Navigating the AI transition in marketing
Our guest
Jean-Philippe Devos has been working for Proximus for 25 years. Trained as an engineer, he started his career in intelligent networks, then moved into IT architecture, before joining marketing. He worked on retention and loyalty programmes before taking on, about five years ago, a hybrid position at the crossroads of business and technology, focused on AI and data. A part of Jean-Philippe’s mission is to evangelise AI within Proximus. His department is responsible for keeping the rest of the organisation informed about the latest developments in AI. They have notably created a monthly internal newsletter for their division, and they also maintain active technology watch practices by sharing relevant articles with their own team as well as with colleagues across other departments.
Proximus has been using predictive models for years. What does today’s AI bring to the table?
Indeed, we have been doing churn prediction for fifteen to twenty years. AI doesn’t replace this work; it enhances it. It lets us enrich models with new data, gives us clearer metrics to evaluate their performance, and tells us when a model needs to be refreshed. AI is essentially a way to amplify business KPIs that already existed. Not all benefits are strictly financial, enabled by retention models for instance. There are softer gains too. Better tools make employees more efficient and more satisfied. That influences how they interact with customers, how they feel at work, even how likely they are to stay. These things are harder to quantify, but they do matter.
What is the main challenge you face when rolling out AI internally?
People are curious about AI. They have a strong appetite to learn, they want to use the tools. The real difficulty is keeping up with a technology that evolves at an unprecedented pace. Our biggest challenge is to increase the AI literacy of our employees. The workshops we gave two years ago on prompting are already outdated. We try to educate people, provide guidance, explain the rules, share best practices, but we are not trainers. And doing this takes time away from project delivery. Another challenge is the abundance of AI tools and educational content. There are thousands of videos, articles, conferences. Finding information is easy. Finding good information takes hours. You can watch ten videos and only one may be worth recommending.
How does Euranova help you address that challenge?
Euranova plays a critical role in structuring and accelerating adoption. Together, we set up a community of ambassadors, the GenAI Heroes. Each month, the ambassadors gather for lunchandlearn sessions, workshops, internal presentations and guest talks. These sessions are inspired by Euranova’s own practices in collective intelligence. The result is a community of practice where employees stay informed, share insights and connect with colleagues working on similar initiatives. Other business units within Proximus have begun replicating the model.
What are the external skills that Euranova bring you?
Euranova’s support extends far beyond community building. They provide profiles capable of bridging business and technical worlds, especially the crucial role of analytics translator. If you put business experts and data scientists in the same room, they don’t always manage to understand each other. The translator helps both sides articulate their needs and challenges, and ensures we capture the full value of a project. Such profiles are rare and difficult to recruit. Euranova helps fill that gap. Without that hybrid expertise, our projects would not move forward at the same pace. They also bring methodological coinnovation, especially in emerging domains like generative AI. On the MAIA project, for example, we defined the evaluation method together. On top of this, Euranova provides targeted expert input whenever the team faces highly technical questions. Sometimes it only takes a short expert report to unblock a situation or challenge an assumption.
Profiles like analytics translators are rare and difficult to recruit. Euranova helps us fill that gap.
How has the collaboration evolved over the past three years?
The relationship has grown alongside the technology itself. We learned together; we adapted together. Euranova is not just a consulting provider but a partner embedded in our transformation. Sophie, the first Euranova consultant to join our team, has become a central piece of our AI project work, bringing ideas, structure and expertise. Her contribution helped Proximus understand the true value of this hybrid role between data scientict and marketeer. Over time, other Euranova collaborators were added to support us, confirming the strong fit between Proximus’ needs and Euranova’s expertise.
Sophie, analytics translator aka bridge-builder
Sophie van Ravestyn is our engineer on a mission at Proximus. She started as project manager within Euranova and then found her marks as analytics translator. She plays a critical bridging role between business stakeholders and technical teams by translating operational needs into data science opportunities and ensuring that AI solutions remain aligned with business value. « Together, we ensure that language models are trained on the most relevant content. We guide model evaluation in business terms and support user adoption by making complex AI outputs understandable, actionable, and trustworthy, » explains Sophie. “Through the communities of practice and ongoing exchanges, we all contribute to a shared body of knowledge. It creates a positive feedback loop that benefits everyone,” adds Sophie on the mutually enriching collaboration with Proximus.