AI Doesn't Fail Because of Technology. It Fails Because of Misplaced Friction

Artificial intelligence is not failing because it lacks capability. Rather, many organizations are struggling because they are removing the wrong kinds of friction from their systems.

Across industries, leaders are investing in AI, launching pilots and generating more data than ever. Yet results remain inconsistent. Some initiatives stall, while others produce activity without delivering meaningful impact.

The typical explanations are familiar. The data is not ready. The tools are immature. The organization needs more time.

While each of these may contain some truth, they miss the underlying issue. AI is not exposing a technology gap. Instead, it reveals a structural problem in how organizations think, make decisions and align their efforts. At its core, this is what I would call misplaced friction.

Friction Isn’t the Enemy

In operations, we are trained to remove friction. Our tools focus on eliminating waste, streamlining flow and reducing delays.

However, not all friction is waste. In many cases, it is essential.

There is a form of tension that improves outcomes. It shows up in the debate that sharpens a strategy, the questioning that clarifies a problem, the back and forth that builds alignment and the discomfort that accompanies learning and change.

Lean practitioners recognize this instinctively. It is present in kaizen events, embedded in A3 thinking and reinforced through the dialogue of catchball—structured, two-way conversations that give employees a voice and clarify how their work supports bigger objectives. It is also reflected in how teams develop over time.

Groups rarely move directly to high performance. They progress through stages often described as forming, storming, norming and performing, where early tension and challenge give way to alignment and results.

This same principle appears more broadly in how growth occurs. Development often requires resistance. There is a reason a butterfly must struggle to emerge from the chrysalis. Without that effort, it does not develop the strength required to fly.

This type of friction is productive. It is what turns activity into learning and learning into results.

The issue is not whether friction exists, but where it exists within the system.

The Failure Mode: Misplaced Friction

In many AI implementations, organizations unintentionally reverse this balance.

They remove friction from the thinking while leaving it embedded in the process. Decisions are automated before the problem is fully understood. Analysis is accelerated without alignment on purpose. Insights are generated without building shared understanding.

At the same time, the day-to-day experience of work remains difficult. Data is fragmented. Workflows are unclear. Communication is slow. Ownership is ambiguous.

The outcome is predictable. AI makes things faster, but not better, leaving teams busier without becoming more effective.  Decisions happen more quickly, but with less clarity.

AI does not fix broken systems. It amplifies them.

This dynamic shows up clearly in practice. In one organization, a team set out to use AI to automate reporting with the goal of reducing manual effort and accelerating insights. Instead, much of their time was spent cleansing data, reformatting inputs and reconciling inconsistencies across sources. The underlying process had never been fully defined, and expectations for how the data would be used were not aligned.

Rather than simplifying the work, the introduction of AI increased the pace at which confusion surfaced. Reports were generated faster, but the team spent more time questioning their accuracy and debating their meaning. The system moved quicker, but the thinking behind it had not improved.

Simply accelerating a task without first designing the process does not eliminate waste. It replaces delay with chaos.

A similar pattern appears outside of operations. Teams can now generate messaging and content at scale using AI, but when that output is not grounded in a clear voice or aligned with broader strategy, something is lost. The content is produced quickly, but lacks coherence and credibility.

In each case, the issue is not the technology itself. It is the placement of friction. Organizations have made it easier to produce output, while making it harder to ensure that the output is meaningful.

Where Alignment Breaks Down

This is where alignment becomes critical and is often overlooked.

Many AI pilots are launched as experiments, but they are not designed as learning systems. The intent is vague. The problem is loosely defined. Success criteria are unclear. Teams are asked to try something rather than solve something.

In these situations, there is a natural tendency to move quickly in order to demonstrate progress. However, as Brian Joiner observed, sometimes the better course is to pause and think before acting. In his words, “don’t just do something, stand there.” Premature action often creates more variation rather than less.

Without clear alignment, organizations remove the friction that would otherwise force better thinking. What problem are we solving? What decision will this inform? What would success look like? How does this connect to the broader strategy?

Instead of slowing down to clarify these questions, teams accelerate into execution. Activity increases, but accomplishment does not.

In lean terms, this reflects a breakdown in strategy deployment. Hoshin kanri, the strategy deployment process, is not about control. It is about alignment through structured dialogue that connects intent to execution. It introduces intentional friction through catchball, reflection and alignment across levels before action is taken.

When that friction is absent, coherence suffers. AI simply accelerates the misalignment.

 What Good Looks Like

High-performing organizations take a different approach. They are deliberate and intentional about where friction belongs within the system.

They preserve friction in problem definition, root cause analysis, strategic alignment and decision-making. At the same time, they remove friction from data access, information flow, coordination and repetitive analysis.

In this environment, AI becomes a true partner. It reduces noise so leaders can focus on thinking. It accelerates learning cycles without bypassing them. It supports alignment rather than replacing it.

This approach is consistent with lean principles. Technology should improve the system, not shortcut the thinking that makes the system effective.

The Real Opportunity

The conversation around AI often centers on capability. What can it do? How fast can it go? How much can it automate?

A more important question is where friction exists within the system and whether it is in the right place.

When positioned correctly, friction is not a constraint. It is a source of energy that drives better decisions, stronger alignment and more durable results.

A Simple Test for Leaders

If AI efforts are increasing activity without improving outcomes, it is worth stepping back and asking a few simple questions.

• Have we removed friction from the thinking?
• Have we preserved it where it improves decisions?
• Have we made it easier to move faster without understanding more?

The answers to these questions will provide more insight than any dashboard.

Closing Thought

High-performing organizations do not eliminate friction entirely. They are deliberate about where it exists.

They remove friction from the experience of work while preserving it in the thinking that drives it.

When you remove the wrong friction, you get comfort without progress. When you preserve the right friction, you create progress without chaos.