The situation
The hypothesis deserved a test, but not at the price of turning the main business away from its own work. Running the pilot inside the existing systems would have spoiled both: the experiment by the old constraints, the company by tools nobody had agreed to keep.
What was built
A separate contour in the cloud, on Kubernetes, holding the pilot's own stack: a cloud CRM, its databases, the new digital marketing platform. Two people moved into it with a KPI set of their own and ran the new business model for three months, without a single change to the systems the company lives on.
The result
A clean experiment instead of an argument: what the new platform really demands of IT, what it really costs per month, and a migration plan for the rest of the company that does not require stopping the business to carry it out.
Kubernetes · Managed cloud services · Cloud CRM · Cost modelling · Migration planning