For Now, Only 11 Percent of Leaders Anticipate AI Job Cuts

A new signal from the executive suite: The AI jobs apocalypse is just not imminent. Just 11 percent of business decision-makers saw AI adoption leading to substantial headcount cuts over the next few years, according to a survey by CRM platform Creatio. What most leaders are projecting, instead, is to put AI to work augmenting their people’s abilities and spinning up new responsibilities as software agents take over routine tasks.

The finding extends a pattern emerging from the rollout of enterprise technology: automation accelerates first in the form of copilots and agents rather than mass workforce replacement. It is also a sign of pragmatic constraints — costs and compliance, to say nothing of accuracy — that may be guiding leaders in favor of augmentation rather than layoffs, at least for now.

What Is Actually Being Conveyed by the Figure of 11 Percent

Creatio’s survey of more than 550 decision-makers suggests a wary, operations-first stance. Eighty-three percent of the participants said they anticipated AI agents would increase productivity for employees — triaging tickets, drafting communications, summarizing data — and that it would give rise to new jobs focused on oversight and orchestration. That fits in with how tool vendors are packaging their wares. This week, for example, Asana announced AI Teammates in beta as virtual coworkers to help teams organize work rather than fully replace humans.

The “for now” qualifier matters.

Short-Term Automation, Long-Term Unknowns

There is evidence that role redesign will come before role removal. An analysis by Indeed of AI exposure in job posts also found that in many professions, it is more likely to shift the skills required or add additional skills than eliminate a profession. At the same time, public anxiety is widespread: a Reuters/Ipsos poll found over 70% of U.S. adults are concerned about AI’s effect on jobs.

Macro-level research underscores the stakes. The International Monetary Fund estimates that AI could impact some 40 percent of jobs globally and around 60 percent in advanced economies, pointing to significant exposure even if displacement isn’t uniform. The World Economic Forum Future of Jobs report predicts job creation in AI-related areas and losses in some routine work, with net impacts depending on how rapidly companies reskill their workers.

Where Leaders Plan to Deploy AI First in Their Operations

Executives are focusing on high-volume, rules-based tasks in areas such as customer support routing and drafting, IT service management, procurement workflow and financial reconciliations. Generative assistants already help with customer operations by summarizing calls and suggesting responses, which frees up agents to tackle harder problems. In professional services, copilots assist with writing proposals and interpreting research, accelerating billable work without taking the human out of the loop.

Real-world deployments show that preference for augmentation. Teams of financial advisers have rolled out AI assistants that surface research for client meetings. Code copilots are used by software teams and boost productivity for experienced engineers but also highlight deficiencies for beginners. These examples adjust the work mix within a role and may increase variation of performance, but they don’t dictate layoffs on day one.

Why Layoffs Aren’t the Immediate Play for AI Adoption

For the moment, three constraints prevent leaders from linking AI to near-term staff cuts. First, trust: hallucinations, the risk of leaking data, and uneven performance don’t make fully automated workflows a safe bet in regulated settings. Second, governance: new laws calling for auditability and human oversight (such as those in the EU AI Act and a changing range of internal AI usage policies) are also making aggregate data analyzable. Third, economics: infrastructure expenses, vendor lock-in and integration work all lead to benefits that are incremental rather than immediate.

There’s also a talent calculus. IBM estimates that approximately 40% of employees will need to be reskilled in the next three years as artificial intelligence redesigns tasks. For many companies, upskilling people already on the payroll is a faster and less risky alternative to large-scale churn. “It’s not like you’re taking this ethical stand that you should be training market-ready employees,” said Peter Cappelli, a management professor at the Wharton School of the University of Pennsylvania.

The Jobs Most Likely to Expand in the AI Economy

The expansion of AI adoption is already generating new specialties: AI product managers who are converting business processes through agent workflows, data governance leads, model risk and compliance officers, AIOps engineers and security personnel focused on early injection and supply-chain threats. Content-heavy jobs are experiencing task reshaping — more reviewing and curating, less first-draft writing — and productivity gains will hinge on clear directives and measurements.

Research from LinkedIn indicates that workers are scrambling to indicate competency in AI, even if they over-check the boxes. It’s a reminder to companies, too: productivity gains take more than simply access to tools. Without it, companies risk uneven performance and burnout as employees try to balance tougher output demands with an additional layer of supervision.

What to Watch Next as AI Moves from Pilots to Scale

The next year may prove whether measured augmentation will lead to broader restructuring. The indicators to look for will be measured productivity deltas at a team level, quality and error rate impacts, new union contracts with AI clauses in them, and regulation moving toward the high-risk spectrum of uses. If reliability and governance improve, that 11% figure could rise — although not consistently by industry or role.

For the moment, the weight has shifted decidedly: executives are using AI to enhance their people and reengineer workflows, rather than to cut jobs. The real job ahead is not replacing work with artificial intelligence but rather rewiring it — systematically, responsibly, and in ways that yield measured returns.