Why employees can miss pre-AI work and want more AI

Employees can miss pre-AI work and still want more AI

Employees can want more AI while finding its current use burdensome. In Adaptavist’s survey of 2,500 knowledge workers, 65% said they “regularly feel nostalgic about how work operated before the widespread adoption of AI.” Yet 67% wanted their organization to increase AI use. For IT leaders considering expansion, employee enthusiasm answers only part of the question: support for more AI can coexist with problems in everyday work.

These percentages describe the workforce surveyed; they do not show which respondents held both views. Adaptavist’s finding of “pre-AI nostalgia” measures feelings about earlier working arrangements, while the expansion finding measures demand for future use. Both reflect reported perceptions rather than measured organizational productivity. So leaders assessing expansion also need to examine the work employees say AI creates.

AI adds verification and repair work to employees’ workloads

The extra work begins with learning tools alongside existing duties. Adaptavist describes how enterprise-wide access to generative AI tools, which produce content such as text, has expanded over the past three years. The company links the added learning responsibility to employees feeling overwhelmed and nostalgic about earlier working arrangements. Adaptavist also has a commercial interest in organizations adopting AI.

Learning takes time at the outset, but using the tools can make continuing demands. In Adaptavist’s survey, 42% said they “spend more time verifying AI output than they save using it.” Generating an answer or draft therefore completes only part of the task: the employee must still judge whether the result is usable. For these respondents, checking outweighs the time saved.

The checking burden also reaches colleagues who receive generated work. Adaptavist found that 52% of workers regularly spend time correcting colleagues’ AI-generated work, which means one employee’s AI use can add to another’s responsibilities. Measuring individual usage could miss this transfer of effort. Managers need to account for the people who receive the output and make it suitable for use.

When colleagues depend on those contributions, correction work can affect delivery. In Adaptavist’s survey, 49% said low-quality AI outputs delay projects. A quickly generated contribution may need further work before others can rely on it, so generation speed alone says little about the time needed to finish the project. These reported delays make output quality part of the delivery assessment.

Project delays sit alongside a broader concern about collaboration: 55% of respondents to Adaptavist’s survey said AI-generated content reduces overall team efficiency. This finding reflects employees’ assessment of team performance, extending the question beyond one person’s checking time. Together with the correction finding, it points to the way generated work passes between colleagues. Measuring usage and output volume leaves those reported costs out.

The work involved can also change how employees experience their jobs. Adaptavist found that 46% said AI-generated content makes their work feel “more repetitive and less meaningful.” The result concerns what employees spend their time doing as well as how quickly they finish. If responsibilities shift toward checking and correcting generated material, greater AI use can move working life away from the improvement the organization intended.

Dissatisfaction also appears alongside changes in tool use. One in three workers, expressed in Adaptavist’s findings as 36%, experienced “AI fatigue,” or weariness associated with using AI, alongside less frequent tool use and active resistance to using AI for routine tasks. These experiences and behaviors are associated; the findings do not establish a fixed sequence in which fatigue causes avoidance. Still, they show why access does not ensure continued use.

Adaptavist links these less visible productivity costs to rushed implementation without adequate training or transparency. Its explanation points leaders toward how learning, verification, and correction fit into assigned responsibilities. Employees expected to absorb these demands alongside unchanged workloads face an extra burden. Assessing deployment therefore means examining the effort needed to make AI useful.

Employees need clear expectations for AI use

That effort becomes harder to manage when employees are unsure what their organization expects. Adaptavist found that 69% believed their organization used AI ethically, while more than a third, 36%, were confused about expectations for AI use in their role. These are different judgments: confidence in organizational conduct and clarity about personal obligations. Confidence in the first leaves the second question open.

Unclear expectations are also harder to resolve when employees feel their concerns receive no response. In a separate finding, 46% said management had not addressed their AI concerns, which Adaptavist presents as another contributor to implementation problems. Employees need to understand their responsibilities and have a way to raise difficulties with them. Tool access does little to meet either need.

Performance assessment gives this uncertainty a further consequence. Half of all workers in Adaptavist’s survey felt their performance was now “directly or indirectly compared to AI-generated output.” Employees facing such comparisons need to understand how managers will judge their contribution as AI use grows. The findings raise distinct management questions about ethical conduct, role expectations, and performance assessment, even though they do not show which respondents experienced all three.

Age and seniority are unreliable guides to AI readiness

Uncertainty about workplace AI also challenges assumptions about who needs support. Adaptavist found that 42% of Gen Z workers preferred the pre-AI world, compared with 26% of Gen X colleagues. This difference challenges the assumption that younger employees are naturally more ready for workplace AI. Treating youth as evidence of readiness risks overlooking concerns among workers managers might expect to need the least help.

