ITOps hits a turning point with agentic AI

For many ITOps teams, any investment in AI might be better than none at all, according to a new report from market research firm Omdia, titled “Modernizing IT Operations in the Agentic Era.”

The IBM-commissioned report reveals that, even among organizations that hand less than 10% of their ITOps duties to AI, 92% observe operational improvements. This finding reinforces an increasingly common strategy: starting with simple integration wins before advancing to more complex AI deployments.

Yet deciding which areas of ITOps to prioritize can be daunting, especially when IT teams are already expected to deliver reduced costs, improved operational efficiency and faster delivery across increasingly distributed architectures and services.

What’s more, AI and agentic workflows have further complicated IT deployments. Some teams struggle to provide models with sufficient proprietary data. Others lack the technical expertise needed to design and maintain new automations and workflows. Meanwhile, without sufficient monitoring and oversight, agents can perform unauthorized or undesired actions, introducing security and compliance risks. And even in simple use cases, misdirected investments in AI initiatives can yield costly errors.

Despite these challenges, AI integration is fueling ITOps innovation at an unprecedented rate. For 39% of organizations, AI now performs at least half of ITOps duties alongside humans. Early adopters report improved threat detection, predictive maintenance, workflow automation and cost optimization capabilities, among other benefits.

A confluence of internal and external pressures is driving organizations to place machine learning and agentic AI at the center of their ITOps strategies, according to the report.

Commonly cited challenges include maintaining steady performance across multi-cloud and hybrid environments, providing 24/7 response coverage for users and operating within tightening personnel and budget restraints.

Meanwhile, some IT professionals say their department is turning to AI to support enterprise-wide digital transformation initiatives, or in response to external drivers, including competitive pressures, compliance requirements and mounting cybersecurity threats.

Notably, 29% of respondents say their enterprise made AI-focused IT investments to improve existing AI programs that didn’t perform as expected. This finding suggests that AI initiatives are not always immediately productive and might require longer-term operational adjustments to deliver value.

IT teams have used the foundations of AI, such as rule-based systems and statistical modeling, for decades. But multimodal agents—which can navigate digital interfaces and autonomously reason through complex, multi-step problems—have emerged only recently.

Despite the technology’s relative novelty, agentic AI adoption has been swift: 53% of organizations say agents already handle at least a quarter of their ITOps tasks autonomously (alongside human-monitored guardrails), while 8% of organizations assign between 75-89% of ITOps tasks to agents.

In addition, IT professionals now list agentic capabilities as their organization’s top integration priority—ahead of other implementation objectives, such as scaling integrations across every IT domain or focusing only on high-impact areas.

Agentic tools are also trending toward greater autonomy, with the majority of organizations predicting their agents will be highly or fully autonomous within the next 24 months. That independence will come in part from agents’ ability to delegate duties and collaborate on particularly challenging IT tasks.

Organizations with less developed AI programs tend to focus on reducing manual, repetitive tasks, which can deliver immediate efficiency and accuracy improvements. Meanwhile, enterprises with more sophisticated AI implementations (who presumably have extensive automation workflows already in place) instead prioritize more complex deployments, such as enhancing predictive capabilities through automation.

As with most transformative technologies, AI integration requires significant technical, workforce and financial investments. IT professionals say one of their biggest expenses is upgrading infrastructure to support AI workloads. Platform licensing and subscription costs are also major budgetary concerns.

On average, organizations use three financial sources to cover AI integration expenses, rather than relying on a single funding pool. Similarly, AI spending decisions typically involve three or four parties, including ITOps management, chief AI officers and security leaders, during the research and evaluation stages.

Regardless of industry, the report notes some common features that organizations tend to look for when assessing their ideal AI solution.

At the top of the list, more than a third of enterprises prioritize platforms that can integrate seamlessly with their existing monitoring and management tools. Other in-demand features include predictive and forecasting capabilities, hybrid and multi-cloud support and automated incident responses.

While AI integration can provide a variety of long-term benefits, embedding AI-driven workflows into existing IT processes can be operationally complex.

More than a third of IT leaders cite “lack of skilled personnel” as one major barrier, suggesting that IT professionals with AI expertise can maintain a competitive edge—even as the industry shifts toward AI-powered pipelines. To help close AI skill gaps, organizations might introduce training and mentorship programs as well as developer sandboxes, where IT professionals can safely experiment with agentic workflows.

Another widely shared concern is that new AI platforms will be unable to communicate with current IT components, leading to incompatibilities and bottlenecks.

Relatedly, AI models might struggle to ingest information trapped inside data lakes and data warehouses, resulting in inaccurate or biased outputs. As a consequence, organizations that successfully harness unstructured data for model training stand to gain an advantage, leading to more refined forecasts and workflows.

Most respondents do not cite a single, shared operational barrier, suggesting that organizations face a diverse range of AI integration challenges rather than a universal problem.

Despite initial operational hurdles, AI has the potential to reshape nearly every aspect of ITOps, according to the report. For example, more than half of respondents say they’ve already experienced (or expect to experience) improvements across autonomous monitoring, problem resolution and security threat detection.

Despite having different investment approaches and operational priorities, most IT leaders remain bullish on the idea that AI adoption will become increasingly important in the next 12 to 24 months. The report suggests that for the majority of ITOps departments, any level of AI investment is better than avoiding AI altogether.

IT leaders are also confident that AI and agents will improve the productivity of entry-level workers and will help improve system reliability. Meanwhile, 86% of respondents believe human oversight “will remain critical,” even as AI systems become more sophisticated and autonomous.

AI is already reshaping AIOps, ITAM/SAM, capacity planning and knowledge management. But the study also describes some surprising areas where AI is still underutilized, hinting at future ITOps trends.

At the same time, like any investment, AI integration carries risk, and investing heavily in the wrong initiative can amount to a costly mistake. The report offers a detailed look into how ITOps teams are allocating AI resources—and critically, which investment areas yield measurable results.

Together, these insights can serve as a roadmap for both early-stage integrators making their first AI investment and advanced ITOps teams that want to expand their AI footprint to gain an edge over the competition.

Download the full Omdia analyst report