Global businesses are riding a new wave of AI-powered innovation to transform IT operations, or ITOps.

Research reveals that more than half of organizations have already deployed AI agents and more than a third are planning to do so. This comes amidst growing executive trust in AI agents as they are increasingly relied on within IT and other business operations.

AI can be a critical part of the IT puzzle by helping to accelerate incident response, reduce downtime and keep customers happy. These gains can shift executive attitudes, as leaders come to see AI agents not just as experimental tools, but as reliable partners in mission-critical situations. It's no surprise, then, that 81% of IT and business executives now trust AI agents to take action during a crisis.

Adding Value for Ambitious Businesses

AI is already having a profound impact on digital operations, helping organizations to innovate faster and work more efficiently. In fact, a recent survey found that 51 % of organizations using AI reported that it had a transformational impact on their operations. AI agents can amplify that further by autonomously executing complex, multi-step tasks, helping to automate entire business processes.

It's part of the reason why we're seeing so many organizations adopt AI agents. In PagerDuty's latest survey, nearly three-quarters of executives (74%) reported that AI has become so essential to operations that they would struggle to function without it. A similar share has already deployed more than one agent. Interestingly, smaller companies seem to be even more reliant on the new technology than their larger peers.

Executives surveyed also revealed that AI is expected to handle as much as a third of their departmental workload in the next 12 months. Many are deploying AI agents for code writing or reviewing, but business processes as diverse as employee onboarding and supply chain optimization could be made more efficient and cost effective through agentic AI.

Challenges Mount

AI adoption is spiking in part because a significant majority of organizations now trust the technology more than they did a year ago, with IT leaders in particular feeling more confident. The improved sentiment stems from leaders observing better-quality outputs and more positive results in practical deployments and pilots.

However, legitimate concerns persist. Organizations are most anxious about outages that may impact their AI systems. 84% of survey respondents reported having experienced at least one outage impacting either internal or external models and tools. A similar share believes their organization needs to improve its ability to detect AI failings to respond to incidents more rapidly.

In many companies, those with AI skills are spread too thin across organizations, meaning fewer experts may be available or on-call to fix issues. Even among those qualified to deal with these incidents, they may find that black box AI systems defy traditional debugging approaches. It's important to remember that many are laboring with siloed incident response/detection systems, which may exclude input from customer support and other functions. Many of these tools aren't designed to monitor complex AI systems with cascading decision layers.

AI Lights a Way Forward

Fortunately, AI can be an important ally as organizations tackle these operational issues. AI embedded into tools can learn from previous incidents to accelerate triage, reduce alert fatigue by minimizing "noise" and review recent changes to proactively surface issues. Generative AI can also add value with automatically generated post-incident reviews for enhanced team learning and with AI-generated runbooks to streamline response. It could also produce timely status updates to keep stakeholders informed of incident developments without impacting first-responder workloads.

Agentic AI can help incident response move even faster, autonomously resolving routine issues so engineers can focus on innovation. This could involve pulling relevant triage data and running diagnostics to make first responders' jobs easier, or recording and sharing incident information for other responders, and even surfacing recommended areas for ongoing improvements.

Predicting and Preventing Failure

The most advanced organizations will use AI to attain a level of operational maturity where AI and automation are used for predictive maintenance and incident prevention. As a result, these organizations can decrease the likelihood of an AI-related system error occurring, helping to ensure customers are happy, services remain uninterrupted and engineers are focused on growth rather than incident response.

It's an aspiration rather than a reality for most enterprises today, but as AI finds its way into more business-critical systems, a preventative approach to ITOps should at least be on the roadmap for teams. A combination of embedded, generative and agentic AI can be key in helping an organization be prepared for the next major outage.