North American companies are now actively deploying agent AI systems that aim to reason, adapt, and act with complete autonomy.
Digitate’s three-year global program data shows that while adoption is universal across the board, regional maturation paths diverge. While North American companies are scaling towards full autonomy, European companies are prioritizing governance frameworks and data management to build long-term resilience.
From practicality to profitability
The story of enterprise automation has changed. In 2023, the main goals for most IT leaders were to reduce costs and streamline daily operations. By 2025, the focus will be further expanded. AI is no longer seen as just an operational utility, but as a benefit-producing capability.
Data supports this shift in perspective. According to the report, organizations in North America saw a median return on investment (ROI) of $175 million from implementation. Interestingly, this financial validation is not unique to the fast-moving North American market. European companies report comparable median ROIs of around $170 million, despite a more cautious and governance-focused approach.
This consistency suggests that although deployment strategies differ, such as a focus on risk management in Europe and a focus on speed in North America, the financial outcomes are similar. All organizations surveyed confirmed that they have leveraged an average of five different tools to implement AI within the past two years.
Generative AI remains the most widely deployed at 74%, but we are seeing a notable increase in “agent” capabilities. More than 40% of enterprises have deployed agents or agent-based AI, moving beyond static automation to systems that can manage goal-oriented workflows.
IT operational autonomy becomes a testing ground for agent AI
While marketing and customer service often dominate public discussions about AI, the IT sector itself is emerging as a key laboratory for these deployments. IT environments are inherently data-rich and structured, creating ideal conditions for models to learn while remaining sufficiently dynamic to require the adaptive inference that agentic AI systems promise.
This explains why 78% of respondents have deployed AI within IT operations, the highest percentage of any business function. Cloud visibility and cost optimization leads the adoption curve at 52%, followed closely by event management at 48%. In these scenarios, rather than alerting humans to problems, technology proactively interprets telemetry data to provide a unified view of spending across the hybrid environment.
Teams leveraging these tools report improved decision-making accuracy (44%) and efficiency (43%), allowing them to handle higher workloads without increasing escalations.
Cost and the human conundrum
Despite the optimism around ROI, the report highlights “cost and human challenges” that threaten to slow progress. This contradiction is straightforward. Businesses are deploying AI to reduce their dependence on labor and operational costs, but these very factors are acting as a major inhibitor to growth.
47% of respondents cited the continued need for human intervention as a major drawback. These agent AI systems are far from achieving the full autonomy of “set-it-and-forget” solutions and require continuous monitoring, adjustment, and exception management. At the same time, implementation costs were the second-highest concern at 42%, driven by costs associated with model retraining, integration, and cloud infrastructure.
The human resources needed to manage these costs are lacking. For 33% of organizations, a lack of technical skills is the primary barrier to future adoption. The demand for professionals who can develop, monitor, and manage these complex systems exceeds the current supply, creating a self-reinforcing loop in which investment increases operational capacity while simultaneously increasing human and financial dependence.
Gap between trust and awareness
Differences in perspective exist between executives and practitioners. Although 94% of all respondents expressed trust in AI, this trust is not evenly distributed. C-suite leaders are significantly more optimistic, with 61% categorizing AI as “very trustworthy” and viewing AI primarily as a financial tool.
Only 46% of non-executive practitioners share this high level of trust. Those close to the day-to-day operations of these models are more aware of the reliability issues, lack of transparency, and need for human oversight. This gap suggests that leadership is focused on long-term review and autonomy, while teams on the ground are grappling with practical delivery and governance challenges.
There are also different views on how these agents work. 61% of IT leaders view agent systems not as replacements but as collaborators that augment human capabilities. However, expectations for automation vary by industry. In retail and transportation, 67% believe agent AI will change the key tasks of their roles, while in manufacturing, the same percentage see these agents primarily as personal assistants.
Full agent AI autonomy is rapidly approaching
We expect the industry to rapidly move towards reducing human involvement in routine processes. Currently, 45% of organizations operate as semi-autonomous or fully autonomous enterprises. Projections suggest this number will rise to 74% by 2030.
This evolution means a change in the role of IT. As capabilities mature, IT is expected to move from being an operational enabler to an orchestrator role. In this model, IT functions manage “systems of systems” and enable various intelligent agents to interact correctly while humans focus on creativity, interpretation, and governance rather than execution.
Avi Bhagtani, CMO of Digitate, said, “Agentic AI bridges the gap between human ingenuity and autonomous intelligence, and heralds the dawn of IT as a strategic, profitable function.” “Companies have moved from experimenting with automation to extending AI for measurable impact.”
Moving to agent AI requires more than just software procurement. We need an organizational philosophy that balances automation and human augmentation. Policies alone are not enough. To ensure transparency and ethical oversight in all decision-making loops, governance must be integrated directly into system design. European organizations are currently leading the way in this area, prioritizing ethical deployment and strong monitoring frameworks as the basis for resilience.
Furthermore, the shortage of technical human resources cannot be solved by recruitment alone. Organizations should invest in upskilling their existing teams by combining operational expertise with data science and compliance literacy.
Finally, reliable autonomy depends on high quality data. Investments in data integration and observability platforms are necessary to provide agents with the context they need to act independently.
The era of experimental AI is over. The current stage is defined by the pursuit of autonomy, whose value comes not from novelty but from the ability to scale agent AI sustainably across the enterprise.
“As organizations balance autonomy and accountability, those that embed trust, transparency, and human engagement into their AI strategies will shape the future of digital business,” concludes Bagtani.
See also: How updates to the MCP specification improve security as your infrastructure grows
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events including Cyber Security Expo. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
