The difficult part of AI is not starting. It is getting through the parts of the journey that cannot be skipped.
Most organizations can launch experiments. Far fewer can turn them into reliable, adopted, production-level value. IBM’s 2025 CEO study found that only 25% of AI initiatives had delivered expected ROI, and only 16% had scaled enterprise-wide. At the same time, 64% of CEOs said the risk of falling behind pushes them to invest in AI before they have a clear understanding of the value.
That is the pattern: urgency first, learning later.
McKinsey’s 2025 research helps explain why the journey is longer than many leaders expect. It found that workflow redesign has the biggest effect on whether organizations see EBIT impact from gen AI, but only 21% say they have fundamentally redesigned at least some workflows.
That is the critical insight. The early phases of AI are often about testing possibilities. The later phases are about redesigning how work actually gets done:
- decisions
- handoffs
- controls
- accountability
- escalation paths
- human-AI interaction
- monitoring and learning
Those steps are slower. They involve more friction. They are harder to present as quick wins. But they are exactly where value starts to become real.
So the journey matters because the biggest gains usually do not come from standing up another pilot. They come from doing the harder work that sits later in the journey and makes value durable.