Agentic AI: The Question Is No Longer Whether to Use It, but How to Redesign Work

Over the past two years many organisations have tried artificial intelligence, launched pilots and invested in new technology. The results have been uneven: some teams gained efficiency, others lost confidence in the outcomes, some processes got faster while others became harder to govern. Top Employers Institute’s report “Designing work in the age of agentic AI” starts from there. A pulse survey run in March 2026 shows that about 31% of respondents already use some form of agentic AI and 41% plan pilots or implementations this year, in line with McKinsey, where 23% say they have an agentic system running in some area of the business and another 39% are experimenting. Adoption has therefore begun. What matters now is redefining processes and workflows, and the study draws on almost 2,500 certified organisations in 131 countries.

First Principle: Design for Trust, Clarity and Adoption

The more tasks AI takes on, the less it matters what the technology can do and the more it matters how people understand it and whether they trust it. According to Upwork, 96% of senior leaders remain confident that new technologies will improve employee performance, yet nearly half of workers do not know how to get those improvements. Adoption therefore has to be designed: it needs clarity on what the system does, on what is expected of people and on how reliable the tool is. 40% of Top Employers organisations continuously assess their AI implementations, looking at wellbeing, job satisfaction, skills development and the quality of the workflow, and 38% involve employees in discussions on AI ethics, so that people do not simply undergo the change but have a say in how it is introduced.

Second Principle: Update Hiring, Training and Pay

The real challenge is not adding a few new skills but understanding how roles, performance and workforce planning change. The market sends a clear signal, since AI skills are scarce and in high demand. Even more important is what happens inside organisations: 70% are already preparing their entire workforce for change, not just a small group of experts. Gaps remain, though. Only 61% have a formal strategic workforce planning process and just 41% consider multiple future scenarios. Planning around stable roles and predictable demand makes it hard to see where capacity really comes from and where new bottlenecks appear. And because results increasingly depend on how a person combines their abilities with the systems around them, pay, recognition of value and fairness need rethinking too.

Third Principle: Ethics as Infrastructure

For Top Employers, ethical AI is not a statement of values but a governance system that supports trust, accountability and scale. It covers cyber risk, bias, explainability, employee wellbeing and how people actually experience AI in their daily work. 65% of Top Employers organisations say they follow responsible implementation practices and 62% have formal AI ethics frameworks. Those that do are more likely to assess how collaboration between people and AI works in practice, to involve employees and to invest in preparing the whole workforce.

What the Most Advanced Organisations Do

The research identifies three recurring elements. The first is shared construction of the rules: HR, IT and operations align on how AI is used, where the limits are and how decisions are made, balancing business results and impact on people. The second is assessing impact over time rather than only at launch, because value, costs and changes to work emerge after the first efficiency gains. The third is leading with values as well as processes, explaining how and why AI is used and promoting adoption through transparency.

What It Means for Companies

Agentic AI is no longer only a technology topic, it is a question of leadership and organisation. It is a field that closely touches Euren’s work: understanding which skills really exist in a company and which are missing through mapping, assessing people’s ability to work alongside intelligent systems, supporting change with development and change management programmes, and identifying the leaders able to guide this transition. As the report concludes, the challenge for HR leaders is not to chase innovation, but to make it understandable, governable and meaningful for the people doing the work.