With each new wave of technology, businesses focus on the tools before focusing on themselves. They invest time in use cases, models, and demonstrations, hoping that a successful deployment will naturally have an impact on their activities and performance. In doing so, they think they are preparing for AI when they have not yet clarified how they operate. They install an innovation in an environment that was not designed to accommodate it, and when we look at what happens next, things quickly go off the rails. We then tend to blame the technology, when in fact it is the business that has failed to adapt itself to what it is trying to integrate.
A long time ago, I quoted Goldratt’s expression that “We should not expect an application to work in environments for which its assumptions are not valid“ when talking about collaborative tools, but this applies just as well to AI, as we will see.
In short:
- Businesses invest in technological tools such as AI without clarifying their own operations, leading to ineffective integrations and failures attributed to technology rather than a lack of organizational preparedness.
- Enterprise design aims to align identity, experience, and operational architecture, which is necessary for AI to truly support an ambition and integrate coherently.
- An organization must clearly structure roles and dependencies to enable technology to have a global impact and prevent gains from remaining isolated.
- Work design identifies the concrete sequences on which AI can build, clarifying the sequence of actions and reducing friction in daily practices.
- AI acts as a revealer of organizational shortcomings, it accelerates awareness of existing inconsistencies and requires a rethinking of the business before considering technological deployment.
Enterprise design: aligning identity, experience, and architecture to embrace AI
Enterprise design is not about drawing up a theoretical diagram or reorganizing teams, but about giving coherence to what the business wants to be, what it wants to offer, and how it organizes itself to deliver on that promise (EDGY: a common language to align identity, experience, and operations). AI cannot support any ambition if the identity of the business remains unclear, if the experience it seeks to deliver is not precisely defined, or if its operational architecture does not reflect these choices. An organization that does not know what it wants to achieve cannot decide what it wants to entrust to technology.
Alignment between identity, experience, and architecture is not a luxury but the foundation on which the ability to integrate innovation rests. When a business has not clarified what it wants to be or what it wants to offer, its activity is organized around local interpretations or even imposed by technology rather than around a shared intention (Efficiency vs. uniqueness: the false dilemma of operations). And when it has not given precise form to what it seeks to deliver, it cannot identify where technology can really lighten the load or improve practices (How management let systems do the thinking for them). AI only highlights this lack of consistency and cannot compensate for or invent what the business has not formulated. AI never compensates for these weaknesses, but exposes them.
Roles and dependencies: the operational foundation without which AI cannot function
Beyond this alignment, a business must be clear about the articulation of its roles and how dependencies structure its operations. When responsibilities are unclear, coordination is lacking, and everyone compensates for the system’s imperfections without addressing their root causes (Work about work: when the reality of work consists of making things that don’t work work), no technology can be properly integrated. AI requires stable ground, not an uneven foundation where everyone is tinkering to compensate for the system’s imperfections.
Clarifying roles is not a conceptual exercise, as it is a prerequisite for redistributing certain tasks, reviewing the balance between teams, and identifying what can be supported entirely or partially by technology. When a dependency or bottleneck is not taken into account, the gain remains blocked where it was obtained. The team makes progress locally, but the rest of the business continues to operate as before without any noticeable improvement (Local optimum vs. global optimum and the theory of constraints: why your productivity gains sometimes serve no purpose and Collective appropriation of AI: the only condition for tangible impact). The benefit is real, but it has no impact on an activity or workflow taken as a whole: it is simply “locked in” because what is happening around it has not been adjusted.
AI draws attention to what we did not want to see and shows what is missing for the business to embrace the new.
Work design: clarifying sequences before transforming them
While enteprise design provides an overall framework, work design shapes day-to-day activities. It involves understanding how actions are linked, how teams collaborate, and identifying tensions and friction points to see how a sequence can be rewritten to improve the functioning of individuals and the group. Without this, AI cannot find a foothold, as it must be based on specific practices, not vague intentions that are not reflected in the reality of the work.
Many organizations are surprised to see individual use progressing without producing a collective effect. A team can achieve results in a matter of moments, but if the next step does not know how or cannot take advantage of them, nothing progresses. A task can be simplified, but if the flow does not adapt, this simplification has no impact. Work design helps to identify exactly where technology can lighten the load, where it can help and where it can eliminate it altogether, and where it would be inappropriate.
AI as a revealer of shortcomings in organization and work
AI is often perceived as a lever, but in reality, it acts as a revealer. With its acceleration potential far superior to the “productivity” tools that preceded it, it reveals a company’s shortcomings much faster than previous technologies. It exposes inconsistencies between identity, experience, and architecture. It highlights poorly defined roles, unknown dependencies, and poorly designed workflows.
This accelerated exposure of the organization’s limitations gives the false impression that a new difficulty has just appeared, when in fact this is not the case. AI simply accelerates awareness, and what was previously hidden or accepted immediately becomes an obstacle.
AI forces us to look at the business and the work before looking at the tool, which is rarely done, regardless of the type of technology involved (Businesses invest in technology, not in work).
Bottom Line
Ultimately, AI does not transform businesses or work. It requires businesses to redesign themselves and work to be clear enough to accommodate what technology has to offer (If your business isn’t designed for AI, it will end up being designed by AI). It highlights the importance of consistent enterprise design and precise work design. Organizations that move forward are not those that deploy the fastest, but those that take the time to shape what they do.
To answer your questions…
AI is often disappointing because businesses focus primarily on tools rather than their own operations. When identity, user experience, and architecture are unclear, technology becomes part of an inconsistent system. AI then reveals these weaknesses instead of creating value. Without a shared purpose or clear framework, it can neither guide decisions nor improve practices. Clarifying what the business wants to be and how it operates is essential to avoiding these dead ends and preparing for a truly useful deployment.
Enterprise design serves to align identity, experience, and architecture to give the business a coherent form. Without this alignment, AI has no reference point to support the business’s ambition. A vague identity or poorly defined experience leads to scattered and ineffective use of technology. An unadjusted architecture then blocks gains. By clarifying these three dimensions, the business determines what it wants to achieve and where technology can really help, ensuring relevant deployment rather than simply stacking tools.
AI can only be properly integrated if roles, responsibilities, and dependencies are clearly established. When everyone compensates for the system’s shortcomings or coordination is weak, technology only exposes these flaws. Clarifying roles makes it possible to identify which tasks can be redistributed or automated. Understanding dependencies prevents gains from remaining localized. This review creates a stable foundation where AI-driven improvements can truly spread throughout the business.
Work design involves understanding precisely how actions are linked and where friction and tension lie. Without this vision, AI cannot be based on real practice. Local simplification has no effect if the next step cannot benefit from it. By clarifying sequences and collaborations, the business identifies exactly where technology can lighten, speed up, or eliminate a task. This makes it possible to adjust flows and achieve a collective impact rather than just individual improvement.
AI speeds up activities and immediately reveals inconsistencies that previously went unnoticed. It exposes unclear identities, poorly defined experiences, ambiguous roles, and faulty workflows. These limitations already existed: AI simply reveals them more quickly. This impression of novelty masks a simpler reality: businesses must first clarify their organization and work before they can take advantage of the technology. It is this consistency that determines any real impact.
Image credit: Image generated by artificial intelligence via ChatGPT (OpenAI)
Related articles:
How can you prepare your organization for AI?
How management let systems do the thinking for them
AI adoption: methods put to the Test
Businesses invest in technology, not in work
AI First is not AI Only: clarify your intentions before transforming your business
Generative AI in the workplace: revolution or illusion?
