July in Review: Show Me the Math on the Back Office

I spend a lot of time listening to founders pitch their "AI-native" platforms. The demos are usually gorgeous. The vision is always transformative. But when I ask, "Who is actually validating the data?" or "How much manual rework does this actually save?", the room often gets very quiet.

Nobody ever buys a benefits platform because they are excited about the implementation process. It is the unglamorous, heavy-lifting part of our industry: the data mapping, the file builds, the endless testing. But if you want to know who is going to win the next three years, stop looking at the enrollment screens and start looking at the back office.

We have reached the point in the cycle where the market is tired of the magic trick. Whether it is a new AI feature or a private equity roll-up, buyers are demanding to see the math. In July, the most important announcements weren't about what the employee sees. They were about how the plumbing works, who owns the transaction, and what happens when coverage fails to travel well.

1. AI and Consolidation Have Reached the Operating Layer

For a long time, the answer to every benefits problem was another point solution, and recently the answer has been an AI chatbot. July brought more evidence that the market is demanding something deeper.

Bank of America's 2026 Workplace Benefits Report, a sponsor-supported survey, found that 87% of its defined employer sample reported using AI somewhere in their organizations, but only 52% reported using it for benefits administration. Saying you use AI is not the same as using it well. Still, the gap tells you something: the appetite is clearly there, benefits just is not where it has landed yet.

That is starting to change, and the change is happening where it counts. bswift announced an AI-enabled implementation workflow designed to interpret plan documents and generate a near-complete initial site build. It aims at one of the least visible but most labor-intensive corners of benefits administration. It is the first AI claim in a while that made me lean in, because it goes after the file builds and testing cycles that quietly eat an implementation budget, not the enrollment screen.

Consolidation is moving the same way, toward the back office. Principal announced an agreement to acquire Beam Benefits, bringing digital ancillary products and AI-enabled underwriting into a far bigger carrier footprint. Shore Capital acquired ThrivePass, reinforcing the private equity appetite for configurable administration platforms.

What I want to know is not whether a company has AI or a slick front end. I want the boring proof: where the technology actually fits, who checks its work, and what got faster or more accurate because of it. Consolidation raises the exact same bar. Buyers are looking for distribution reach, modern underwriting, and configurable administration. As I said last month, all of this rewards the thing that is hardest to fake: proof you can measure, not a demo you can admire.

2. The Benefit Is Becoming a Transaction

A benefit is not a single moment anymore. It is a handoff, from eligibility to funding to payment to claims, and every seam is a place it can break. The announcements this month show carriers and platforms trying to own more of that chain.

UnitedHealthcare announced a post-tax Lifestyle Spending Account integrated directly with UHC Store, so the money and the checkout finally live in one place. Paytient announced an expansion into prescription benefits, pairing cash-pay pricing with its employer-sponsored payment capability. Sun Life U.S. and Claritev announced a program that uses medical-claims information to identify employees who may be eligible for supplemental health benefits. And on the regulatory side, CMS published a timeline for No Surprises Act remittance-advice remark codes, moving compliance from a legal question to a claims-operations deadline.

The UHC, Paytient, and Sun Life announcements are promises, not proof yet. But they rhyme. The question I keep asking founders is not who has the longest feature list, it is who owns the workflow when the sequence breaks and who the employee calls when it does. Put funding, navigation, and payment in one place and the winner is whoever makes it genuinely simpler for the employee, not whoever adds one more login.

3. Coverage Has to Travel Better

All that back-office work only matters if coverage survives the handoff to the person using it. This month's evidence is a blunt reminder that affordability and portability, the part the employee actually feels, are still real problems for the workforce.

Gallup released research in July (from a survey fielded in late 2025) pointing directly to the workforce consequences of employer-sponsored coverage. Within its defined study population of employed adults with employer-sponsored insurance, Gallup reported that 24% stayed in a job they wanted to leave because they feared losing their health insurance. That is an estimated 23 million adults, and Gallup noted it is 8 percentage points higher than the same measure in 2021. The survey is self-reported, and it does not prove anyone actually quit, but the number is hard to wave away. A quarter of that covered workforce feels stuck by the way their benefits are built.

At the same time, we are seeing just how localized the alternative can be. Colorado's Division of Insurance reported that carriers requested an average 11% increase in preliminary 2027 individual-market filings. Peterson-KFF looked more broadly and found a median proposed increase of 14% (individual requests ranged from 1% to 52%) among 77 insurers in 16 states and the District of Columbia. Those are preliminary requests, not approved rates, and they do not speak to the group market. But the direction is hard to miss.

Choice without the local rate, network, and subsidy behind it just moves the cost onto the employee. For employers weighing ICHRA or another defined-contribution approach, this is a reminder to test the local math before assuming portability solves anything. ICHRA is a local operating model, not a universal answer. The market does not need another way to shift cost. It needs coverage that travels well without breaking the employee's bank.

What I Keep Coming Back To

Three announcements about plumbing, one about job lock, and a stack of rate filings turn out to be the same story. The back office is where the next three years get decided, but it only earns its keep if it delivers something the employee can feel: a plan that fits, coverage that travels, and a number that does not break them. The platforms that win will be the ones that can show the math all the way down to that person, not just to the buyer. If you are building for 2027 and thinking about how to do that well, I would love to connect at info@revgem.llc.