The Second Wave of AI Will Be About Value, Not Adoption

For the past two years, one question has dominated boardrooms across industries: "How fast can we adopt AI?"

It was the right question to ask. Today, it is no longer the most important one.

A different conversation is beginning to emerge — one that is quieter, less visible, but ultimately far more consequential for the future of enterprise AI: "How do we know whether all this AI is actually creating value?"

That shift may seem subtle, but it signals the beginning of a new phase in the AI revolution. The first wave rewarded speed. Companies rushed to deploy copilots, experiment with large language models, automate repetitive tasks, and launch internal AI initiatives before competitors did.

The second wave will reward discipline.

Not because AI has become less important, but because it is becoming something much more familiar: infrastructure.

Every technological revolution follows a similar pattern. The first years are driven by excitement and experimentation. The following years are shaped by governance, operational excellence, and economic reality.

Cloud computing followed that trajectory. What began as an exercise in flexibility eventually forced organizations to develop FinOps — a discipline dedicated to understanding cloud consumption, optimizing costs, and ensuring technology investments translated into business outcomes.

Artificial intelligence appears to be approaching a similar inflection point.

Only this time, the challenge extends far beyond infrastructure costs.

It reaches the very way organizations create, measure, and govern value.

The enthusiasm surrounding AI has been extraordinary. Deloitte's latest State of AI report shows enterprise adoption continuing to accelerate, while companies expect the number of AI initiatives in production to grow significantly over the next year. Yet, the same research reveals an uncomfortable contradiction: although productivity gains are becoming increasingly common, only a minority of organizations report that AI is fundamentally reshaping their business. Most are making existing work faster rather than redesigning how value is created.

That distinction matters more than it appears.

Increasing productivity is not the same as improving profitability.

Automating tasks is not the same as building competitive advantage.

And deploying more AI does not necessarily mean creating more business value.

Many organizations are beginning to discover this firsthand.

Individual teams report time savings. Employees appreciate their new tools. Pilot projects produce encouraging demonstrations. Yet when executive teams ask a much simpler question—"What impact has this had on revenue, margins, customer experience, or EBITDA?"—the answers often become surprisingly difficult.

Not because AI failed. Because most organizations never defined what success was supposed to look like.

The first generation of AI adoption was driven by fear of missing out.

The next generation will be driven by accountability.

This transition is already becoming visible.

Research from Gartner suggests that while enterprise AI adoption has accelerated rapidly, fewer than half of organizations have implemented financial guardrails or AI FinOps practices to manage AI investments, even as executives face increasing pressure to demonstrate measurable returns.

That should not be surprising.

For many companies, AI spending is no longer limited to a single technology budget. Marketing teams subscribe to one platform. Developers use another. Legal departments experiment with different assistants. Customer service builds specialized agents. Business units purchase AI capabilities independently while employees create their own workflows using public models.

Gradually, organizations accumulate dozens of disconnected AI initiatives.

Some overlap.

Some duplicate existing capabilities.

Some create value.

Some simply create activity.

Few companies have a complete picture.

This is remarkably similar to what happened during the early years of cloud computing.

Organizations loved the flexibility of cloud services until they realized nobody truly understood where the money was going. Costs became fragmented across business units. Consumption became invisible. Infrastructure expanded faster than governance.

Cloud computing did not become more valuable because organizations bought more servers.

It became more valuable because companies learned how to operate it.

AI is entering that same stage.

The challenge, however, is considerably broader than cost optimization.

The real question is no longer how much organizations are spending on AI.

It is whether they understand what they are buying.

More importantly, whether they understand what they are building.

An AI strategy should not be measured by the number of copilots deployed or the quantity of automation projects launched. Those are activity metrics.

Business leaders should be asking different questions.

Which AI initiatives are generating measurable economic value?

Which ones are simply reducing effort without improving outcomes?

Where is AI creating entirely new capabilities instead of making existing processes marginally faster?

Perhaps the most uncomfortable question of all is this: If we stopped paying for half of our AI tools tomorrow, would anyone notice?

That question is less about technology than organizational clarity.

Because the greatest risk organizations face may not be overspending.

It may be confusing experimentation with transformation.

One reason this distinction matters is that AI is quickly becoming embedded in every layer of the enterprise. McKinsey recently observed that as organizations move from isolated pilots to enterprise-wide deployment, AI spending increases dramatically while visibility often decreases. Individual business units adopt their own tools, employees build autonomous workflows, and token consumption expands faster than governance mechanisms. The result is a familiar pattern: organizations accelerate adoption before developing the operating discipline required to manage it.

This is precisely where the next generation of enterprise opportunity is emerging.

Not in building another language model.

Not in launching another chatbot.

Not even in creating another AI agent.

The next generation of value may come from helping organizations answer a much more fundamental question:

How do we operate AI as a business capability?

That question extends beyond technology teams.

It belongs equally to CFOs trying to justify investment decisions, CEOs redefining competitive strategy, boards overseeing enterprise risk, and operating executives responsible for execution.

Governing AI will require organizations to understand not only cost, but adoption, business impact, knowledge management, vendor concentration, security, compliance, and organizational learning.

In other words, the conversation is shifting from AI implementation to AI operations.

This evolution may prove particularly relevant for Latin America.

Across the region, companies have demonstrated remarkable speed in adopting AI despite resource constraints. Yet many organizations continue to operate with highly decentralized decision-making, informal processes, and institutional knowledge concentrated in key individuals. Adding AI to that environment without redesigning operating models risks accelerating inefficiency rather than eliminating it.

The opportunity, therefore, is not simply to deploy artificial intelligence.

It is to design organizations capable of extracting sustained economic value from it.

That requires moving beyond technology metrics and embracing business metrics.

Beyond productivity toward profitability.

Beyond automation toward operating leverage.

Beyond adoption toward accountability.

The companies that create the greatest value over the next five years may not be those announcing the largest AI investments.

They may simply be the ones asking better questions.

Not, "Which model should we use?"

But, "Which capability are we trying to build?"

Not, "How many employees are using AI?"

But, "What measurable business outcome has AI improved?"

Not, "How quickly can we deploy another agent?"

But, "Which decisions become fundamentally better because this capability exists?"

The first wave of enterprise AI rewarded organizations that moved first.

The second will reward those capable of proving that AI creates value — not just activity.

Because eventually, artificial intelligence will become another layer of enterprise infrastructure.

When that happens, competitive advantage will no longer belong to companies that merely adopted AI early.

It will belong to those that learned how to govern it, measure it, and turn it into an enduring business capability.