How enterprises can move beyond AI innovation theatre

By Ismail Amla, Senior Vice President, Kyndryl Consult

In short order, artificial intelligence has become the core enterprise ambition — and anxiety. 

Across the global economy, business and technology leaders are racing to harness the technology’s potential to accelerate innovation, simplify the complex, and unlock new value. Yet behind the headline-grabbing announcements and pilot projects, a quiet reckoning is underway. For every successful deployment, many more stall in what can only be described as innovation theatre: a flurry of proofs-of-concept that look impressive on the surface but deliver little lasting impact.

The data tell the story. AI investment continues to surge, with many organizations increasing spending by double digits, according to 2025 Kyndryl Readiness Report. Yet nearly six in 10 projects never make it past the pilot stage. Enthusiasm is real — but so are the obstacles. Many are working with fragmented IT estates, something punctuated by the fact that 70% of CEOs reportedly feel they arrived at their current cloud environment by accident, rather than by design. Add atop that challenge the fact that much data is trapped in silos, and security strategies are often unfit for the speed and scale of machine learning. These are the primary barriers that thwart attempts to move from experimentation to execution.

It’s against this backdrop that enterprise leaders must decide whether AI will remain a collection of pilots, or become a scalable, secure, and measurable capability.

Unready by design

Too often, enterprises are layering AI onto fragmented enterprise systems, all built after years of tactical decisions that were made to solve for evolving business and market challenges. In that sense, AI doesn’t just herald innovation potential, it magnifies architectural shortcomings.

These patchwork IT environments are often not conducive to what AI requires — a connected, flexible, data-rich foundation. Without intentional design, integration, and governance, enterprises risk multiplying complexity instead of value.

The path forward

Breaking free from the pilot trap requires more than bold vision. It demands a strategic rethink and a thoughtful shift to systems that are designed for greater interoperability, observability, and adaptability. Here are three steps leaders should consider, according to findings of the 2025 Kyndryl Readiness Report.

• Invest in plumbing before prototypes. AI ambitions are outpacing the systems meant to support them. With nearly 60% of AI initiatives stalling before scaling, there is a clear need to organize data sources. Organizations who lead in this regard are solving this by consolidating platforms, improving data hygiene, and building observability into their estates. To put a finer point on it, the readiness dividend comes from connected data, interoperable systems, and architecture that’s designed for scale.

• Have the C-Suite sign off on a single AI ambition statement. Readiness isn’t owned by IT alone. It’s a leadership alignment issue. An overwhelming percentage of leaders — some 90% — say they have the right tools to rapidly test and scale new AI ideas, and yet fewer than one-third believe their teams are ready. Leading organizations tether their AI strategies to clear business outcomes and workforce enablement from the start. Aligning around a strategy and then translate it into measurable goals for data, operations, and people will pay off in spades. Alignment turns AI from a technology experiment into a business transformation plan.

• Treat every pilot as a steppingstone to production, not showcase. Investment in AI remains heavy by enterprises across the board, yet 60% of projects never make it beyond the pilot stage. This has led to a lot of innovation theatre. Leaders can solve for this by more methodically identifying operational owners of each project, the integration points, and the KPIs that will define success. Measure progress by outcomes deployed, not prototypes demonstrated.

AI readiness is not a technology milestone; it’s an organizational one. The most advanced enterprises are treating AI as a core capability, one that connects strategy to execution. They measure success not by the number of pilots launched, but by how AI changes how work gets done: how it accelerates product cycles, strengthens resilience, and redefines customer experiences.

When AI becomes part of the enterprise fabric — designed to scale, governed to stay secure, and aligned to measurable outcomes — innovation stops being theatre. It becomes actual transformation.