To explain the generational difference, Adaptavist suggests that entry-level workers may face greater disruption while more experienced employees gain new opportunities. This is the company’s proposed explanation. It distinguishes familiarity with technology from confidence about one’s place at work: a younger employee can understand a tool while remaining concerned about what its use means for their career.

Employment concerns extend across the workforce. Overall, 54% of respondents to Adaptavist’s survey said they were “concerned AI could reduce the need for their role within the next five years.” An employee can know how to use AI and still worry about whether their role will remain necessary. Training and communication can address deployment difficulties while leaving that concern unresolved.

The distribution of job-loss concerns also complicates assumptions about seniority. Adaptavist separately reported figures of 23% for junior employees, 29% for C-level executives, 13% for mid-level employees, and 12% for senior employees. Junior workers and C-level executives had the highest figures in that organizational-level set. Concern at both ends of the hierarchy gives leaders reason to assess readiness directly rather than infer it from experience or authority.

For some leaders, this concern has prompted consideration of major career decisions. In Adaptavist’s survey, 47% of C-level executives were exploring a change of industry, a change of career, or stepping away because of possible AI displacement; 36% of directors were considering the same options. These figures describe exploration and intentions rather than completed departures. Even so, the perceived consequences reach beyond learning a new system.

Career concerns also coexist with active preparation for change. Across Adaptavist’s survey respondents, 74% said they were actively learning new skills to remain relevant, rising to 85% among C-level leaders. Skill development shows action, but it does not establish confidence about the outcome. The executive findings likewise describe group-level responses, so managers should assess concern, career plans, and learning separately when deciding what support people need.

Implementation needs clearer guidance and work redesign

Because employees can be learning while remaining concerned, management needs to address implementation and employment questions together. Jobin Kuruvilla is field CTO at Adaptavist, which produced the survey report and has a commercial stake in AI adoption. “Transparency is critical to truly drive AI engagement; organizations must establish clear guardrails and maintain an open dialogue around AI use and employee choice where workers feel they are being listened to,” he told CIO. His proposal combines clear boundaries for AI use with continuing opportunities for employees to influence how it affects their work.

For that dialogue to address employees’ concerns, it needs to cover career prospects as well as tool use. Adaptavist recommends explaining whether and how AI affects careers, with the aim of reducing fatigue. It also calls for implementation roadmaps, plans for introducing AI over time, alongside guidance and training to help employees use it effectively. These are proposed responses to the gaps employees reported, and their value depends on the experience they produce.

Training takes time too, so managers need to account for the responsibilities employees retain. If learning AI is added to an unchanged workload, training adds to the burden it is intended to ease. IT leaders and their executive colleagues therefore need to decide how work will change. They must establish how employees will learn, apply, and check AI-assisted work within the duties they are expected to complete.

Making room for learning and checking matters because the intended benefit depends on changing how employees spend their time. Anand Unadkat, a senior solutions architect at Atlassian, describes that ambition from a company with a commercial interest in organizational AI adoption. “Ultimately, by automating the mundane tasks that make work feel repetitive —organizations can refocus their specialists on high-value creativity, transforming AI from a source of fatigue into a powerful engine for meaningful human achievement,” he said. His proposed sequence is to automate routine tasks, release specialists’ time, and direct that time toward more valuable creative work.

But this ambition sits in tension with employees’ reports of more repetitive and less meaningful work. Managers need to examine whether automation frees time for work that employees and the organization value, or shifts effort toward reviewing generated material. They can make that assessment by identifying which duties disappear, which checking tasks are added, and what employees can do with any time released. Improving those responsibilities still leaves a separate question to address openly: how AI changes the organization’s future need for each role.

Key takeaways for decision-makers

• Assess support and friction separately: Adaptavist found that 65% of employees missed pre-AI work, while 67% wanted more AI use. These workforce-level findings give IT leaders reason to assess demand for AI alongside employees’ experience using it.
• Measure checking and correction work: 42% said verifying AI output took more time than AI saved, and 52% regularly corrected colleagues’ AI-generated work. Managers assessing productivity need to track review effort, handoffs, and project delays across the people receiving that output.
• Clarify role and performance expectations: 36% were confused about expectations for AI use, while half felt their performance was compared with AI-generated output. Managers need to explain acceptable use, quality standards, and how AI affects performance assessment.
• Assess readiness across age and seniority: Younger workers expressed greater preference for pre-AI work than older colleagues, and job concerns reached both junior employees and executives. HR and department heads need to assess skills, confidence, and career concerns separately when planning support.
• Make room for learning and redesign: Adding AI training to unchanged workloads creates another demand on employees’ time. Department heads need to allocate learning time, identify which duties automation removes, and account for new checking responsibilities while explaining potential career effects.