Solution
The mission led us to rethink what change management means in the age of AI.
Deploying AI is not like rolling out a standard tool or a target process through a top-down program. AI starts much closer to the ground: a business pain point, a task that could be automated, a complex decision, or a situation in which teams feel they could work better.
EMERGENCE OF USE CASES
We organized the emergence of use cases around personas drawn from both core business activities and support functions. For each persona, workshops brought together a small group of employees, supported by a pair of experts combining change and adoption expertise with technical AI expertise.
That combination was the strength of the approach: bringing together those who understand the reality of the work with those who can translate a business need into an actionable solution — a prompt, an agent, a secure internal tool, or a specialized solution.
This approach built a portfolio of more than 100 use cases, which were then prioritized according to business impact, complexity, technical maturity, change effort, and potential for scale. Fourteen initial use cases were selected for deployment across business units.
During the mission, several use cases showed how much AI value is created in very concrete business situations. In customer-facing and sales roles, for example, AI can help teams retrieve relevant and personalized information faster, consolidate the data needed at the moment of interaction, and free up time for higher-quality advice. The value is not limited to productivity gains: it also affects the quality of the customer experience, the relevance of the recommendation, and the potential for additional sales.
An AI portfolio must therefore be managed dynamically, not like a traditional roadmap. What is not a priority today may become one tomorrow, depending on business maturity, the pace of technological development, data availability, or a use case’s ability to prove its value in real work.
This was one of the project’s key lessons: AI is not only a productivity lever. It can become a decision-making capability in times of uncertainty.
SCALING
The next challenge was scale: how do you turn a use case built with a handful of employees into a practice that can be adopted quickly by teams around the world?
AI adoption cannot simply be mandated. It has to prove itself in the flow of work. People adopt a prompt, an agent, or a solution when they immediately see its usefulness: faster analysis, a sharper brief, a more reliable simulation, a task avoided, a decision better prepared.
Adoption was therefore organized through communities of champions and peer-to-peer sharing. When a use case worked in one business unit, it was shared with teams doing the same type of work elsewhere in the group. Not as a top-down instruction, but as a concrete proof point: this is what it changes, this is what it enables, this is how you can make it your own.
When a use case proves its value, adoption can move fast.
AI change management relies on networks, business communities, collective intelligence, and informal relays that help promising practices circulate close to the field.
IMPACT
By the end of the mission, the group had its first set of group-level AI use cases, shared criteria to assess business value, and a diffusion model supported by champions and business communities.
The group had also gained velocity: the ability to detect faster what matters, decide faster what deserves to be accelerated, and circulate faster what works. It had strengthened its ability to deal with uncertainty by using AI to simulate, compare, and inform decisions when the business context shifts abruptly.
In the age of AI, leading change is less about pushing a tool than about organizing the emergence, proof, and circulation of use cases. It is about enabling the intelligence of the field to meet the right technology — and making that value travel quickly through the organization.
