June in Review: The Oral GLP-1 Wave, a Conflict-Free Data Play, and a Consolidation Reckoning

Every month I sit down to write this and try to sort what actually matters from what just made headlines. June was a quiet month for splashy launches and a busy one for the structural stuff that decides who wins over the next few years. Three things stayed with me, and they all circle the same question: when the market keeps handing everyone more options, who is actually helping the employee choose well?

Here is what I watched in June, and why I think it matters.

1. The Oral GLP-1 Wave Is Coming, and It Changes the Math

For most of the past year, the GLP-1 conversation has been about injectables and sticker shock. In June, the story shifted to what happens when these drugs come in pill form and get easier to start.

A Business Group on Health survey released in June put real numbers behind the worry. Among the 105 large employers surveyed, 87% expect that the availability of an oral GLP-1 will drive higher demand overall, and only 9% expect the price to come down. Put those two numbers next to each other. Almost everyone expects demand to climb, and almost no one expects the price to fall. That is a cost problem you can watch coming from a year out.

What struck me most was not the demand number, it was how little plan design has caught up. In the same survey, 83% of employers said they apply the same standard cost-share to GLP-1s that they use for any other medication. After two years of headlines, most employers still have not engineered this benefit at all. They are absorbing a category that behaves nothing like a typical drug using tools built for typical drugs. And the decisions are getting escalated: nearly eight in 10 employers now pull executive leadership into GLP-1 coverage calls, which tells you this has moved from a pharmacy line item to a boardroom conversation.

GLP-1s are really just the sharpest test of something that is true across every category: good plan design is not about what you cover, it is about how you design around it. Coverage without data behind it and without a support program around it is how you end up with runaway demand and no real idea whether anyone is getting healthier. The alternative is design that is built in rather than bolted on: clinical eligibility, a real support program alongside the drug, and formulary choices grounded in what actually works. That is what steers an employee toward the right path instead of the one that is simply easiest to start. Most plans have not done that work yet. They are still treating these like any other prescription, and the oral wave is about to show just how little design sits under the coverage.

2. Independent Data Finally Has a Serious Backer

This is the thread I cannot let go of, and in June it got sharper for me personally. Ben Yomtoob and I have spent the month launching the Benefits Guidance Consortium, and we gained over a hundred followers in the first few weeks, faster than either of us expected. Nearly every conversation lands in the same place: AI is now showing up inside every tool an employee touches, and no one can see where any of it is getting its information.

Picture a fairly ordinary employee. They enroll on Empyrean, their carrier is UHC, they have Healthee as a care navigation tool, and their employer runs Jellyvision for education and decision support. That is four different products, and any one of them may now put an AI in front of the same question: is this the right plan for me? Each model can be trained on different data and tuned for a different outcome. So the employee asks in four places and can get four different answers, with no way to tell which one is the source of truth. And that is before you count the “Shadow Agent”, the public LLM they open in another tab that has no view into their actual plan at all.

None of these tools has to disclose what it was trained on or what it is steering toward. The employee is left refereeing a debate between systems they cannot see into, on a decision that touches both their health and their paycheck. That is the risk hiding underneath all the AI enthusiasm, and it is a big part of why the Consortium exists.

The market is starting to answer this one level up, at the data itself. Peterson Philanthropies announced a $50 million commitment to launch Peterson Health Analytics, a public-benefit company built to give employers independent analysis of what they are actually paying for care, with no financial ties to health plans, health systems, or benefits consultants. A press release is not a track record, and the proof will be in whether employers use it to renegotiate. But the instinct behind it is right. The scarce thing in benefits is no longer another AI that generates an answer. It is a source of truth that no vendor owns.

For everyone building in this space, that points to something uncomfortable but clarifying: the market is going to start pricing in independence. A tool whose AI only ever produces answers that happen to favor its maker will lose trust over time, and the employee is the one who pays for that bias first. The products that last will be the ones willing to show their work, name their sources, and say plainly what they are optimizing for.

June gave a small preview of what that can look like in practice. When bswift announced a licensing agreement with Mayo Clinic to embed clinically vetted content into its recommendation engine, the noteworthy part was not the AI, it was the named clinical source sitting behind it. Tying guidance to a source an employee could actually recognize and trust is a concrete version of naming your sources, and it points in the direction this whole conversation needs to go.

3. Consolidation Comes for Everyone. Proof Is the Price of Staying.

For a decade the answer to every benefits problem was another point solution. A tool for weight management, one for MSK, one for fertility, one for mental health, one for cancer navigation. June brought more evidence that employers have hit their limit.

A spring survey of 166 benefits brokers by the Council of Insurance Agents & Brokers found that clients are running three to four point solutions on average, and that nearly three-quarters of brokers say their clients are actively consolidating vendors or seriously considering it. The buying mood has flipped from collect them all to prove it or lose it.

That same pressure is landing on ben admin, which is where a lot of my head has been this month. The question I keep turning over is whether there is still room for a new AI-native platform to break in. When a tech advisor recommends a new system, is the employer they are advising really buying the tooling or the service, and does the service genuinely get better if AI runs the back engine, the implementation, the renewals, the file builds? Even a newcomer with a gorgeous AI core still has to answer the oldest question in this business, which is who else runs on you and did it go well, and with no customers yet that is a brutal cold start. Meanwhile the big HCMs keep widening their benefits modules, and every feature they absorb is one more reason for an employer to stop paying for a standalone. I do not think the standalone model is finished, but the bar to justify one has climbed a long way.

What all of this rewards is the thing that is hardest to fake: evidence. When the buyer's default was more, brand and buzz carried the day. Now that the default is fewer, the players left standing are the ones that can show a measured result, a clean ROI story, and proof the tool actually changed something for the employee using it. That is a healthier place for the market to be, even if it is a harder one to sell into.

What I Keep Coming Back To

Three different stories, one connective thread. The oral GLP-1 wave, the arrival of independent data, and the point-solution reckoning are all really about the same shift: the era of adding more is ending, and the era of choosing well is beginning. More drugs, more dashboards, more vendors did not make anyone's benefits program better. Better decisions will.

Underneath all of it is the person I keep coming back to: the employee at the end of the chain, trying to make a good call about their health and their money with far less information than everyone selling to them. The solutions and carriers that keep that person at the center are the ones that will earn the next few years. If you are building for 2027 and thinking about how to do that well, I would love to connect.