Berlin – September 15, 2026 -- Broken business processes are costing companies an average of $1.55 million each as artificial intelligence initiatives collapse under outdated workflows, according to a new report from Camunda.
Process failures drive 72% of AI project setbacks, study finds
The AI Process Gap report, commissioned by Camunda and conducted by Sapio Research, found that 72% of enterprises have seen AI projects fail due to process-related challenges. Another 72% warn of spiraling costs if they fail to overhaul their processes, while 82% believe AI investments will flop without deep process re-design.
The core issue: 79% of companies say bolting AI onto existing workflows meets less internal resistance than redesigning them from scratch. Yet 82% estimate it will take up to five years to make core processes truly AI-ready, and 61% admit their process re-design efforts cannot keep pace with AI adoption speed.
Compliance failures trace back to broken workflows in 84% of cases
Governance risks are mounting alongside the process gap. Some 40% of companies faced an AI-related compliance or governance issue in the past 12 months, and 84% of those incidents stemmed from process-related problems. Nearly half (48%) of firms fear AI agents could spiral out of control once process changes are made, while 79% of employees worry their own AI use could trigger compliance violations. Almost all respondents (96%) expect process challenges to fuel further compliance problems ahead.
Workforce trust gap widens as 46% consider quitting over poor AI rollouts
A perception gap has emerged between leadership and staff: 89% of companies believe AI has boosted team productivity, but only 64% of employees agree. Nearly half (48%) of companies have already scaled back AI use after seeing negative effects on staff workability, and 46% of employees are now considering leaving their jobs due to poorly implemented AI tools.
Additional findings show 69% of employees were not adequately consulted on how AI would support their daily work, 65% believe they could use AI far more effectively under different implementation, 44% have had to manually correct AI outputs due to flawed underlying processes, and 41% used AI solely to satisfy internal mandates rather than improve outcomes.
Kurt Petersen, SVP Customer Success at Camunda, said employees are not rejecting AI but are rarely included in decisions about how the technology should function in their specific work environment, resulting in tools that fail to help them and a culture where AI use becomes a compliance checkbox rather than a driver of business value.
Camunda surveyed 1,000 decision-makers and 5,000 employees across four markets
The findings draw on two Sapio Research surveys conducted in July and August 2026: one covering 1,000 process decision-makers (500 US, 250 UK, 125 Germany, 125 France) at companies with at least 1,000 employees, and a second covering 5,000 employees (2,000 US, 1,000 UK, 1,000 Germany, 1,000 France) at similarly sized firms that have rolled out AI or automation.