Something has shifted in enterprise IT, and it’s not getting enough attention.
Today, boards, CEOs, CIOs, and risk leaders increasingly want to know one thing:
Can we trust artificial intelligence to assume greater operational responsibility?
I've spent the past year deep in three parallel bodies of research that, taken together, have made this issue impossible to ignore. My first survey, Redefining Modern Service Management, queried 204 IT leaders on how their organizations manage the convergence of IT service management, IT operations, and AI. The second, The Reality of Observability Unification, completed with my colleague Shamus McGillicuddy, surveyed 356 IT tool experts on the current state of observability unification in IT operations. In the third survey, ITAM in the Age of AI—publishing soon—I asked more than 200 IT pros if IT asset management practices are mature enough to support the AI workloads organizations now run at scale.
Although these studies examine different aspects of enterprise IT, they all help draw the same conclusion. Artificial intelligence doesn’t expose a technology problem. It clearly exposes a trust problem.
The 2025 ServiceOps research confirms AI adoption is real and accelerating: 71% call automation a C-suite mandate, 60% use autonomous AI for communications and escalations, and 64% have generative AI in production or pilot. But the challenges are also real, with 53% saying they lack adequate real-time discovery, dependency mapping, ITAM, and CMDB data and 63% requiring human approval before acting on AI recommendations. The top AI adoption challenge cited was governance and guardrails—fundamentally a trust problem. The Reality of Observability Unification survey showed that AI delivers the greatest value after organizations establish shared operational data, standardized workflows, meaningful service context, and effective cross-functional collaboration. Finally, my ITAM for AI research shows that AI does not fix untrusted or fragmented asset data. Instead, it depends on trustworthy data to function.
The 2026 Trusted AI Operations: EMA’s 2026 Service Delivery Readiness Report I’m starting now examines whether operational foundations can support AI safely, assessing data currency and confidence across cloud, SaaS, containers, and AI-specific assets. The picture is troubling: many organizations run automated, sometimes consequential, decisions on periodic, siloed asset data with no clear governance owner—data they wouldn't trust for a manual audit. Together, these studies show AI capability is scaling faster than the operational trust foundation required to support it.
IT research on AI tends to focus on adoption rates, feature inventories, and vendor comparisons—useful, but it leaves unanswered whether organizations have built the operational foundations that make AI trustworthy over time.
This new study will benchmark the operational disciplines behind trustworthy AI: process maturity, continuous discovery, configuration and asset data quality, workflow automation, operational visibility, governance, and business alignment. It will also examine how organizations balance human oversight with AI-assisted decision making, and which capabilities most strongly correlate with executive confidence.
The goal: a practical framework for evaluating AI readiness.
The market has no shortage of AI maturity studies. This research is a deeper dive into whether service delivery's foundations are solid enough to make AI trustworthy, governable, and safe to expand. ServiceOps research showed convergence happening; ITAM research showed gaps in the data underneath it. My future research will define what “ready” looks like and give organizations a high-level framework for self-assessment.
Productivity and cost savings will always remain real drivers, but organizations continue to invest in tooling and processes that reduce risk, strengthen governance, and build trust.
If you work in IT service management, IT operations, enterprise service management, or anywhere in the automation and AI governance space, I'd welcome your perspective as this research develops. The best industry benchmarks are shaped by the practitioners who live these challenges every day.
This article is part of an ongoing research series exploring the operational foundations, governance, and trust required for organizations to adopt AI successfully. Future articles will examine operational trust, AI governance, and the evidence organizations need before expanding AI autonomy.