Imagine a bank in Libreville deploying an artificial intelligence (AI) assistant to help its advisors respond to clients faster. Three months later, messages are sent more quickly. However, complaints are increasing. Why? No one specified which responses should be reviewed, which files should be escalated to a manager, or how to measure service quality.
The same scenario can play out in a university, a small or medium-sized enterprise (SME), or an administration. The tool is there, teams are trying it out, but the organization hasn’t really changed. This is where management becomes crucial.
Recent articles
Rio Tinto accelerates five mining startups, the proof remains to be seen
Offshore: Oxy and Valaris awarded, HSE AI must remain human
MOL solar and battery branch, the cost remains to be proven
Certificate
Professional Certificate in Governance and CSR — CSR Director
What is it, concretely?
The idea is simple: an AI can help draft, summarize, classify, or analyze certain information faster, but it does not decide for you the best way to work. If a company maintains the same validation circuits, the same gray areas, and the same poorly chosen indicators, AI is likely to only accelerate the existing disorder. The key point is therefore not just to buy software, but to review tasks, roles, controls, and training. Nothing insurmountable: an organization can proceed cautiously, starting with a few simple, measured, and framed uses.
Concrete case: what to do and what not to do
Questions to Ask Before Acting
- What specific task do we want to improve with AI?
- Who remains responsible if the tool produces an inaccurate or inappropriate response?
- What data are we allowed to input into the tool?
- What results will we track: time, cost, quality, client satisfaction?
- Do managers know how to explain the use of AI to their teams?
- Is a testing phase planned before generalizing the use across the entire department?
- What do we do if employees are already using unauthorized tools?
French 
English 