Celonis survey: 9 in 10 IT leaders have concerns about AI adoption

A new global study by Celonis, a global leader in process mining, has found that while most organisations are rapidly embracing Artificial Intelligence (AI), 90% of IT leaders remain concerned about its growing use within their companies. Top worries include increased security risks, pressure to prove return on investment, and a widening skills gap.

The 2025 Celonis AI Process Optimisation Report, based on a survey of 1,620 business leaders from around the world, including 10% from Australia and 25% from the broader APAC region, reveals that nearly half of IT leaders (45%) cite security and data breaches as their number one concern. This is followed by challenges in demonstrating ROI and shortage of relevant skills (both 37%). Notably, nearly a third of respondents say expectations for AI are often unrealistic.

Despite these concerns, few leaders are focusing on the underlying processes that enable AI to succeed. Pascal Coubard, APAC Lead at Celonis, said, “Process shortcomings ranked surprisingly low, which suggests many IT leaders may not yet recognise how critical optimised processes are to effective AI deployment. Issues like data quality and system inefficiencies, which are key drivers of AI failure, can often be solved through process intelligence.”

While 25% surveyed were in the IT sector, 25% were from the Supply Chain sector, 25% from Finance/Shared Services and 25% from Operations/Process Improvement. A company needed a minimum revenue of $US500 million to participate with 80% of companies in the $US2 billion to $US10 billion range.

Key industry findings in the other sectors were:

• Supply Chain: AI is already widely used for automated inventory management (55%) and warehouse automation (49%). While only 40% currently apply AI to transport network and route planning, 96% plan to adopt it within the next 12 months, signalling a strong forward-looking investment.

• Finance and Shared Services: Leaders here are optimistic, with 70% having a mandate to invest in AI to improve efficiency. Today, more than half (54%) are automating routine tasks, with growing interest in areas like fraud detection, cash flow optimisation, and regulatory compliance. By mid-2026, at least 90% expect to have deployed AI across every major use case.

• Process Improvement and Operations: Over 50% of leaders in this sector say they are already using AI for continuous improvement and process automation. Even in lower-adoption areas such as identifying new optimisation opportunities, most plan to scale up in the next year. Nearly 9 in 10 believe intelligent automation will deliver the most business value over the next five years.

“AI is no longer experimental. It’s being embedded into the fabric of the way businesses operate,” Coubard added. “But to fully realise its potential, companies must feed AI with the process knowledge and business context provided by process intelligence. Otherwise, they risk automating inefficiency and missing the opportunity entirely. And when processes work, AI works.”