AI investment has accelerated rapidly, but that investment isn’t necessarily translating into more capacity for IT teams. In many organizations, IT operations feel as constrained as ever, if not more so.
The issue may be less about how much organizations are willing to spend and more about where that investment is going. Despite growing IT budgets, 48% of corporate IT teams cite a lack of time as a primary barrier to moving new initiatives forward. Higher spending alone cannot solve structural workload issues, particularly when funds are directed toward additional software that requires significant onboarding, implementation and integration effort from already stretched-thin IT teams before delivering efficiency gains.
The real question for IT leaders then becomes: what are organizations paying for, if not productivity? And how can IT leaders establish the groundwork needed to effectively operationalize AI?
AI Optimism Versus Operational Reality
According to Auvik’s 2026 IT Trends Report, nearly 70% of IT professionals are optimistic about AI’s future impact on their work, yet only 5% report AI is currently core to their daily operations.
One persistent barrier to realizing AI’s full potential in IT operations is tool sprawl. More than a third of MSPs and over a quarter of internal IT teams now manage ten or more tools in their technology stacks. Each new tool promises to save time and improve efficiency, yet each requires time investment to implement, manage, integrate, and train teams on how to use it. The fragmented data created by these disconnected tools also makes it harder for AI agents to operate effectively.
The problem will only continue to snowball if left unaddressed. Gartner projects that the average Fortune 500 enterprise will be running over 150,000 agents by 2028, a stark increase from fewer than 15 in 2025. This sudden surge could hit IT teams before they’ve even finished integrating the tools they already have. Without a unified way to manage AI agents and the data they rely on, agent sprawl can quickly become an even greater challenge than the tool sprawl IT teams have been contending with.
If ten separate dashboards were difficult to govern, imagine the challenge that comes with thousands of AI agents operating across systems, accessing data and taking actions with varying degrees of autonomy. Increasing AI budgets without addressing the operational complexity underneath them risks scaling the problem alongside the technology.
Building the Foundation for AI ROI
Realizing the value of AI requires more than deploying new capabilities. IT leaders need to simplify the environments those systems operate in, establish policies that reflect how AI is actually being used and ensure clear visibility across the infrastructure AI depends on.
IT leaders may be envisioning autonomous systems that identify and address problems before users are affected. Meanwhile, teams on the ground are still managing fragmented tools, incomplete visibility and alert fatigue. Building toward greater automation starts with addressing those fundamentals.
- Audit the existing tool stack. Before making additional investments, organizations should understand the tools already in use, including shadow AI applications employees may have adopted independently. You can’t consolidate or optimize what you can’t see, and it’s critical to close this gap at the starting point rather than later in the integration or adoption phase.
- Treat AI governance as an ongoing practice. An AI policy should not be a static document that is written once and forgotten. 76% of leaders surveyed in Auvik’s IT Trends report say they have an AI policy in place, but only 42% of frontline workers agree, highlighting a critical gap in communication and implementation. IT leaders should treat policy rollout as an ongoing conversation with regular check-ins. Policies need to evolve alongside new technologies and be communicated in a way that reaches the employees actually using these tools.
- Build the visibility AI needs. AI is only as useful as the context it can access. IT leaders should ensure agents have access to accurate, relevant and up-to-date operational data so they can understand what’s happening across the environment, identify how different events and systems relate to one another and support faster, more automated resolution. Without access to that contextual data, the best AI agents can do is offer generic how-tos.
There’s no doubt that the next few years will bring a wave of innovation in how AI agents manage our complex and time-consuming IT tasks. Yet without the groundwork to turn AI’s promise into operational reality, those advances will only magnify the gap we see today: organizations investing more in technology that can do more, without necessarily getting more from it. Closing that gap requires investing not only in what AI can do, but in everything around it that makes it work.
