The challenge
The leadership team of a mid-sized Swiss AG had read every report on generative AI and attended several conferences, yet six months later nothing had landed inside the business. Department heads were running disconnected experiments with consumer chat tools, IT was nervous about data leakage, and the executive committee could not see how to translate enthusiasm into measurable outcomes without taking on uncontrolled risk.
What they needed was not another vendor pitch. They needed a structured, vendor-neutral process that would identify the few use cases where AI could move a real KPI, weigh those against operational and compliance constraints typical for a Swiss enterprise, and equip an internal core team to carry the work forward without ongoing external dependency. The brief to our digital transformation consulting team was deliberately concrete: deliver a backlog the executive committee could approve and a team that could execute.
Our approach
We ran the engagement as a focused eight-week workshop programme. Week one was a discovery sprint across the five largest functions, mapping current workflows, data availability and the specific moments where staff were waiting on information or repeating the same tasks. From roughly forty candidate use cases we shortlisted twelve and scored each on expected business value, data readiness, regulatory exposure and time-to-pilot.
In weeks two and three we evaluated tools against the shortlisted use cases, comparing hosted Swiss-friendly options, EU-region cloud services and self-hosted open-source models. Selection criteria were not driven by hype but by data residency, audit trail, integration cost and total cost of ownership over twenty-four months. Output was a one-page recommendation per use case that the CIO could defend in a board setting.
Weeks four through seven moved to hands-on AI integration consulting: two pilot workflows were built end-to-end with the customer's own data and the customer's own people pairing with our engineers. The final week was reserved for enablement — a working playbook, an internal prompt library, governance guard-rails, and a coaching session for the newly formed AI core team of six employees.
Outcome
The executive committee received a prioritised backlog of seven approved pilots with business cases, owners and clear stop or scale criteria. Two pilots had already shipped and were generating measurable internal time savings of at least forty percent on the targeted task within the first quarter. Just as importantly, the core team of six left the programme able to run their own discovery and evaluation cycles.
One year on, the client continues to add use cases on the same scoring framework. The engagement is referenced internally as the moment AI moved from a slide to a line item in the operating plan.
Stack used
Duration
8 weeks