The Off-Ramp

For years, the mental-health-benefit conversation was organized around access. Could employees find a provider? Would they use the benefit? Could the organization offer something more credible than a referral line that most people never called?

Now the question arrives with a budget attached. Cost is the most important factor in benefits decisions for 54% of employers, according to Lockton's 2026 National Benefits Survey, up from 38% the prior year. NFP reports that average employer spending per employee on mental-health resources has fallen by roughly 7% year over year.

The easy response is to reduce the program. It is also the response most likely to create a new problem: employees experience a benefit they trusted as being taken away, while the employer returns to a lower-engagement model that was already failing to reach the workforce. The distinction that matters is between reduction, which removes access to lower the cost, and redesign, which changes the model so that the care employees depend on survives the change.

Employers know that is not a clean trade. In the same Lockton research, 81% say employee impact is a primary consideration when evaluating benefits changes. And Spring Health's 2026 research exposes the measurement problem underneath the decision: only 9% of HR professionals say their current mental-health solution is clearly reducing health-plan spend.

Benefits leaders are being asked to cut more carefully and prove more clearly. That is the real context for the next renewal.

When a benefit becomes difficult to sustain, a price comparison feels like the responsible place to start. It is not enough.

The question is not whether a different program has a lower headline cost. It is whether it preserves the parts of the experience that made employees engage in the first place, and whether it gives the organization a credible way to measure the result.

This is not a semantic distinction. Employees can tell the difference between a benefit being removed and a care experience being redesigned around how they actually seek help.

The standard should be clear before a benefits team begins comparing options.

Any proposed replacement that fails this test may save money. It is still a downgrade.

Thoughtful starts with a different premise: people do not need the same kind of support at every moment. Some want an immediate, private place to work through a problem. Some want provider care first. Some need both, with the balance changing over time.

T, Thoughtful's AI mental wellness support, offers short, time-bound wellness sessions at any hour. Its skills practice draws on principles from approaches such as CBT and DBT. It is not therapy, and it is not intended to replace a provider when a provider is needed. It gives a member a low-friction place to begin when stress is present but an appointment does not yet feel like the right or available first step.

Its differentiator is not simply availability. T carries relevant context from one interaction to the next: themes the member has been working through, goals they have set, skills they are practicing, and patterns they want to change. That longitudinal memory lets support become more tailored with use. Members do not have to re-explain their circumstances at every session or receive generic support that ignores what came before.

When provider care is appropriate, members have direct access to Spring Health's network of more than 16,000 licensed therapists, coaches, and psychiatrists. With member consent, providers can use T session summaries to inform their work and recommend skills exercises for members to continue between appointments. The result is one connected experience rather than separate tools that do not know about each other.

This is the positive case for a change in benefits strategy. It is not a swap from human care to AI. It is a way to preserve human care while adding immediate, continuous, and increasingly personalized support around it.

The decision to redesign a benefit needs more than a compelling product story. It needs evidence that each layer of care can stand up on its own and work together.

Spring Health's provider-care outcomes are peer-reviewed and independently validated. Thoughtful's own randomized controlled trial included 229 participants; after two weeks of use, the intervention group showed statistically significant improvements across eight dimensions of psychological health, including emotional wellbeing, loneliness, self-esteem, behavioral activation, and help-seeking attitudes.

The evidence does not guarantee the same financial or workforce result for every employer. It does establish what a serious evaluation should include: access to care, clinical progress, member experience, safety performance, and the ability to connect program data to the business questions leadership is already asking.

That is particularly important when only 9% of HR professionals say their current mental-health solution is clearly reducing health-plan spend. The gap is not only between investment and cost. It is between investment and evidence.

Price matters. It is not the first design decision.

The commercial question is not whether engagement has a cost. It does. The useful question is whether the organization understands the cost of the care experience it is choosing to preserve, and what happens if more employees actually use it.

That shifts a renewal conversation away from a single headline rate. A benefits leader and Finance partner should be able to model expected participation, understand what is included for a participating member, and test the range of spend at different levels of engagement. They should also be able to see whether a lower price is being achieved by narrowing provider access, adding friction, limiting continuity, or simply returning to a model fewer people use.

Thoughtful uses a participation-based model: employers pay for active participants, with AI sessions and provider sessions included for that participant. The intended effect is to make the cost of a member's care legible once they engage, rather than add a new unit charge every time they need more support.

That does not make one contract structure right for every employer. It makes the comparison more honest. The question is not only “what is the PEPM?” It is: “What care experience are we protecting, what happens if participation rises, and can we keep supporting that outcome?”

Employees notice a change in a benefit they use. The internal narrative therefore has to be true before it can be persuasive.

Employees need to understand what will remain: access to licensed providers, a clear safety pathway, and confidential support outside working hours. They should also understand what changes for the better: a more personalized experience that builds on the member's own history and goals, continuity between immediate and provider-based support, and a model the employer can fund without quietly limiting care.

That is the difference between an announcement of a cut and an explanation of a better operating model.

1. Identify what employees actually value in the current program: provider choice, access, confidentiality, support outside work hours, or continuity across care.
2. Test every option against the preservation standard: human care, timeliness, clinical quality, safety, continuity, personalization, and sustainable economics.
3. Request current evidence for every claim: provider availability, clinical methodology, outcome measurement, safety protocols, privacy protections, and contract economics at different participation levels.
4. Define the measures that will tell leadership whether the redesign worked before the contract is signed.
5. Build the employee communication before announcing the change. If the new model cannot be described honestly as preserving or improving care, it is a cut.

The off-ramp from an expensive program should not lead back to an unused one. It should lead to care that employees can rely on, benefits leaders can explain, and employers can sustain.

Thoughtful offers AI mental wellness support alongside access to Spring Health's provider network. Its model is built for people who want immediate, low-friction support, people who prefer provider care, and people whose needs change between those paths. T's longitudinal memory lets support become more personalized across time; with member consent, that context can strengthen provider care rather than sit beside it.

Thoughtful's position is not that AI alone is the answer to workforce mental health. It is that employers should not have to remove high-quality human care in order to address cost pressure. A blended, clinically accountable model can preserve care while making it more continuous, personalized, and financially sustainable.

For what to measure once a change is made, read The New Scorecard, and for why low usage is a design problem rather than a demand problem, read The Silent Majority.

Learn more about how Thoughtful works for employers.

Market figures come from Lockton's 2026 National Benefits Survey, NFP's 2026 U.S. Benefits Trend Report, and Spring Health's 2026 Workplace Mental Health Annual Report. Product, network, access, safety, pricing, and outcomes statements must be validated against Thoughtful's current approved claims before publication. Provider-network and Thoughtful clinical evidence do not guarantee a specific financial or workforce outcome for every employer. This report does not compare Thoughtful with a named competitor or make unsupported claims about another company's price, clinical quality, or access.