What if the next AI risk for businesses was no longer just moving too slowly, but moving faster than the organization itself?
In the sample studied by the Boston Consulting Group, 61% of CEOs surveyed believe their board of directors is pushing artificial intelligence transformation too quickly. This finding comes from the global study CEOs and Boards Align on AI in Theory but Divide in Practice, published on May 4, 2026, and conducted among 625 executives, including 351 CEOs and 274 board members.
This figure does not reflect executive resistance to artificial intelligence. Rather, it highlights a deeper problem: a possible gap between the speed at which technological capabilities advance and the speed at which organizations can actually absorb, govern, and convert them into performance.
We have talked a lot about the risk of not adopting AI. Perhaps we now need to consider its symmetrical risk:
adopting faster than we transform.
A company can quickly purchase solutions, multiply experiments, and launch dozens of projects while remaining unable to convert this new technological capability into sustainable economic results.
The problem appears when technological speed durably exceeds the organization’s transformation capacity.
The risk of inaction has not disappeared
However, it would be wrong to conclude that companies should slow down.
In its AI Radar 2026, based on a survey conducted among 2,360 executives across 16 markets and nine sectors, BCG observes a strong increase in investments dedicated to artificial intelligence and growing CEO involvement in defining AI strategies.
Artificial intelligence is no longer just a technological subject. It now affects processes, business models, skills, customer relationships, knowledge management, and decision-making. The arrival of AI agents reinforces this organizational dimension: the more systems acquire action capabilities, the more questions of responsibility, control, and integration into processes become structural.
The risk of inaction therefore remains real. But considering that any acceleration is necessarily positive constitutes another risk. Leaders must move fast enough not to undergo transformation, while maintaining enough discipline not to commit their company to a succession of projects whose value, governance, or industrialization capacity remain uncertain.
The question then becomes less “Are we moving fast enough?” than:
“Do we know why we’re accelerating?”
When FOMO enters governance
The BCG study reveals a particularly interesting result: board members who consider themselves less competent in AI matters than their peers are also more likely to fear that their company is not adopting the technology fast enough. BCG explicitly links this situation to a form of FOMO, this fear of missing a major transformation.
The phenomenon is understandable. C-suites and boards of directors are daily exposed to new models, new agents, spectacular investments, and announcements of productivity gains. In this environment, immobility seems dangerous.
But this result opens a broader governance question:
can the fear of being left behind eventually become itself a decision criterion?
Constant comparison with other organizations can subtly modify the starting point of strategic reasoning. Instead of asking “What problem do we need to solve?”, the company begins to ask “Why aren’t we doing what others are doing yet?”
These questions lead to very different strategies. The first starts from the company’s priorities; the second, from the dynamics of the technology market. In one case, AI becomes a means to serve an identified transformation. In the other, technology risks becoming itself the justification for change.
A company can move fast and transform very little
We still too often confuse deployment speed and transformation speed.
An organization can distribute several thousand licenses of a generative assistant, launch a multitude of pilots, and experiment with AI in almost all its functions. This does not necessarily mean it is transforming its operational model.
Transformation truly takes shape when AI modifies the way a process works, a decision is made, a responsibility is organized, or economic value is created.
At this stage, difficulties change in nature. Data quality, integration with existing systems, access rights, cybersecurity, control mechanisms, human responsibilities, skills, and criteria for evaluating results must be addressed.
Technical deployment speed is therefore not a sufficient measure of an organization’s AI maturity.
The use case collection trap
The first years of generative AI have naturally been those of experimentation. Marketing, finance, human resources, customer service, IT, legal, operations: each function has sought its applications. This exploratory period was essential.
But it is now reaching a limit.
An organization does not transform because it has several hundred AI use cases.
It can even accumulate independent pilots, different vendors, multiple architectures, and dispersed responsibilities without having a sufficiently precise vision of the value produced.
The maturity shift occurs when the question evolves. It is no longer just about asking:
“Where can we use AI?”
but:
“Which processes do we need to rethink because AI now makes it possible to organize them differently?”
The first question produces possibilities. The second imposes choices. It requires prioritizing opportunities, concentrating investments, transforming certain processes in depth, and abandoning other initiatives.
The next phase of AI in business could therefore be less that of generalized experimentation than that of strategic discipline.
Moving from experimentation to transformation then leads to another question: under what conditions can an organization truly accelerate?
Four dimensions appear decisive: value, organizational capacity, governance, and adoption.
Before examining them, two tensions highlighted by BCG help understand why this question is becoming urgent: governance and return on investment.
The board is neither the accelerator nor the brake
BCG observes a significant perception gap between CEOs and board members on understanding artificial intelligence and on the appropriate speed of transformation. In a publication devoted to the role of boards of directors, the firm emphasizes the need to focus the AI agenda on capabilities and value pools likely to truly modify the company’s competitive position, rather than multiplying pilots.
This question extends a reflection already developed on Yunova Consulting in the article “AI is no longer an IT subject. It is becoming a board responsibility”: making AI a governance subject does not mean transferring operational transformation leadership to the board. Its role is to exercise informed judgment on strategy, value creation, risks, and the organization’s actual ability to execute.
Read the article on Yunova Consulting
Understanding AI, for a board of directors, therefore no longer simply means following model evolution or knowing the main technology players. One must be able to distinguish what is technically possible from what is economically relevant and organizationally achievable.
This requires understanding where value can be created, what data is necessary, what risks emerge, what operational transformations will be essential, and to what extent certain decisions can be entrusted to automated systems.
