Enterprise adoption of agentic AI is hitting a pivotal moment, with new research from Dynatrace showing that firms are ramping up investment but scaling cautiously until reliability can be assured.

Polling more than 900 senior leaders responsible for agentic AI implementation, Dynatrace’s Pulse of Agentic AI 2026 report found that companies are stalling on advanced AI projects, not because they doubt their value, but because they cannot govern, validate, or safely scale these autonomous systems.

The research found that around 50% of agentic projects are stuck in the proof-of-concept or pilot stage, though early adoption has grown, with 26% of organisations having eleven or more projects. 

Of those, 50% already have agentic AI projects in limited production, 44% are scaling across select departments, and 23% report mature, enterprise‑wide integration.

Among the top priorities of senior leaders when deploying agentic AI, improving decision-making with real-time insights is foremost (51%), followed by gains in system performance and reliability (50%) and reducing operational costs (50%). 

The main barriers to agentic AI production revolve around security, privacy or compliance concerns (52%), followed by technical challenges around managing and monitoring agents at scale (51%).

These concerns mean that human guidance remains a central part of firms’ agentic AI strategy, even as they build toward greater autonomy.

The report shows leaders expect a 50/50 human–AI collaboration for IT and routine customer-support applications, and a 60/40 human–AI collaboration for business applications, with 87% of organisations building or deploying agents that require human supervision.

Over half (64%) of firms said they deploy a mix of autonomous and human-supervised agents, relying on validation methods that include data quality checks (50%), human review of agent outputs (47%), and monitoring for drift or anomalies (41%).

Meanwhile, only 13% of organisations use fully autonomous agents, and 44% reported using manual methods to review communication flows among AI agents, highlighting the need for governed oversight mechanisms.

“While human oversight remains essential today, organisations are increasingly preparing for more autonomous, AI-driven decision-making. The focus is now on building the trust and operational reliability needed to scale agentic AI responsibly,” said Alois Reitbauer, chief technology strategist at Dynatrace.

Recommended reading

• Will AI Agents Be the Enterprise Disruptor of 2026?
• Microsoft Tests AI Agents in Simulated Marketplace
• 80% of Firms Say Their AI Agents Have Taken Rogue Actions

According to the report, this shift means observability is now a crucial intelligence layer that provides visibility across every stage of the agentic AI lifecycle.

Nearly 70% of businesses reported already using observability to gain real-time visibility into agent behaviour, with the highest adoption during implementation (69%), followed by operationalisation (57%) and development (54%).

“Observability is a vital component of a successful agentic AI strategy,” said Reitbauer. “Observability not only helps teams understand performance and outcomes, but it provides the transparency and confidence required to scale agentic AI responsibly and with appropriate oversight.”