Finance leaders face a critical artificial intelligence (AI) opportunity despite more than half lagging in process automation.
In a recent report, intelligent document processing (IDP) firm Rossum found that, in spite of clear belief in AI’s transformative potential, 54.2 percent of finance leaders are yet to fully automate their processes, losing out on efficiency savings and fraud reduction.
The Document Automation Trends Report 2026: From Hype to Hard Returns surveyed 450 finance leaders across three markets (the UK, US, and Germany).
While technology exists to automate end-to-end processes, most businesses are not close to being able to implement it or reap the rewards, the study suggests. For example, one in 10 respondents said their processes still rely on manual data input and excel documents.
Finance leaders face critical AI opportunity
The appetite for process automation is strong, with finance leaders optimistic about AI’s potential, the study found. Nearly half believe the benefits of automation outweigh the security concerns, with 43.8 percent agreeing that automation and AI present more opportunities than risks for their business’s security.
The report also points to the rising threat of financial fraud, driven in part by duplicate invoices, concealed delivery shortages, and fraudulent supplier requests. Traditional detection methods often uncover issues only after the damage is done, whereas AI can identify patterns and anomalies that are difficult for humans to spot.
From analytics and document-use forecasting to invoice deadlines and payment cycles, AI’s ability to stay one step ahead gives finance leaders a crucial advantage. Nearly one-third of respondents (31.3 percent) are turning to AI to improve the accuracy of financial reporting and analysis, reinforcing their top priority of enhancing the reliability of financial data.
“In the world of finance, the measure of AI’s value is whether it can deliver accuracy, foresight, and protect against fraud – not just speed,” said Petr Baudiš, CTO and co-founder at Rossum. “This maturity gap is a stark reminder that legacy tools simply can’t handle today’s complexity. It’s been proven AI can eliminate manual data entry, but the next step is achieving verifiable accuracy and predictive power.”
Register for All Access: AI in PEX 2026!
AI adoption hurdles persist
The results also reveal that a trust gap remains when it comes to leaders’ confidence in AI’s broader value. More than a third of finance leaders (34.2 percent) gauge AI’s success using specific operational KPIs such as cost per invoice, processing time, and error rates. This reflects a highly granular approach, focused on performance and cash-flow outcomes rather than broader strategic impact or anecdotal indicators.
The PEX Report 2025/26, based on a global survey of more than 200 professionals, found that data quality and availability are the biggest barriers to AI adoption, cited by 52 percent of respondents. The next most prevalent hurdles are lack of internal expertise (49 percent), regulatory or legal concerns (31 percent), and resistance to change (30 percent).
These are followed by the high cost of implementation (27 percent), lack of clear ROI (23 percent) and ethical concerns (15 percent). Other challenges include clearly defined use cases, strategic alignment, scaling, and time constraints.