Polish managers expect artificial intelligence to lift productivity, but many also believe it could reduce staffing needs. According to SAP Poland’s ‘AI Readiness Index’ study, 56% of surveyed managers expect employees to be required to complete more tasks in the same amount of time as AI becomes more widely deployed. At the same time, 54% believe the technology could lead to fewer positions in their department within the next three years.
The expectations are moving faster than the actual transformation. Nearly half of respondents — 49% — say AI implementation is progressing more slowly than previously announced and has not yet fundamentally changed how their teams operate.
Higher productivity expectations, but slower transformation
The findings show a gap between installing AI tools and redesigning how an organisation works. Companies may be able to provide access to new systems relatively quickly, but changing workflows, responsibilities and employee habits takes longer.
Katarzyna Rusek, vice-president and HR director at SAP Poland, argues that AI deployment is not only a technology project. Organisations also need to build new skills, create clear rules for using AI and develop trust in systems that increasingly support day-to-day work.
This distinction matters because the productivity gains expected by managers will not automatically appear simply because employees are given access to generative AI or other automated tools.
Training is becoming part of AI adoption
Among managers working in organisations that have already implemented AI, 65% say their companies provide training or upskilling programmes related to artificial intelligence.
As AI tools become more capable, employees need more than basic technical familiarity. They increasingly have to know how to delegate tasks to systems, assess the quality of generated results, identify errors and decide when human judgement is still necessary.
This also changes the role of managers. Supervisors need to determine which processes can be automated, how performance should be measured and where accountability remains with people even when part of a task is performed by AI.
From AI assistants to autonomous agents
The next stage of adoption is expected to involve AI agents capable of carrying out sequences of actions rather than supporting only individual tasks. Today, many systems are used as assistants for research, writing, data analysis or other isolated activities. Agent-based systems can go further by executing multi-step processes under predefined rules.
That shift could change the division of work between people and software. Repetitive actions may increasingly be handled by systems, while employees focus on tasks requiring experience, creativity, negotiation, responsibility or expert judgement.
The potential effect on employment therefore depends on how companies redesign processes rather than on the capabilities of AI alone. A reduction in headcount is one possible outcome, but so is a change in job composition and a reallocation of employees toward higher-value work.
Survey covered managers in companies already using AI
The SAP Poland ‘AI Readiness Index. The Road to an Autonomous Enterprise’ survey was carried out in August 2026 using the CAWI method by UCE Research. It included 411 managers from companies employing at least 50 people.
Respondents had influence over IT spending decisions in their teams or business areas, and participation required that their organisation already use at least one AI solution in everyday work.
The results suggest that Polish companies are moving beyond the question of whether to adopt AI. The more difficult issue is now how deeply the technology will be integrated into operating models — and how quickly that integration will translate into measurable productivity gains, new skills and changes in staffing.
