"AI agents are no longer experimental, they're inevitable," Jarek Kutylowski, DeepL's chief executive, told respondents, and his warning is supported by the numbers. DeepL surveyed 5,000 executives across the United States, United Kingdom, France, Germany and Japan and found 69 percent expect agentic AI to reshape business operations in 2026, with 44 percent forecasting major transformation and 25 percent saying change is already underway. The UK leads on measurable AI performance at 80 percent. The practical response for firms is clear: redesign jobs around tasks and invest in AI skills and tooling, while policymakers debate whether taxes and incentives should nudge decisions between hiring and automation.

The claim that agentic AI will remake work next year isn't a prediction from a single enthusiast. It's the consensus of executives surveyed by DeepL and the basis for concrete advice from consultants and HR platforms on how to adapt. Jarek Kutylowski, DeepL's chief executive and founder, put the point plainly: "AI agents are no longer experimental, they're inevitable." The survey of 5,000 leaders across five major economies found 69 percent expect transformation in 2026, 44 percent expect major transformation, 25 percent say change is already underway, and only 7 percent expect no change.

Rethink roles around tasks and capabilities

The managerial demand is simple. Organisations must stop treating jobs as fixed titles and start treating them as bundles of tasks that can be reassigned between humans and agents. That's exactly the shift KPMG UK describes: moving from job-title-based allocation to skill and task-based frameworks, and redesigning workforce planning, learning and rewards around capability rather than fixed roles. Workday UK argues leaders should invest in employee experience, tools and training so staff can work with agents rather than be displaced by them.

Practical steps from consultants map directly onto that framework. Bernard Marr sets out three actions professionals can take now: build AI literacy, decompose roles into tasks to identify what can be handed to agents, and map the available agentic platforms for role-specific delegation. Marr gives concrete examples. Firms can use agents to orchestrate email campaigns, or automate early-stage recruitment steps, shifting routine coordination and candidate screening tasks away from human time. The implication isn't that people vanish. DeepL respondents were split: more than half said AI will create more new roles next year than it replaces, and 52 percent said AI skills will be required for most new hires.

The DeepL data explains why companies are moving: measurable ROI and efficiency were cited by 22 percent of executives as a top driver for deploying agents. Workforce adaptability and enterprise readiness followed at 18 percent apiece. The main barriers, meanwhile, are cost at 16 percent, workforce preparedness at 13 percent, and technology maturity at 12 percent. Those figures make clear where managers must focus: demonstrate return on investment, upskill staff, and pilot agent deployments until maturity and integration reduce the cost friction.

Mastheads

The workplace reorganisation doesn't happen in a policy vacuum. Former prime minister Rishi Sunak argued in The Times that employer national insurance contributions raise the immediate cost of hiring while deploying AI carries no equivalent tax, creating a fiscal bias that could push firms toward automation.

Employer NIC raises more than £100 billion annually for the Treasury, a revenue base Sunak and others say must be rethought in an era where automation is an economically workable alternative to recruitment.

Sunak used a mixture of national and international data to frame the issue. He noted that independent research cited by him found no clear rise in unemployment in AI-exposed roles since the arrival of generative models, but it did identify a slowdown in hiring, most notably among 22 to 25 year olds. Industry data from the British Standards Institution reported 41 percent of businesses saying AI is enabling headcount reductions and that nearly a third now consider AI solutions before hiring. Internationally, Sunak pointed to the scale of job disruption in the United States, where more than 1.17 million jobs were cut in 2025 amid post-pandemic restructuring, to argue this is a broader labour-market shift.

The policy responses proposed in the material run three ways. First, some suggest cutting employment taxes to encourage hiring. Second, others argue for levies on firms that replace jobs with automation, to recapture revenue and slow an immediate swing toward robots over recruits. Third, several sources recommend an AI economics institute with access to live job-market data, a body that would track hiring, displacement and the fiscal effects of automation in near real time. No timetable for such an institute was specified in the reporting, but the proposal would give policymakers the data they currently lack.

Those who oppose urgent tax changes point to the optimistic judgments inside companies. Many executives report measurable performance gains and expect net job creation through AI. That's a legitimate counter-argument. The work from DeepL itself shows firms see ROI and readiness as strong drivers, which suggests adoption is often demand led rather than purely cost led. But the survey and industry data together also show a real risk: where hiring costs are high and agent deployments are immediate and untaxed, the incentives to automate first and hire later are powerful. The consequence isn't simply a mechanical loss of jobs. It's a change in career pathways, with younger cohorts facing slower entry and fewer on-the-job learning opportunities in routine tasks that are now handed to agents.

For managers the response is operational. Map tasks, upskill for AI collaboration, measure ROI and treat adoption as a capability play not a one-off saving. For policymakers the choice is fiscal. Either accept a structural move toward automation and redesign tax and social support around that outcome, or alter incentives so hiring remains attractive where human judgement, learning and entry-level experience matter most.

Related Articles

DeepL frames 2026 as the tipping point when agentic systems move from pilots to enterprise production. That makes this a policy as much as an operational deadline: firms must redesign roles and invest in skills now, and governments will need to choose whether tax settings should preserve hiring incentives or accept a structural shift toward automation.

This article was created with AI assistance.