Would You Work Harder for a 2% Raise? Omnisend Is Betting Its Whole AI Strategy on It.

Most companies talking publicly about AI adoption right now are talking about it through the lens of who becomes unnecessary. Omnisend, a Lithuania-based marketing automation platform with roughly 250 employees, decided to run the experiment in the opposite direction. Starting in April, the company began offering staff a 2% to 4% raise, layered on top of its existing quarterly review cycle, tied specifically to demonstrated AI impact. More than half the company qualified in the very first quarter the program existed.

The Triad of Measurable AI Impact

The mechanics matter more than the headline number, because they're the part any other employer could actually replicate. To earn the raise, an employee has to show measurable value in at least one of three categories: efficiency, meaning a genuine, quantifiable time or cost saving; impact, meaning a workflow that visibly moved a real business metric or improved output quality; or scalable adoption, meaning something the employee built that other people or teams picked up and started using on their own, without being told to. The finance team qualified by lifting invoice-automation coverage from 36% to 85%. The legal team qualified through AI-assisted vendor assessment and legal-intelligence workflows that measurably cut manual review time. Several engineers earned the raise not for their own personal output but specifically because a workflow they built spread across other teams unprompted — the adoption criterion doing precisely what it was designed to reward. Managers, not a centralized committee, make the final call on whether the bar has been cleared, which keeps the judgment close to the actual work being evaluated.

Cultural Groundwork: Play Before Pay

Omnisend didn't lead with the money, and that sequencing turns out to be a meaningful part of why the program worked as well as it did. Before tying any pay to AI use, the company spent time building familiarity first: internal games and short videos to get staff comfortable with the tools, and monthly "AI days" where employees cleared their calendars specifically to experiment, with no output expectation attached to the day at all. Only once that foundation existed did leadership introduce the incentive structure on top of it. CEO Rytis Lauris has been candid that the program cuts both ways — it's explicitly designed to reward what he calls highly effective people, and he's said plainly that compensation across the company is already beginning to diverge based on who leans into the tools and who doesn't. Bernard Meyer, Omnisend's head of AI operations, is equally candid about the harder part: measuring AI's actual business impact is, in his own description, honestly difficult, and unevenly so depending on the function. A sales workflow that saves demonstrable hours is a much easier case to make than a creative or strategic role where the value AI adds is real but far harder to isolate and quantify.

Solving the 'Stick' Problem: Removing the Fear of Efficiency

What makes this case worth studying rather than simply admiring is the specific failure mode it's designed to prevent. Wharton's own research on AI incentives has pointed out that most employees are still quietly asking a very basic question — what's actually in this for me — and that without a clear answer, workers who genuinely do find real AI-driven time savings often avoid reporting them at all, worried they'll simply be asked to do more with the same headcount, or worse, that the savings will eventually be used to justify eliminating their role entirely. That's the "stick" version of AI adoption, and it's the default posture most organizations are running whether they intend to or not, because saying nothing about incentives is itself a kind of incentive. Omnisend's bet is that naming the carrot explicitly, and attaching a real number to it, gets a company further than either silence or a mandate would.

Scalability Limits: From Boutique Agility to Enterprise Bureaucracy

The honest caveat is that Omnisend is a specific kind of company — roughly 250 people, real budget constraints, a founder-led culture where the CEO can credibly claim to know which employees are genuinely driving impact. Whether a 2% to 4% raise moves behavior the same way inside a 50,000-person multinational, where the distance between an individual's AI-driven workflow and the executives deciding compensation is considerably longer, is a genuinely open question rather than an assumption this case study can settle on its own. What it does settle is narrower and more useful: that at least one company has tried paying for AI adoption instead of mandating it or threatening layoffs over it, and the early results suggest that approach is worth more organizations actually testing for themselves, rather than dismissing on principle before they've tried it.