (flag rates, usage rates, exception volume) be measured instead of guessed at. You can’t manage what you can’t measure.

API-first ecosystem integration

Direct API connections to eligibility, enrollment and claims systems let data move automatically between them. Nothing stops for a manual handoff, which cuts busywork, improves accuracy and gets money to people faster.

Secure, enterprise-ready platform

Network-level security, regulatory compliance and enterprise-grade controls, built specifically for healthcare and government benefit programs, let organizations run regulated benefits with confidence, without building or maintaining compliance infrastructure of their own.

Put these six together and one idea holds them up: compliance that never gets in the way. Modern benefits infrastructure lets organizations check eligibility automatically, show participants their own balances in real time and launch new programs faster, all without adding headcount. Companies including evermore and SMI use SoFi’s Galileo to modernize benefits distribution on exactly this basis.¹

3. What Makes a Modernization Land

Technology is rarely what decides whether a benefits upgrade pays off. The teams that get it right build the business case, the compliance plan and the communication plan with the same care they give the platform. Five things separate the rollouts that work from the ones that stall.

Building the business case in CFO terms

Feature lists rarely survive a budget review. Dollar figures do. Lead with the framework above, stated in dollars rather than plan counts and fill it in with the organization’s own flag rate, usage rate and exception volume instead of an industry average. A case built that way is hard to argue with, because the numbers belong to the people in the room.

Compliance and regulated benefit types

Not every benefit is regulated the same way. HSAs, FSAs and government-funded assistance each come with their own eligibility, reporting and audit rules. The strongest plans map which categories are regulated, which rules apply to each and how the new platform’s controls meet them, all before the first dollar moves. Done up front, that map becomes the approval path instead of the bottleneck.

Participant communication and adoption

A better platform only pays off if participants understand it, trust it and use it. The organizations that see usage move are the ones that explain what changed and make balances easy to see at the same time they switch systems. Clear communication plus real-time balances is what turns a technical upgrade into the adoption the project was meant to deliver.

Point-of-sale controls from day one

Automatic, category-level eligibility checks at the point of purchase make every other control easier. Built in from day one, they mean the program never has to catch and undo purchases a looser setting technically allowed but the program was never meant to cover. They’re also far cheaper to set up at launch than to add later.

Vendor and platform selection

Benefits platforms vary widely in how directly they connect to the eligibility, enrollment and claims systems already running a program. A long feature list isn’t what separates a strong platform from a mediocre one. The real test is how directly it connects, by API, to the systems that already decide who’s eligible for what. Platforms that rely on manual file uploads or batch reconciliation give back most of the speed and savings the upgrade was meant to deliver.

The Benefits Value-Recovery Scorecard

Start with a scorecard, not a vendor demo. It’s a simple way to work out what the current program gives back today, and therefore what a better one is worth, before any technology decision gets made.

Score each question from 1 to 5. 1 = not measured or not available. 5 = measured, owned and actively managed.

• Do you know what share of benefit transactions clear eligibility at the point of purchase, rather than getting flagged after the fact in an audit?
• Can you put a dollar figure on the benefit funds that go unused each year, using your own data rather than an industry estimate?
• What share of transactions clear without anyone touching them by hand, and how many staff hours does the rest take each month?
• Can participants see their own balances in real time, without asking an administrator?
• What share of your workforce reports a clear understanding of the benefits available to them?
• Can you launch a new benefit type on your current platform without building a new, separate system?
• Do you have a written map of which benefit categories carry specific eligibility or compliance requirements?
• Do your controls check eligibility automatically at the point of purchase, and do you measure how often they do rather than assume it?
• Does your platform connect directly, by API, to the eligibility, enrollment and claims systems already running the program?
• What would prove a benefits modernization worked: fewer compliance flags, higher utilization, fewer admin hours, faster new-program launches or all of the above?

A low score isn’t a reason to wait. It shows where the business case still needs building, and that’s exactly where the upside is largest. A high score means the organization already knows its numbers, its risk points and what success looks like, which is the shortest path to a rollout that works.

Measurement note: There isn’t much published industry-wide data on flag rates, exception rates and unused balances for benefit card programs specifically. The figures above are illustrative, modeled at representative volumes rather than drawn from a single published benchmark. Organizations using this framework should swap in their own flag, exception and usage rates, which vary a lot by benefit type, workforce and how mature the program is.

^{1. Galileo Benefits Use Case Positioning Brief↩}