The Next Enterprise Transformation Must Be Cultural, Not Technical
The bottleneck is culture, not technology
GenAI tools hold the potential to dramatically reduce friction and cognitive overload. The real question is whether enterprises can evolve their traditional practices quickly enough to bring that potential to life.
- Cognitive load: “The total amount of mental effort used in working memory at any given moment.”
-
Friction:
“The resistance force that opposes motion between two surfaces in contact.”
In organisational terms: any process, gap, or dependency that slows progress toward a goal.
Imagine you are sitting in a room. Someone walks in and asks you to build a system that can predict human behavior, as it pertains to individuals rather than segments, and deliver recommendations at scale, in real-time.
In most enterprises, this can take years. Why? Because the concept of tracking human behavior and real-time AI is relatively new, and enterprise systems are not built to scale in a way that can accommodate individualized personalization.
The task seems gargantuan because the cognitive distance between the words they just heard and the knowledge they would need to act is enormous. The paralysis that ensues is the product of two forces: friction and cognitive overload.
Cognitive load theory, first developed in the 1980s, distinguishes between the effort required to process new material and the effort wasted on navigating unnecessary complexity.
The insight that matters here is that much of the mental effort people expend in daily life (and especially in organizational life) is not directed at the actual problem but rather at the scaffolding around the problem — understanding the jargon, waiting for approval, reconstructing context that was lost in a handoff.
Friction, in the context of human behavior and organizational systems, operates in a similar way: it is everything that sits between intention and action.
Some friction can be useful. It produces the deliberate pause before a consequential decision, forcing critical thinking and clarity. But the friction that permeates most large institutions is not of the useful kind. It is the friction of inertia, diffused accountability, and systems so complex they consume more energy to navigate than they return in value.
Friction and cognitive overload are the reason a major South African bank can spend R800 million and four years attempting to implement a CRM system and still not go live. They are the reason enterprises accumulate 1,800 legacy technology systems. They are the reason a data-pipeline project quotes two years and R120 million, when, with today’s available technologies, four days and a fraction of the budget would do.
The Cost of Waiting
Enterprise culture is, in many ways, a system engineered to produce waiting. Someone needs approval from someone else, who is waiting on someone else, who has escalated to a committee. Each handoff adds friction and compounds cognitive load.
Reducing cognitive load and friction is the core problem that generative AI is built to solve. When well-designed, AI does not just automate tasks, but removes friction between intention and execution so completely that the cognitive distance collapses to near zero.
However, if enterprise culture cannot align to foster the true potential of GenAI, its capabilities may be limited to personal productivity tools rather than organizational-level projects.
Consider what it takes today to build an Accounts Payable (AP) system. First you must understand the domain well enough to document what you need. Then locate a vendor. Then evaluate options. Then design tests. Then negotiate procurement. Then onboard developers, architects, and engineers. Then establish standards. Then wait (possibly years) for something usable to emerge.
“If enterprise culture cannot align to foster the true potential of GenAI, its capabilities may be limited to personal productivity tools rather than organizational-level projects.”
The Promised Land
Now consider what generative AI makes possible: you say what you want, in plain language, and it exists. The domain knowledge is embedded. So the friction disappears and the cognitive load drops to near zero — and suddenly the organisation can focus on the only thing that actually creates value: deciding what to build, and why.
Old Model of Work
- Procedural execution
- Managing handoffs and approvals
- Navigating complex tools
- Waiting on dependencies
- Cognitive load as the default state
New Model of Work
- Problem framing and intent
- Critical evaluation of outputs
- Domain expertise and judgment
- Creative direction
- Psychological ownership
Generative AI holds the potential to shift cognitive load and friction away from the practicalities — the question “How do we build this?” — to a more important one: “Do we understand our business problem well enough to describe what we need?”
Conclusion
Incorporating generative AI to reduce friction and cognitive load in meaningful ways requires redesigning the culture in which the interface lives, creating environments where curiosity is rewarded, where independent judgment is trusted, and where the distance between a question and an answer is not a career risk.
Generative AI does not eliminate the hard work of thinking, but rather strips away everything that was mistaken for it.
It is essential that enterprise culture adapts to the new possibilities afforded by generative AI. Otherwise, it may well be the case that the lack of tools is no longer the cause of stifling friction and cognitive load — but the cultural system itself.
