Three ways that Asia’s enterprises are adopting AI—and where they are falling behind

Asia’s AI race won’t be won by early adopters. It will be won by businesses that rebuild processes, data, and governance around AI at scale.

Asia’s boardrooms are full of AI ambition. Over the past three years, companies have launched pilots, tested AI assistants, and explored use cases across nearly every function. Business leaders are no longer asking whether AI works, but instead whether it’s materially changing the economics of their organizations.

The next phase of enterprise AI will expose an uncomfortable truth. Companies won’t struggle because they don’t have access to powerful models. Instead, they will struggle because they treat AI as a tool to bolt onto old ways of working. McKinsey’s latest global survey argues that the companies with the strongest bottom-line impact aren’t simply deploying more AI. They are redesigning workflows, governance, and decision-making around it.

In Asia, businesses are under pressure to do more while facing tighter margins and less tolerance for delay. AI may now be an operating necessity for many organizations—but necessity alone won’t lead to transformation.

The first wave of enterprise AI has largely focused on assistance. Today’s employees are surrounded by dashboards and beset by emails, yet struggle to get the right insight at the right moment. This is one area where AI has proved immediately useful. It surfaces relevant context and flags anomalies, helping employees act more quickly. When AI helps a finance team detect anomalies before they escalate or enables customer service teams to resolve issues faster, it shows that AI’s value isn’t just theoretical.

The second stage is automation, where AI begins to alter the economics of how work gets done. Traditional automation worked when tasks were repetitive and rules were clear. AI can now expand that range by handling variable and more unstructured tasks with far less manual intervention that before.

The real payoff will be the removal of friction. When approvals move faster, organizations become faster and more efficient. Over time, that can reshape how the business scales.

The third, and ultimately most strategic, benefit is augmentation, where AI begins to expand what the organization can realistically do. It allows companies to coordinate decisions at a scale that would have been difficult to manage manually. AI won’t just improving existing processes, but also make new operating models possible.

Singapore offers a useful glimpse of what that looks like in practice. SMRT, Singapore’s leading public transportation provider, and Oracle are piloting JARVIS, an AI-enabled platform designed to bring together maintenance and operations data, identify potential issues earlier, and help engineering teams intervene before disruptions occur. SMRT’s rail network supports more than two million passenger

journeys a day, which makes the operational stakes obvious. This is a strong example of AI in action where value is created by solving real-world challenges; AI creates value when companies use it used to act before problems become obvious.

That is why the next chapter of AI will be written by companies integrating it into their processes, but rather than treating it as a standalone tool.

The question isn’t whether AI belongs in the enterprise. It does. Instead, the question is whether the organization is prepared to redesign work so AI can deliver value.

Business bottlenecks are a good place to start. Leaders should ask where delays, errors, poor handoffs, duplicated work, or slow decisions are costing the organization money and trust—then they should ask how they need to change to allow AI to remove that friction.

Companies need to also trust AI to take the lead. If AI produces an insight but companies still need manual approval to act on it, value will leak away. It’s the workflow, not the model, that determines whether the transformation succeeds or fails.

Finally, companies need to treat governance as something that enables value. A company that can’t trust its AI won’t use it in consequential decisions.

The next competitive divide won’t be between companies that adopted AI early and those that adopted it late, but rather between those that integrated AI into their workflows and those that kept it at the edges. The latter will be stuck with disconnected pilots and isolated tools that won’t change enterprise performance.

Executives needed to experiment with AI to understand what it could do. But to succeed in the next phase of AI, they need to stop asking where they can deploy AI, and start asking how much they’re willing to change to adapt to it.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

This story was originally featured on Fortune.com