Artificial intelligence (AI) is gradually redefining the way we use the Modern Workplace, through automation, augmented collaboration and new security challenges. According to Cédric Minne, Senior Manager Modern Workplace, this transformation marks the arrival of a new generation of smarter, more predictive and human-centred work environments. He discusses practical uses of Microsoft Copilot, the challenges surrounding its adoption and how it is evolving.

How does the native integration of AI transform the “Modern Workplace 2.0” into a truly intelligent and predictive ecosystem, as opposed to traditional workplace environments?

In recent years, the “classic” Modern Workplace has already represented a major transformation: Teams, SharePoint and OneDrive have fundamentally changed the way we collaborate, remotely, in real time and securely.

Yet we were still operating within a tool-based logic: we had to open applications, search for information, piece together conversations and follow up with colleagues. The environment was connected, but essentially reactive.

With the native integration of AI into Microsoft 365, we are moving into a different dimension. AI becomes a contextual layer capable of understanding meetings, documents, conversations and working habits in order to make suggestions even before they are requested. The environment no longer simply hosts information: it helps turn that information into action.

A concrete example: instead of spending 20 minutes reconstructing the history of a client relationship before a meeting, Copilot can summarise previous exchanges, identify decisions that have been made, highlight open points and suggest an agenda structure, as well as follow-up actions. This is not “magic”; it is, above all, a new interface to work, one that is closer to a conversation than a treasure hunt.

One point remains essential: AI does not forgive disorder. Copilot relies on existing data and permissions. If the environment is poorly structured or poorly governed, AI amplifies those shortcomings rather than creating value, and may even generate “hallucinations”. As the saying goes, “better safe than sorry”: AI running on poorly governed data is like putting a turbocharger on a car without brakes.

From eliminating repetitive tasks to enhancing creativity and decision-making: which concrete uses of Microsoft Copilot are transforming employees’ working days and creating the most value?

The use cases that create the most value are not necessarily the most “spectacular”. They are primarily those that simplify everyday work and reduce friction. I see three major contributions.

The first concerns the automation of repetitive tasks: summarising emails, drafting meeting minutes, extracting action items from a meeting, rephrasing a sensitive message or structuring notes. This delivers an immediate time saving and is often the starting point for demonstrating a return on investment.

The second is about enhancing creative capabilities. In Word or PowerPoint, Copilot helps move from a blank page to a first structured version: an outline, an argument, a summary. AI accelerates the formatting and structuring process, but humans retain the essential role of adding nuance, judgement and critical thinking.

Finally, the third contribution is improving decision-making. When working with a corpus of documents or data, Copilot can help identify trends, inconsistencies and points requiring attention, and then produce an actionable summary. What previously required half a day can become material that is ready for decision-making.

For example, in a workflow involving several versions of documents, meetings, emails and approvals, Copilot can summarise, compare, draft an email highlighting discrepancies, produce minutes and feed information into a presentation. The value comes from accumulation: saving a few minutes ten times a day eventually adds up to hours.

One principle remains fundamental: “Copilot, not Autopilot.” If we let AI drive on its own, we expose ourselves to risk. If we put it at the service of human judgement, we gain speed and quality.

Companies are looking to provide increasingly seamless and mobile work experiences while strengthening their security requirements. How can these seemingly conflicting objectives be reconciled?

Security and productivity have always been in tension: too many visible controls frustrate users and slow down adoption; too much invisible freedom increases risk. AI intensifies this tension because it relies on data and access rights.

The answer today lies in the Zero Trust model: never trust by default and always verify based on context: identity, device, location and risk level. This is precisely what makes it possible to remain mobile without being naïve: security is applied “in the right place, at the right time”, rather than through uniform rules that penalise everyone.

In practice, this translates into adaptive security: stronger authentication when accessing resources externally or from a non-compliant device, but a seamless experience in a trusted environment. Security becomes less of a wall and more of a safety net, discreet when everything is fine, robust when needed.

With Copilot, this question becomes even more important: AI does not “create” access; it uses the access rights that already exist. If permissions are too broad, sensitive data is poorly classified, or data loss prevention and retention mechanisms are not in place, Copilot can surface sensitive information very quickly, sometimes even unintentionally.

This leads to a systematic rule: we do not deploy AI “on top of” a disorganised environment. We follow a structured approach: assess, secure and govern, then adopt and deploy at scale. Without trust, there is no adoption, and without adoption, there is no value. “Trust does not exclude control”, it depends on it.

How can companies overcome the cultural and psychological barriers to adopting Modern Workplace 2.0 and turn concerns about AI into engagement?

The main barriers are rarely technical, they are human. Fear of losing control, concern about being judged, or simply fear of the unknown. As the saying goes, “what we don’t know scares us”; this is true in technology just as it is elsewhere.

The first step, therefore, is to establish a climate of trust: protect data according to its sensitivity, clarify the rules and demonstrate that AI is part of a responsible governance framework. It is also essential to establish a very clear framework: AI assists, but humans remain responsible for decisions and deliverables. It is a simple message, but an extremely reassuring one.

The second step is to consider adoption as a discipline rather than a launch. Train, coach, create champions, share concrete use cases by business function and, above all, provide long-term support. In the technology sector, training is continuous. AI is evolving at an unprecedented pace, much faster than previous waves of technology; yet people need time to become comfortable with the tool and turn it into a genuine habit.

Finally, at C-level, we need to speak frankly: AI comes at a cost. It is therefore essential to measure the benefits: time saved, improved quality, satisfaction or faster decision-making. In my experience, engagement happens when AI is linked to tangible benefits, rather than presented as an abstract promise.

Are we on the verge of a structural change in the way we work? How do you envision the evolution of the Workplace over the next five years?

Yes, and I even believe that we are still underestimating the scale of this transformation. We have experienced the era of “humans + tools”. We have entered the era of “humans + assistants”. We are now moving into the era of Agentic AI: agents capable of carrying out complete tasks, collaborating with one another and executing end-to-end processes under human orchestration. This is at the heart of the Frontier Firm concept: an organisation where AI becomes an operational foundation rather than a gadget.

Over the next five years, this model will probably become widespread: hybrid teams in which humans frame, supervise and make decisions, while agents execute, analyse and optimise. Every department will have its own agents: HR, finance, sales, IT, legal. Employees will increasingly become conductors rather than executors.

We will see the emergence, and this is already visible in some trajectories, of startups built around a radical model: a single leader surrounded by a constellation of agents capable of handling the majority of operational work, from market analysis to content creation, support, reporting and automation. The leader no longer does everything; they set the direction, establish guardrails, make decisions and manage the performance of an agent-based system.

Is this a good thing? Everyone can form their own opinion. But history is consistent: with every technological revolution, those who learn to integrate it intelligently move further and faster. As with industrialisation, the question is not whether change is coming, but how we frame it, how we protect people and how we turn technological power into sustainable value.

Many companies in Luxembourg and Belgium have already begun this transformation. At Business Elements Reply, we support these organisations throughout this strategic transition. To best meet their needs, we are currently developing a partnership that makes a great deal of sense and will bring real added value and additional expertise to our clients.

This interview was originally published in French by The Dots, a Luxembourg-based platform covering technology, innovation and digital topics.