The board’s role is neither to systematically press the accelerator nor to become the transformation brake. Its role is to determine where acceleration is truly justified.
Adoption, productivity, and value do not measure the same thing
To this speed question is added that of economic return. The BCG study highlights a gap between CEOs and boards of directors in their perception of expected performance from AI investments. Leaders can thus find themselves simultaneously confronted with two expectations: accelerate and quickly demonstrate that this acceleration produces an economic result.
However, deep organizational transformation does not necessarily follow the timeline of a software deployment. Data sometimes needs to be restructured, processes rethought, skills developed, and responsibilities redefined.
This pressure can favor projects whose results are immediately visible at the expense of more structural transformations.
It especially invites distinguishing three notions too often confused:
adoption, productivity, and value creation are not three ways of measuring the same thing. They are three different levels of transformation.
A high number of employees using a tool measures adoption. Time saved on a task can reveal a potential productivity gain. But neither automatically constitutes value creation.
One must still demonstrate that this improvement translates into costs, revenues, quality, risk reduction, execution speed, innovation, or customer experience.
It is precisely in this conversion that much of the transformation takes place.
The four conditions for AI acceleration
There probably is no universal speed at which all companies should move. An organization can legitimately accelerate strongly in certain areas while remaining cautious in others.
Before accelerating an AI transformation, four questions should therefore be asked.
- Value: do we know precisely why we are using AI?
An initiative should begin with a value hypothesis, not only with a technological possibility.
What result are we really seeking? Reduce a cost? Accelerate a cycle? Improve a decision? Reduce a risk? Increase revenue? Create a new value proposition?
The vaguer the economic or strategic objective remains, the greater the risk of developing a project impressive on the technological level but marginal for the company.
- Organizational capacity: can the company truly absorb the transformation?
A model’s performance is not enough to determine an organization’s ability to exploit it.
AI can help bring together or exploit fragmented data, but it does not by itself correct problems of quality, governance, or data access. Similarly, a poorly defined process remains difficult to automate efficiently, and an ambiguous responsibility can become more problematic when a system intervenes in the decision chain.
The question is therefore not only: “What can the technology do?”, but also: “What are we truly capable of integrating?”
- Governance: who decides, who controls, and who assumes responsibility?
As systems become more autonomous, this question becomes central.
Who authorizes the use case? Who evaluates the risk? Who controls the results? Who determines the acceptable level of autonomy? Who decides that a project should be stopped? Who assumes responsibility when a system intervenes in a critical process?
Effective governance does not consist of systematically adding an additional layer of control. It must enable faster decision-making because responsibilities, rules, and intervention thresholds are clearly defined.
- Adoption: can the business units truly transform their practices?
A technology does not transform an organization because it is available. It transforms it when it is integrated into daily practices, decisions, and processes.
This requires skills, but not only. Employees must understand what changes in their role, where the system’s limits are, what responsibilities remain human, and how cooperation between people and technologies evolves.
Adoption is not the final step of a technology project. It is part of the transformation’s business model.
From these four conditions follows a simple rule:
An organization should not accelerate at the pace of technology. It should accelerate at the pace of its ability to convert technology into value.
AI maturity could be measured by the ability to say no
This evolution also leads to revising our definition of maturity.
For several years, it has often been associated with the number of projects launched, investments committed, or number of equipped employees. These indicators remain useful, but they are no longer sufficient.
A mature organization is perhaps also one that knows which projects it will not launch, one that stops an experiment when its value is not demonstrated, and one that distinguishes a spectacular prototype from a truly industrializable process.
It knows how to differentiate a technically possible automation from a strategically desirable automation.
As technological capability becomes abundant, selection capability becomes strategic.
When almost everything becomes testable, scarcity changes in nature. What becomes rare is no longer just technology. It is management time, skills, relevant data, execution capacity, and the organization’s collective attention.
Maturity therefore no longer consists only of knowing how to accelerate. It consists of knowing where to accelerate.
Artificial intelligence creates capability. Organizations create value.
Artificial intelligence rapidly expands the field of possibilities. It can produce, analyze, search, predict, recommend, and progressively participate in executing certain actions.
But a technological capability is not yet value creation.
Between the two lies the organization.
It is companies that choose the problems to solve, define priorities, structure data, organize responsibilities, transform processes, develop skills, and determine the respective place of automation and human judgment.
Artificial intelligence creates capability. Organizations create value.
During the first years of generative AI, the dominant risk was that of inaction. This risk has not disappeared. But a second one appears: confusing technological speed with transformation speed.
The next competitive advantage will probably not come from the company that adopts each innovation first. It could belong to the one that knows where to accelerate, where to experiment, and where to give up.
The question is therefore no longer only:
“Are we moving fast enough?”
It becomes:
“Is our organization truly capable of converting this speed into value?”
Main references
Boston Consulting Group, CEOs and Boards Align on AI in Theory but Divide in Practice, May 4, 2026. Global study conducted among 625 executives, including 351 CEOs and 274 board members.
Boston Consulting Group, BCG AI Radar 2026: As AI Investments Surge, CEOs Take the Lead, January 15, 2026. Survey conducted among 2,360 executives across 16 markets and nine sectors.
Boston Consulting Group, Five Things Boards Need to Get Right with AI, February 24, 2026.
Boston Consulting Group, The CEO’s Guide to Closing the AI-Knowledge Gap with Your Board, August 4, 2026.
