Agentic AI is gaining serious traction across North America, where enterprises are moving beyond basic automation and toward systems capable of reasoning, adapting, and acting on their own. New findings from Digitate’s three-year global programme show that while companies worldwide are adopting AI at a fast pace, regions are taking different routes. North American firms are racing toward full autonomy, while European organisations are prioritising governance, oversight, and long-term resilience.
Ryan Daws
AI shifts from cost saver to profit driver
AI’s role inside the enterprise has changed quickly. In 2023, most IT leaders focused on cutting costs and streamlining routine work. By 2025, adoption patterns tell a broader story. Companies now see AI not only as a tool for efficiency but as an engine for profit.
The data is striking. North American organisations report a median return on investment of 175 million dollars. European enterprises, despite taking a more cautious and governance-first approach, report a similar median ROI of about 170 million dollars. This suggests that whether the goal is speed or risk management, the financial payoff is largely the same.
Every organisation surveyed has deployed AI in the past two years and uses an average of five distinct tools. While generative AI remains the most common, with a 74 percent adoption rate, agentic systems are rising fast. More than 40 percent of enterprises now use agentic or agent-based models that can manage goal-driven workflows instead of simply following static rules.
IT operations emerge as the proving ground
Although customer service and marketing often dominate AI headlines, the biggest wave of deployment is happening inside IT. Environments with rich, structured data give models the conditions they need to learn, while fast-moving systems demand adaptable reasoning.
It is no surprise that 78 percent of respondents have rolled out AI inside IT operations. Cloud visibility and cost optimisation lead adoption at 52 percent, followed by event management at 48 percent. In many cases, the systems are no longer just sending alerts. They are interpreting telemetry data, mapping costs across hybrid environments, and giving teams clearer insight into system behavior.
Users of these tools report real gains. Decision accuracy has improved for 44 percent of teams, and efficiency for 43 percent, allowing IT departments to take on more work without increasing escalations.
The cost-human paradox
Even with solid ROI, progress is not without friction. Enterprises face what Digitate calls the “cost-human conundrum.” AI is meant to reduce reliance on human labor, yet human labor is often the limiting factor.
Almost half of respondents say the ongoing need for human oversight is a major challenge. These systems are not yet operating with full autonomy. They require continuous tuning, monitoring, and exception handling. At the same time, the cost of implementation is the second-largest concern at 42 percent, driven by retraining needs, integration demands, and cloud resource requirements.
Talent shortages remain a pressing issue. One in three organisations cites a lack of technical skills as the top barrier to further adoption. As agentic systems become more complex, demand for individuals who understand both AI and enterprise operations continues to grow.
A widening trust gap
Most organisations say they trust AI, but attitudes vary by role. While 94 percent of total respondents express trust, only 46 percent of non-C-suite practitioners consider AI “very trustworthy.” By contrast, 61 percent of executives hold that view. Leaders tend to see AI as a strategic and financial opportunity, while operators are focused on reliability, transparency, and day-to-day risk.
Expectations differ across industries as well. In retail and transportation, 67 percent believe agentic AI will change the core tasks of their roles. In manufacturing, the same share views these tools as personal assistants designed to support, not replace, human work.
Autonomy is approaching faster than expected
Nearly half of enterprises already operate with some level of semi- or fully autonomous processes. By 2030, that figure is expected to climb to 74 percent. This shift is likely to reshape the role of IT. As systems gain more autonomy, IT teams will take on orchestration responsibilities, ensuring that interconnected agents cooperate while humans focus on interpretation, strategy, and oversight.
Digitate’s CMO, Avi Bhagtani, describes agentic AI as the connection between human creativity and scalable autonomous intelligence. He says enterprises have moved past experimentation and are now pursuing measurable business outcomes.
- Woolworths
To reach reliable autonomy, organisations will need more than new tools. Governance must be embedded into system design, not bolted on later. Europe currently leads in building strong oversight frameworks that prioritize ethical deployment and resilience.
Upskilling is another necessity. Hiring alone cannot fill the talent gap. Companies must strengthen internal teams by combining operational knowledge with data science and compliance skills. High-quality data will also be essential. Without strong integration and observability, autonomous agents cannot operate with clear context.
Ryan Daws
Agentic AI is entering a pivotal phase. The experimental era is ending, and enterprises are now focused on scaling these systems in a sustainable, accountable way. Those that balance autonomy with trust and human engagement will be the ones that shape the next chapter of digital business.
