Enterprise AI is solving the efficiency question. In 2026, Harvard researchers say we finally have to face the meaning question.
Most of the enterprise AI conversation in 2026 is, understandably, about performance: faster claims processing, better underwriting accuracy, reduced settlement errors, lower operating costs. What gets discussed far less is a quieter, harder question that Harvard Business School researchers are now raising directly: what happens to the meaning people find in their work, when AI takes over the parts of a job that used to connect them most directly to helping another human being?
A Harvard Business School professor and co-principal investigator at the Digital Data Design Institute frames the shift precisely: 'A lot of work is currently being done on the fi rst-order eff ects, meaning: how do people work with AI? How do we design AI so people can work more effi ciently? But in 2026, we also need to start thinking very carefully about the second-order eff ects: how does AI change my experience of work and its meaning to me?' The prediction is blunt: 'We're going to experience the first glimpses of what the future looks like when work becomes less meaningful because of AI.'
The Customer Service Example, and Why It Generalizes
The professor's illustrative example is customer service, and it's a useful one precisely because it generalizes so easily across industries. In the past, if a customer had a problem with a service or a claim, they reached out to a person — and that employee, in solving the problem, got to directly experience the satisfaction of having helped someone. Increasingly, those conversations are handled fi rst, and often entirely, by AI. The effi ciency gain is real and measurable. What's harder to measure, and easy to overlook, is that fewer employees now get to directly experience how their work positively impacts another person — and for many people, that experience was never incidental to job satisfaction. It was the job satisfaction.
This pattern isn't unique to customer service. It applies to claims adjusters whose caseload increasingly consists only of the complex, ambiguous cases an AI system couldn't resolve — meaning their daily work becomes an unbroken stream of difficult, often emotionally taxing decisions, with the more routine, satisfying resolutions handled invisibly by a system. It applies to HR professionals whose role shifts from actively helping employees navigate benefits and career questions toward supervising an AI system that answers those questions instead.
This Isn't an Argument Against AI Adoption
It's important to be precise about what this research is, and isn't, saying. The HBS researchers are not arguing that organizations should slow down AI adoption to preserve meaning artificially — that would be both impractical and, in many cases, would deny employees the genuine benefits of AI removing tedious, repetitive work. The argument is more specific: organizations that treat AI adoption purely as a process-optimization exercise, without deliberately designing for how it reshapes the human experience of work, are likely to see costs show up later as disengagement, attrition, and burnout — costs that don't appear on the same dashboard as efficiency gains, but are just as real.
The research offers a constructive reframe for how to approach this: a 'shift to process design from process optimization.' Most AI implementations today aim to improve an existing workflow. The more powerful, and more difficult, opportunity is using AI to fundamentally reorganize how decisions and work happen — including deliberately preserving or redesigning the parts of a role that give people a genuine sense of purpose and connection to outcomes, rather than treating every task an AI can technically automate as a task that should be automated.
What This Looks Like in Practice
For organizations deploying AI across claims, underwriting, HR, or customer-facing roles, this suggests a genuinely useful design question that goes beyond 'can AI do this task': should AI do this task, if doing so removes the parts of the role that give employees a sense of meaningful contribution? In many cases, the better design is one where AI absorbs the genuinely repetitive, low-judgment work — freeing employees to spend more time on the complex, relational, or judgment-intensive parts of their role that both create the most value and provide the most meaning. That's a very different design outcome than simply automating whatever can technically be automated and leaving employees with whatever is left over.
The Bottom Line
As AI reshapes how work gets done across every industry, the organizations that will retain their best people — and build genuinely sustainable AI-enabled workforces — are the ones asking not just 'what can we automate,' but 'what should we automate, and what do we need to deliberately preserve or redesign to keep work meaningful.' The efficiency question has largely been answered. The meaning question is just getting started.
References & Sources
• Harvard Business School Working Knowledge. "AI Trends for 2026: Building 'Change Fitness' and Balancing Trade-Off s." library.hbs.edu/working-knowledge/ai-trends-for-2026-building-change-fitness-and-balancing-trade-offs