AI Spending Is Accelerating. Why Is ROI Still So Difficult To Find?

For the past two years, organizations have been encouraged to think of AI as the next great business transformation opportunity. Vendors have promised productivity gains, operational efficiencies and entirely new ways of working. Boards have approved new budgets. Executive teams have launched pilots. Technology leaders have raced to evaluate platforms, copilots and increasingly sophisticated AI agents. Yet despite the extraordinary momentum behind AI, a question is beginning to emerge in boardrooms around the world. After all the investment, where is the return?

This question sits at the heart of Forrester's The CIO's Guide to AI Readiness, authored by Vice President and Principal Analyst Frederic Giron. The report arrives at a particularly important moment because the conversation around AI is changing. Two years ago, organizations were focused on understanding what AI might be capable of. Today, they are being asked to demonstrate what it is actually delivering. AI spending continues to rise across infrastructure, platforms, models and talent, yet many leadership teams still struggle to articulate how those investments are translating into measurable business outcomes. The pressure to invest has not diminished, but the pressure to prove value has increased significantly.

What makes Forrester's research particularly useful is that it challenges one of the most common assumptions in the market: that AI readiness is primarily a technology challenge. Most organizations instinctively evaluate readiness through a technical lens. Do we have the right infrastructure? Do we have access to the right models? Have we hired the right skills? While these questions remain important, Forrester argues they represent only part of the equation. Organizations that successfully operationalize AI tend to demonstrate maturity across leadership, governance, operating models, business ownership and trust. In other words, readiness is not simply about whether an organization can deploy AI. It is about whether it can absorb AI, govern AI and translate AI into measurable business outcomes.

This distinction helps explain why so many organizations remain trapped between experimentation and transformation. Many already possess modern cloud environments, capable analytics teams, sophisticated data platforms and access to leading AI technologies. Yet despite these investments, they struggle to move beyond isolated pilots and demonstrations. According to Forrester, the issue is often not technical capability but organizational capability. Executive sponsorship may be inconsistent. Ownership may be unclear. Governance may be disconnected from operational processes. Business teams may not trust the information supporting AI-driven decisions. Success may be measured through activity rather than outcomes.

The report highlights a number of characteristics commonly found in organizations that are successfully operationalizing AI. Executive sponsorship is visible and sustained. Accountability for outcomes is clearly established. Governance is embedded into business processes rather than added after deployment. Collaboration exists between business and technology teams. Most importantly, AI initiatives are directly connected to business objectives and measurable value. These observations may appear straightforward, yet they represent some of the most significant barriers organizations face today.

Forrester also highlights the importance of organizational adaptability. As AI becomes increasingly embedded into workflows, decision-making and customer engagement, organizations must be capable of continuously evolving the way they operate. Governance models must mature. Ownership structures must adapt. Decision-making frameworks must change. The challenge is no longer introducing AI into the business. The challenge is ensuring the business can evolve alongside AI. At Data Tiles we call this “AI as an experience”, where AI is invisible to the organization. It is the enablement that is experienced.

This is why the report feels particularly relevant in 2026. Executive expectations are changing rapidly. Boards are demanding evidence that AI investments are translating into competitive advantage. CFOs are increasingly focused on measurable outcomes and sustainable returns. Investors want confidence that AI initiatives are improving productivity, profitability and growth rather than simply increasing technology expenditure. The market is moving from experimentation to accountability.

And this is where I believe the conversation becomes particularly interesting.

Forrester correctly identifies readiness as one of the defining challenges facing organizations today. However, I would argue that the opportunity is even larger than the report suggests. Much of the market still approaches AI as a way to improve existing processes. The underlying assumption is that AI should help organizations perform the same work faster, cheaper or more efficiently than before.

That view is understandable.

It is also limiting.

One of the greatest mistakes organizations can make is treating AI primarily as a replacement strategy. Human-led projects have often struggled with delays, complexity, competing priorities, inconsistent outcomes and escalating costs. The temptation is to assume AI can simply replace those activities and produce better results. Yet if a process is fundamentally flawed, replacing people with AI does not create transformation. It simply automates the failure.

The greatest opportunity presented by AI is not the replacement of human effort. It is the opportunity to rethink how work happens altogether.

Too many organizations begin their AI journey by asking where people can be removed from a process. The more valuable question is how that process would be designed if trusted information, business context and intelligent automation were available from the outset. That shift in thinking changes everything. It moves the conversation away from cost reduction and toward reinvention. It encourages organizations to stop replicating yesterday's operating model and start designing tomorrow's.

This is where the conversation around ROI often becomes confused.

Organizations do not invest in AI because they want AI. They invest because they want better outcomes. They want cost out or revenue up. They want stronger growth, improved customer experiences, greater operational efficiency, faster decision-making, reduced risk and more effective use of resources. AI is simply the mechanism through which those outcomes are expected to be achieved.

The problem is that organizations struggle to connect AI activity to business outcomes. A pilot may be successful. A model may achieve impressive levels of accuracy. An agent may perform exactly as designed. Yet leadership teams are often left asking the same question: what changed? Did revenue increase? Did costs decrease? Did productivity improve? Did customer retention improve? Did decision quality improve?

Without clear answers, AI risks becoming another technology investment that generates activity rather than transformation.

This is why I often describe AI readiness as decision readiness. The organizations creating the strongest returns from AI are rarely those deploying the most technology. They are the organizations creating the shortest path between trusted information and business action. They establish clear ownership. They align business definitions. They embed governance into the point of decision-making. They create trusted information assets that both humans and AI can rely upon with confidence.

At Data Tiles, we increasingly see organizations moving beyond the idea of becoming data-driven and toward becoming decision-driven. The distinction is important. Data-driven organizations focus on access to information. Decision-driven organizations focus on confidence in action. They recognize that data alone does not create value. Value is created when trusted information is connected to business outcomes and can be acted upon with confidence.

This is where trusted data products become particularly powerful. A trusted data product is not simply a dataset. It combines information, business context, ownership, governance, lineage and accountability into a reusable business asset. It provides both people and AI with confidence that the information supporting a decision is trusted, explainable and relevant. When these foundations exist, AI stops being an experiment. It becomes operational. More importantly, it becomes measurable.

The organizations that will generate the greatest value from AI over the next decade are unlikely to be those deploying the largest number of agents or investing in the largest number of platforms. They will be those that create entirely new ways of operating around trusted information, intelligent automation and decision velocity. They will use AI not to replicate human-led inefficiencies, but to design better systems from the beginning.

That is where transformation occurs.

And ultimately, that is where ROI is found.