AI reality check

Three years of AI hype are behind us. The big promises are gradually giving way to the real question: how do you get results from AI without losing control?

That is the question at the heart of the third edition of our AI event, once again at Technopolis in Mechelen. An evening with a sharp keynote, in-depth sessions on security, operationalisation and adoption, and plenty of time to catch up with peers.

We focus on really putting AI into practice and keeping it manageable. We explore that in three sessions:

• AI and security: how do you keep AI secure while you scale it up? In many organisations, employees are already working with ChatGPT, Gemini or Claude, while IT has no idea which company data ends up in them. Bram De Corte (Modern Workplace Expert at Xylos) and Abdelmalek Akkad (Senior IT Consultant at Xylos) show how to make that AI use visible with Microsoft Security and decide which tools are allowed.

• AI and operationalisation: how do you keep the costs of AI in production under control? The bill almost always turns out higher than budgeted. You pay per request to a model, and an agent pays per step. Peter Verrykt, Business Director Professional Services at Xylos, shows where those costs sit and how to keep them low.

• AI and adoption: how do you bring thousands of employees along on a scalable AI learning journey? Johan Peeters, Business Unit Manager Learning Services at Xylos, outlines where to start: first measure where your people stand, then build a programme that fits their role. Then Matthias Meul, Data & AI Capabilities Manager at Ageas, takes the floor and explains how the international insurance group tackled this with a playful AI change programme, which choices they made and where they had to adjust course.