Interview with Julhiet Sterwen: “The New Challenges of the Employee Experience”

To coincide with the publication of our study on the new challenges facing the employee experience, we spoke to David Gautron, Associated Partner – Employee Experience at Julhiet Sterwen. He discusses the key findings of this survey, including the transformation of employee expectations, changing working patterns, the importance of commitment and quality of life at work, and the need for organizations to make the employee experience a genuine driver of performance, attractiveness and talent retention.

Question 1: In 2026, your study highlights two dynamics that are accelerating in parallel: the lasting installation of hybrids and the rise of AI. What has really changed in organizations this year? Are we still in a phase of experimentation, or have we already entered a more structural changeover?

The Phygital Workplace 2026 Barometer effectively confirms a twofold underlying trend. Firstly, hybrid working is now established. Secondly, AI has become a social fact in organizations. What is changing in 2026 is not so much the nature of the transformations as their degree of irreversibility and the growing tension between individual practices and collective structuring.

Hybridization, as this ^10th edition clearly shows, is no longer seen as an exception or a “modified” organization, but as a foundation: most companies have the necessary tools, agreements and mechanisms in place. And the subject has shifted. The focus has shifted from “where to work” to “what to work for”. Similarly, the focus has shifted from the quantitative (number of days) to the qualitative (value of the collective, role of the office, managerial time).

AI has also changed status, moving from an emerging to a massive subject. In 2026, 62% of employees claim to use it, compared with 38% in 2024, mainly for writing, summarizing, translating or searching for information. This figure marks a clear break: AI is no longer perceived by employees as a technology “yet to come”, but as anchored in daily work life. For them, experimentation is already a thing of the past. However, the maturity gap between populations is striking: 85% of managers use AI, compared with 44% of employees, with similar gaps in training. This reflects a “learning gap” that is becoming an organizational risk: heterogeneity of practices, feelings of injustice or downgrading, concerns about the business impact.

As far as organizations are concerned, the movement is still lagging behind. 57% of companies have begun deployment, but often at an experimental stage. Only 31% are already using AI agents to automate certain processes. The key change is therefore not adoption, but the structural mismatch between rapid individual use and slower organizational transformation. The Barometer thus clearly underlines that, for the majority of companies, AI remains in the pilot or exploration phase. The obstacles are well identified: decision governance, data security, process quality, IS integration, prioritization of use cases. In short, AI has become a personal performance tool, often adopted without any formal framework.

Question 2: We’re seeing rapid adoption of AI: 57% of organizations have already deployed solutions, but with wide disparities in terms of training, usage and appetence. Are we witnessing the emergence of “two-speed AI” in companies? And today, what is really holding back adoption: technology, employee support, managerial culture…?

Your reading of the Phygital Workplace 2026 Barometer is very accurate, and the question you pose is now widely shared by general, HR and digital management. Yes, we are indeed witnessing the emergence of a form of “two-speed AI”. The results of the barometer clearly highlight a double divide.

Firstly, there is a divide between individuals and organizations. In 2026, 62% of employees claim to be using AI, while only 57% of companies have initiated deployment, very often still experimental and not widespread. In other words, usage is progressing from below, through individual initiatives, faster than structuring from above, i.e. with a strategy, governance, changes to the information system.

Similarly, the barometer reveals considerable heterogeneity within organizations:

• 18% advanced users,
• 41% in a learning phase,
• 29% on the sidelines, sometimes worried about the impact on their business.

This polarization is even more marked between managers and employees:

• 85% of managers use AI, versus 44% of employees
• 75% of managers have received training, compared with 37% of employees.

This is what characterizes “two-speed AI”. The gap is becoming entrenched within the workgroup itself, with a real risk of fracturing efficiency, autonomy and even power.

If we take a closer look at the obstacles, we quickly realize that technology is no longer the main limiting factor. The use cases are now well identified (writing, synthesis, research, translation) and the tools widely available. In fact, 31% of companies are already using AI agents to automate certain processes in the most mature organizations. The real hard part remains support and structuring. The barometer speaks explicitly of a lack of support: lack of training, lack of a clear framework, lack of prioritization of truly industrializable use cases…Yet the deployment of AI at scale presupposes reliable data, controlled access, rethought processes, sometimes an overhaul of the IS… The gap between individual adoption and organizational transformation is therefore naturally explained.

Faced with this situation, managerial culture is a decisive factor. Managers are both the most frequent users and the main prescribers, but also, sometimes, the ones who first capture the value of AI. And they are the ones who will be able to convey the value ad hoc. We all know that if AI is perceived as a control tool, a subject reserved for experts, or an HR risk to be contained, it mechanically slows down the spread of uses. To ensure that scaling up works, the most advanced organizations treat it as a lever for autonomy and skills development. They see it as a collective issue, integrated into day-to-day management practices.

Question 3: Looking ahead to 2030, what do you see as the main challenges facing organizations, in terms of both technology and people?

The findings of the Phygital Workplace 2026 Barometer and forward-looking studies on the future of work suggest that the most structuring challenges will be played out simultaneously on the technological and human levels, and above all in the articulation between these dimensions.

First and foremost, we need to make the transition from tool-based AI to “structured-systemic” AI, shifting the scale from experimentation to industrialization. In 2026, 57% of companies had begun to deploy AI, often still on an experimental basis, while only the most mature are already using agents to automate certain processes. The first challenge between now and 2030 will therefore be to move away from the logic of pilots and integrate AI at the heart of business processes, decision chains and information systems. This presupposes robust data architectures, clear governance and the ability to measure ROI, which is still uneven across functions.

Secondly, we need to truly master technological complexity, without being subjected to it. Forward-looking studies all agree that AI will automate a growing proportion of tasks, but not all jobs. The challenge is therefore not so much automation as sustainable cohabitation between humans and intelligent systems, with key issues such as algorithm supervision, human-machine arbitration, responsibility in the event of error… This requires organizations capable of thinking of AI as a socio-technical system, and not simply as a layer of software.

It will also be necessary to provide a framework for responsible use and compliance. The more AI is integrated into sensitive processes, such as HR, finance or managerial decision-making, the more central the issues of transparency, traceability and ethics become. Moreover, research highlights the risk of deploying AI too quickly without consultation, generating tensions and mistrust. I think it’s fair to say that, by 2030, AI governance will become a strategic issue for senior management, on a par with cybersecurity or regulatory compliance.

The natural counterpart to these challenges is, of course, the need to support employees. Human challenges are at the heart of the transformation. As I pointed out earlier, usage is already polarized, with a minority of advanced users, a majority still learning, and a significant proportion of employees on the sidelines or even worried. the risk of fracture By 2030, the main human risk is not technological, but social and organizational: to see lasting gaps in productivity, employability and autonomy between employees. It will be essential to rethink training and skills development on a massive scale, to go beyond “tool training”. It will be a question of “training to work with AI”: understanding its limits, knowing how to question it, maintaining control of professional judgment. To achieve this, prospective studies insist on a shift in skills towards creativity, critical thinking and the ability to cooperate with automated systems. And all this, of course, while reducing the gap that currently exists between managers and employees…