Plan an AI pilot using a customer-service case from a financial-services organisation. Work out who provides the required evidence, when an approval decision is needed and which preparatory tasks can already begin.
In the financial-services case, AI is proposed to prioritise customer service requests. Existing risk approval requires evidence of data quality and human review before a pilot. Assign evidence, deadlines and decision rights, and identify preparatory work that can begin. Reassess the process if internal sorting becomes a customer-facing decision.
Take away an implementation board with responsibilities and control handovers. The responsible authority decides on approval; specific legal obligations depend on the jurisdiction and are reviewed separately.
What you will work on
Define the business objectives and control needs of related AI initiatives
Assign evidence requirements and the involvement of control functions to decision points
Coordinate preparatory work and required risk approval in the case using existing procedures
Who it is for
Transformation and programme leaders
Leaders in financial, healthcare or infrastructure organisations
Risk and compliance managers
Topics
Programme mandate and intended use
Existing controls and responsibilities
Evidence requirements and approvals
Dependencies between pilot and operation
Changes in intended use and regulation
Your result
Your implementation board shows when evidence and resources are needed and which control role receives the handover. Use it to coordinate through the existing approval process.
Prerequisites
Experience of organisational change and access to information about related initiatives and their control functions sufficient for planning. An anonymised or synthetic organisational case can be used.
Format and approach
The self-paced course is planned for 10–12 hours. Develop a plan and revise it as the work progresses.
An AI tool is meant to sort incoming requests. Employees demonstrate that urgent cases are classified as routine and no timely second review is in place. Assess their evidence and change the workflow before planning further rollout.
If customers can increasingly produce reports themselves using AI, a provider must reconsider its business. Should it offer existing reports more cheaply or develop a new service with ongoing expert-reviewed insights? Compare the options and test a work sample for the new service.
A service agent promises high value but needs data approval that is not yet available and the same scarce specialist as other initiatives. An internal search tool can be tried in a limited way with approved documents. Decide which AI initiative should start first.