Which roles within Foyer are already using artificial intelligence, or are even undergoing transformation?
Rémy Els. – “I would first distinguish an initial phase of AI, the one we have experienced so far: AI as an assistant, as a means of automation, which helps everyone in their work. It therefore enhances quality and efficiency, but does not fundamentally change the process or the nature of the work. With the advent of agent-based AI, we are entering a more disruptive phase, involving changes to processes and interactions between colleagues. This impact is much more recent. As a CIO, I see a shift particularly in software engineering and development, with smaller teams but software engineers whose skills and responsibilities are expanding. The way we interact with business units is also taking place within much shorter cycles.
Do you have an example of how it’s used?
“When it comes to handling motor claims, for example, we have integrated a solution that allows customers to take photos of the damage to their car. The tool provides an initial assessment and an initial estimate. In straightforward cases, this replaces the need for an assessor to visit. Otherwise, it at least assists the assessor and, in some cases, changes the way the claim is handled. Customers save time and the process is more cost-effective. Especially as assessors are quite difficult to recruit.
These changes may give rise to concerns – how can we gauge how staff are feeling?
“We have launched a comprehensive programme to foster a culture of AI, which has just got underway with two days of events called ‘AI Demo Days’, open to all staff. This has enabled some of them to realise just how much we have already developed within the Group, but also – and above all – that there is a great deal one can do without necessarily being an expert or a technician. For this programme, we’ve adopted the image of a tandem bike. It’s a bit like our motto: the human holds the handlebars and steers, whilst the AI, sitting behind, provides support by pedalling.
We have also built up a pool of around forty AI ambassadors: tech experts and non-tech experts alike, people from all walks of life, junior staff, senior staff… This also humanises AI. Because, let’s not be under any illusion: in some cases, the challenge associated with using AI is not a technical one. It relies more on business expertise than on technical specialists. But the need for support is constant. We have therefore also developed an internal barometer that gathers feedback and insights from staff on specific topics. We strive to stay in touch with what’s happening on the ground.

How do we decide what can be handled by an agent-based solution, and what should remain ‘human’?
“Ever since the dawn of digitalisation, some twenty years ago, we have incorporated the concept of the emotional level into the design of our solutions: if an interaction between the customer and the insurer is highly emotional, then we prioritise human contact. If the emotional level is low, then there is potential for automation. AI does not call this principle into question; quite the contrary.
Our experience has, in fact, reinforced our belief in this approach. A few years ago, we tried using an app to complete an accident report form. It was a failure because, immediately after a car accident, what a customer needs most is reassurance and support, not to fill in a form on their smartphone. That is why, in such situations, we prioritise direct contact with an agent or the assistance service, so that the customer can be reassured and supported. Conversely, when it comes to fraud – a common issue in the insurance sector – AI can help identify correlations and detect suspicious activity more quickly and effectively.
If the challenge is more about people than technology, how should a CIO approach it?
“The technological challenge is my job. An engineer’s job is to ensure that technology becomes a convenience for the user. So, the more of it there is, the happier we are. But the human challenge is very real. The speed at which this technology is evolving places us in a phase of constant transformation and uncertainty. Our challenge, as CIOs and as leaders, is to steer our organisations and our staff through these phases of uncertainty, which are just as much phases of opportunity that we can seize.
I don’t think I’ve ever brought as many technical topics to the board in my career as I have over the last six to twelve months. My challenge is to put myself in the shoes of the various roles around the table so that I can explain the technology in layman’s terms, foster dialogue and have a constructive discussion on these topics. This applies to C-level executives, but also to other staff members, who will need to take a greater interest in technology than in the past.

Which skills is it absolutely essential to ensure are retained?
“Creativity, for which agent-based AI can be a tremendous aid and tool. In software engineering, for example, it is useful to have teams augmented by AI, which act as ‘hackers’ to come up with the right ideas and prototype them quickly; but we also need a second team made up of engineers, architects, security specialists and legal experts, who can build these ideas on a solid foundation to ensure they can be scaled up. This is a model we have tried out and which works well.
Rigorous execution is another essential skill. These two skills may seem to be at odds with one another, but combining them is a good solution. This involves cooperation and collaborative development – and AI does not render this obsolete! Finally, decision-making must remain in human hands; this must be a core principle. At Foyer, it is enshrined in our Pact for Responsible AI.
How can we prevent everyone from experimenting ‘on their own’, which is not without its risks?
“We have put in place a three-tiered process and a governance framework. It is important to bear in mind that with AI – particularly agent-based AI – safeguards are necessary to prevent data breaches or incidents. Internally, we have also developed our own AI, MIA, which stands for My Intelligence Augmented. It has been operational since 2026 and has become the brand name for all the AI solutions and tools we offer our staff.”
This article was written for the December 2026 issue of Paperjam magazine, published on 18 November. The content is produced exclusively for the magazine. It is published on the website to contribute to Paperjam’s comprehensive archive. Click on this link to subscribe to the magazine.
Is your company a member of the Paperjam Club? You can apply for a subscription in your own name. Please let us know via club@paperjam.lu.



