72% of Organizations Say Process-Related Challenges Have Caused AI Initiatives to Fail and it is Costing Them Millions

8 in 10 organizations say their AI investments will be a bust without more investment in process re-design

BERLIN & SAN FRANCISCO--(BUSINESS WIRE)--New research from Camunda, the enterprise platform for agentic orchestration, has revealed how broken business processes are causing AI initiatives to fail at a staggering rate. According to the report, The AI Process Gap, 72% of organizations say that process-related challenges have caused AI initiatives to fail at an average cost of $1.55 million per business. Furthermore, another three quarters (72%) say that AI costs will spiral unless they gain better control over their business processes, with 82% believing that their AI investments will be a bust without more investment in process re-design.

The report highlights how many organizations are attempting to bolt AI onto business processes, which weren’t built in the AI era, vs taking the outcome-centric approach and re-designing them from scratch. That’s because 79% of organizations state that adding AI onto existing processes creates less internal resistance, with 82% saying adapting all their most important processes for AI will take up to five years.

“Organizations are rapidly adopting AI. But they are applying it to processes designed for a world before AI, then wondering why the return on investment falls short,” said Kurt Petersen, SVP Customer Success at Camunda. "Almost two-thirds (61%) of respondents say business process re-design simply can't keep pace with how fast they need to move on AI. So it's no surprise that under pressure to operationalize AI quickly, most organizations reach for the quick fix of bolting AI onto processes that were never built for it."

Growing Governance and Compliance Exposure

The report reveals that legacy or poorly designed processes are a leading contributor to the AI-related compliance and governance failures organizations already face.

• 40% of organizations have suffered an AI-related compliance or governance issue in the past 12 months. And process-related issues have contributed to the vast majority (84%) of them.
• Almost half (48%) of organizations fear AI agents going rogue if they make changes to a process.
• 79% of employees fear that their use of AI will lead to a compliance issue in their organization and almost all (96%) expect process challenges will contribute to future AI-related compliance issues.

“Recognizing the risk has not translated into action. Compliance concerns are also slowing the re-design work that could reduce exposure, with fears of AI operating outside its intended boundaries making organizations reluctant to change the processes it supports. Those concerns are not unfounded, particularly given recent incidents of AI agents acting alone, but innovation is moving too quickly for organizations to stand still,” added Petersen.

The Leadership-Employee Gap

Employees are also feeling the impact of AI being introduced into processes that were never designed for it. Many are left to absorb the friction, quietly working around the AI tools that were supposed to make their jobs easier. So while 89% of organizations say AI is making their teams more productive, only 64% of employees agree.

In fact, almost half (48%) of organizations say they have rolled back on the use of AI because it’s had a negative impact on employees’ ability to do their job. And the consequences go further, with 46% of employees saying they would consider changing jobs due to poorly implemented AI.

The report also reveals the following:

• 69% of employees were not fully consulted about how AI would be used to support their day-to-day work.
• 65% of employees say they could use AI a lot more effectively if it was implemented differently.
• 44% of employees say they have had to manually override AI outputs because the underlying process wasn't set up correctly.
• 41% of employees say they have used AI to artificially comply with their organization's AI mandate when there was no need to do so.

“Employees aren’t rejecting AI, but they are rarely fully involved in decisions about how it will be used in their work. This can leave employees being pushed to use tools that don’t help them, creating a culture where AI is adopted to satisfy a corporate mandate rather than help employees do more of what actually matters. The bottom line is that processes designed before the AI era cannot support the technology effectively without being re-engineered, no matter how much is spent on agents and models,” said Petersen.

For more information:

- Download the report.
- Explore more around the survey on our blog.
- Register for the webinar The $1.55 Million Mistake: Why AI Is Failing on Old Processes on September 29, 2026, 10:00 AM EDT / 4:00 PM CEST.

About the report

Camunda commissioned Sapio Research to conduct two surveys. The first covered 1,000 process decision makers (500 US, 250 UK, 125 Germany, and 125 France), comprising senior IT, operations, and transformation leaders involved in automation and AI strategy and execution at organizations with at least 1,000 employees. The second covered 5,000 employees (2,000 US, 1,000 UK, 1,000 Germany, and 1,000 France) whose day-to-day work has had AI or automation introduced into it, at organizations of the same size. Both surveys were conducted online in July and August, 2026.

About Camunda

Camunda is the enterprise platform for agentic orchestration, enabling organizations to coordinate AI agents, people, and systems across complex, end-to-end business processes. With built-in governance, auditability, and human oversight, Camunda gives enterprises the control they need to move AI from pilots to production — safely and at scale. Trusted by over 700 organizations worldwide, including 9 of top 10 US banks, Camunda helps enterprises boost operational efficiency, accelerate time-to-value, and deliver better customer experiences. For more information, visit camunda.com.