UiPath survey says enterprises still piloting agentic AI

UiPath has published survey findings on enterprise adoption of agentic AI, showing that many large organisations remain in the pilot stage.

The survey covered 590 C-suite and IT practitioners at companies with annual revenue of at least USD $1 billion and a minimum of 1,000 employees, excluding public sector organisations. Fewer than one in three respondents, or 31%, said AI was fully embedded in their business.

Among organisations already using AI, 35% said deployment was limited to selected teams, while 11% said they were still in pilot mode and conducting limited experimentation. The figures suggest many companies have moved beyond initial trials but have yet to roll out the technology more widely.

Respondents identified three main obstacles to improving deployments. Data quality and readiness was cited by 38%, integration with existing workflows and systems by 37%, and governance and compliance by 33%.

Those barriers have become central as businesses try to move AI from isolated use cases into routine operational processes. The findings suggest technical preparation and internal controls remain as important as model selection when companies assess whether AI programmes are delivering returns.

Raghu Malpani, Chief Product and Technology Officer at UiPath, said the gap between experimentation and broader deployment often reflects a failure to address those basics.

"The gaps between experimentation and enterprise deployment are known, and often come down to data, integration, and governance challenges, and the necessary enterprise context, that keep ROI out of reach," Malpani said.

UiPath's research placed particular emphasis on what it described as business orchestration, which it defined as the layer connecting data, systems, workflows, people and AI agents. It argued that this layer is becoming a differentiator for companies trying to expand their use of agentic AI.

According to the survey, 89% of respondents at organisations that had fully embedded orchestration across the enterprise said their agentic AI implementations had met or exceeded ROI expectations. That compares with much lower levels of adoption overall, with only 29% saying orchestration was fully embedded in their workflows.

The survey also asked respondents about the benefits they associated with orchestration. The most commonly cited result was giving employees more time for higher-value tasks, named by 51%, followed by increased integration among applications at 47%, improved oversight of business workflows at 44%, increased responsiveness to customers at 42%, and automation of complex business workflows at 37%.

Scaling challenge

The data points to a recurring pattern in enterprise technology adoption: companies can demonstrate proofs of concept but struggle to extend them across departments and legacy systems. In the case of agentic AI, that challenge appears more acute because the technology must work across data sources, rules, human oversight and existing automation tools.

Malpani said orchestration should sit at the centre of that effort.

"Resolving these challenges depends almost entirely on solving those three things. To begin, build an understanding of where the application of agentic AI can deliver the most value. From there, focus on the missing layer of business orchestration as the connector between data, systems, workflows, people, and agents. And finally, make sure governance and compliance are built in as guardrails. Together, these three factors will close those gaps and allow AI to scale across the enterprise," Malpani said.

The research also examined where companies expect AI agents to be used next. More than one in three respondents, or 36%, said they expect agents to play a significant role in enterprise workflows over the next 12 months.

When asked which kinds of workflows would receive the most attention, 52% said they were focusing on hybrid workflows that combine static and dynamic processes. That matters because such workflows are typically harder to manage than linear, rules-based processes, especially when they require both automated actions and human intervention.

Operational focus

The findings suggest companies are shifting attention from simple task automation to more complex operational environments. As businesses look to apply AI in areas where processes vary, involve several systems and require supervision, the underlying structure that coordinates those actions becomes more important.

UiPath has been positioning itself around that part of the market as companies assess how to organise agents within established business processes rather than deploy them as standalone tools. The survey indicates that while interest in agentic AI remains high, many large enterprises are still working through the operational groundwork needed to turn pilot projects into routine use.

"Orchestration serves as the foundation for agentic deployments, connecting agents with the data and systems they need to operate effectively, along with automation to execute and humans to oversee. Those who prioritise orchestration will surge ahead because they'll be able to ensure agents work within existing processes and guardrails to create real enterprise impact," Malpani said.