A new survey on enterprise artificial intelligence (AI) adoption has found a stark disconnect between ambition and execution. Nearly three-quarters of organizations acknowledge their agentic AI implementations fall short of expectations.
The 2026 State of Agentic Orchestration and Automation report, released this week by Camunda, a leader in agentic automation, reveals that while 71% of organizations claim to use AI agents, only 11% of agentic AI use cases have actually reached production in the past year. Perhaps more concerning, half of respondents believe poorly governed agentic AI risks amplifying existing automation failures.
The research, gleaned from 1,150 senior IT leaders, identified trust as the primary obstacle preventing wider AI agent deployment. A staggering overwhelming 84% of organizations expressed concern about business risks when IT lacks appropriate controls, while 80% cited insufficient transparency around AI use. Additionally, two-thirds worry about compliance issues related to AI agents.
Trust deficits are shaping conservative adoption patterns. Currently, 80% of deployed AI agents serve as basic chatbots or assistants that summarize information rather than handling mission-critical functions. Nearly half operate in isolated silos rather than integrating into comprehensive business workflows.
“The promise of agentic AI is undeniable, but trust remains the key barrier to adoption,” said Kurt Petersen, senior vice president of customer success at Camunda. “Without clear guardrails and visibility, agents will stay at the edge of the business.”
Despite implementation challenges, the report documents significant automation benefits: 95% of organizations reported business growth from process automation over the past year, up from 87% previously. Companies have automated an average of 48% of their processes and anticipate reaching 64% automation.
Investment reflects this optimism, with 79% of organizations planning to increase automation spending by an average of 20% over two years. However, growing technology complexity threatens these gains. More than three-quarters report exponentially increasing endpoint volume and diversity, while 85% say they need better tools to manage process intersections.
The report introduces agentic orchestration as a potential path forward, combining structured process orchestration with dynamic AI reasoning. While 88% agree AI must be orchestrated across business processes for maximum benefit, and 90% recognize the compliance necessity of treating AI as another managed endpoint, 85% admit they haven’t achieved the process maturity required for effective implementation.
“Agentic orchestration, not standalone agents, is the key to closing the AI vision-reality gap,” Petersen said, positioning integrated orchestration as the bridge between today’s experimental AI deployments and tomorrow’s business-critical capabilities.
