Adaptive learning sounds great in a product brief and uncomfortable in a works-council meeting. The pitch is that each employee sees training calibrated to their individual risk profile. The concern is that the platform builds a behavioural dossier on every employee.
Both perspectives are valid. The reconciliation is in the implementation, and the implementation choices most vendors make are different from what we believe is sustainable for the European market.
There is a reasonable middle ground. Personalisation can be cohort-level rather than individual — finance versus engineering versus executive — without singling people out. Risk segmentation can be derived from role and historical cohort performance without involving any monitoring of day-to-day behaviour. Individual-level personalisation can be available, with explicit consent and explicit opt-in defaults rather than opt-out.
We default the platform to cohort-level segmentation. Individual-level personalisation is a tenant setting that requires explicit configuration; the platform does not turn it on silently because the customer wants "more personalisation".
The compromise on training effectiveness is smaller than vendors typically claim. Cohort-level adaptive paths capture most of the benefit of personalisation; the marginal benefit of individual-level personalisation is real but modest, and the works-council friction is substantial.
Where individual-level personalisation is wanted, we provide the customer with template documentation for the works-council consultation in each of the major European jurisdictions. We accompany the consultation process through our customer-success team. We have completed five such consultations in 2025 and have lost zero of them; the conversations are productive when the platform behaviour is defensible.
More vendors should take this approach. The default-opt-in posture works in markets where works-council friction is minimal; it does not work in continental Europe, and pretending otherwise produces sales cycles that fail late.