AI is everywhere in the company, but rarely managed. Between scattered experiments, vague governance and promises of never-measured gains, the real subject is no longer technology
Artificial intelligence is progressing at high speed in businesses. Too quickly, sometimes. As tools multiply, a question now arises for managers: who is really driving the AI strategy?
According to an MIT study, only 5% of companies that have deployed AI solutions are now seeing measurable productivity gains. Conversely, Goldman Sachs estimates that organizations capable of structuring their uses can achieve 15 to 25% gains. This large gap is not explained by the technology itself, but by the way in which AI is controlled… or not.
AI everywhere, but rarely controlled
In the majority of companies, AI has spread opportunistically. Marketing is testing content generation, HR is automating certain processes, sales teams are tinkering with their own assistants, finance is exploring predictive analysis. These initiatives respond to real needs, but they are rarely coordinated.
The result is now well known: shadow AI. Fragmented AI, sometimes insecure, often unmeasured, whose return on investment remains unclear and the risks are very real: legal, human, organizational. AI is not absent from businesses; she is disorganized.
The blind spot of governance
This situation is due to a persistent blind spot: AI has not yet found its place in governance. The IT department manages the infrastructure, the business lines manage uses, HR is concerned about human impacts, and general management arbitrates without always having a consolidated vision.
However, AI is neither a simple technological subject nor one innovation project among others. It simultaneously affects strategy, operations, data, skills and decision-making methods. Treating it in silos is like flying a plane where everyone has access to the cockpit.
The Chief AI Officer, a steering role
It is in this context that a still recent but increasingly necessary role emerges: the Chief AI Officer. Contrary to some popular belief, this is not a tool expert, nor a senior data scientist. Its role is above all strategic and organizational.
The Chief AI Officer’s mission is to define an AI vision aligned with business challenges, to prioritize use cases with real impact, to structure clear governance, to coordinate initiatives to avoid dispersion and to fight against shadow AI. Its objective: to bring about a truly operational, measurable and useful AI.
A key function, including timeshare
Not all companies are looking to recruit a full-time Chief AI Officer. For SMEs and mid-sized companies, the time-sharing Chief AI Officer model appears to be a pragmatic response. It provides access to a strategic vision, arbitration capacity and cross-functional management, without organizational rigidity.
Even in large groups, this role can be part of a temporary logic: structuring an AI trajectory, securing a transformation phase or supporting a management committee before internalization.
From gadget AI to operational AI
The real challenge is no longer to multiply experiments, but to make AI operational. Useful AI is not measured by the number of tools deployed, but by its ability to reduce friction, improve the quality of decisions and free up time on low-value-added tasks.
Without management, AI remains a stack of tools. With clear governance, it becomes a lever for sustainable performance.
As AI becomes established in all layers of the enterprise, one question becomes unavoidable: is there a driver in the AI strategy?
The power of algorithms will depend less on the answer than on the ability of organizations to maintain control.
