Workplace wellbeing has never lacked investment. For two decades, employers have added benefits layer by layer: a helpline here, a meditation app there, a coaching session bolted on after a weak engagement survey. The assumption was simple: give people enough options and someone will use them when needed. That assumption is now failing.
Behavioral health claims among US employer plans rose 17% in 2025 alone, according to Brown & Brown's healthcare cost outlook, now ranking among employers' top-five spending categories. Meanwhile, the problem that spending should fix is worsening. The UK's CIPD recorded average sick leave of 9.4 days per employee in 2025, the highest in 15 years, with mental ill health now the leading cause of long-term absence. The World Health Organization estimates depression and anxiety cost the global economy $1 trillion a year in lost productivity. Spending is rising alongside the problem it is intended to address, a combination any finance leader would flag as a program failing its purpose.
But if employee wellbeing can't be addressed by money alone, how can it be solved? Tim State, Founder and CEO of Altus, has seen the issue from both sides, as a solutions provider and a CPO at enterprises, and he thinks the answer lies in reexamining the system itself.
What Increased Spend Doesn't Equal Wellbeing
The clearest symptom that spending isn't translating into employee wellbeing is a number HR has tolerated for years: EAP utilization. State notes that "well over 80% of large employers offer these EAPs," which can be genuinely useful for the issues behind rising absence, but, as he explains:
"famously, the utilization of those programs tends to be in the lower single digits."
That's no exaggeration. Mental Health America puts EAP usage at roughly 4% of eligible employees, unchanged for more than a decade despite near-universal availability. A benefit nearly every employer buys and almost no one uses isn't a strategy; it's a compliance line with a budget attached.
Part of this is structural: wellbeing is delivered as "a broad menu" that asks "the individual person to navigate" it exactly when they have the least capacity, State argues. UK labor research backs this up: a deterioration in mental health affects absenteeism more than three times as much as an equivalent physical decline, so those least able to self-navigate a complex portal are disproportionately the people the system is meant to serve. Systems built around individual initiative fail hardest for people in genuine crisis.
The second driver is more unsettling: a shift in who employees trust. State observes that "more workers than ever have begun talking to an AI chatbot about their stress," while trust in direct leaders is "moving" in the other direction. A 2025 survey covered by Inc. found that more US workers would be completely honest about work stress with a chatbot, at 20.9%, than with their manager, HR or leadership combined. Fullstack HR found that 97% had sought advice from a chatbot instead of their boss. Employees aren't refusing to discuss stress. They're choosing where to do it, and increasingly, it isn't the system their employer built.
A third driver follows: employers measure the wrong signal. "The industry is very good at measuring utilization and risk after the fact and not good at recognizing deterioration early," State says. Dashboards track logins, claims and survey completion, arriving after a problem has materialized rather than capturing what precedes it. With UK long-term sickness at a decade-plus high and billions spent with, in State's words, "mixed results," the issue isn't a shortage of data. It's that the data describes what already happened.
How Businesses Can Build a Better System
State's answer isn't another benefit for the pile. It's a change in what the pile is for. He says employers should
