Why Corporate Treasurers Are Hesitating to Adopt Artificial Intelligence
Amsterdam, Tuesday 28 July 2026
A new study reveals that despite 46% of corporate treasurers evaluating artificial intelligence for cash forecasting, only 4% have deployed it, primarily due to persistent data quality concerns.
The Adoption Gap: Interest Versus Implementation
The “AI in Treasury: Accuracy, Intelligence and the Future of Cash Forecasting” study, published on 27 July 2026 by Treasury Intelligence Solutions (TIS) and EuroFinance, highlights a stark contrast between corporate curiosity and actual technological deployment [1][3]. While 46% of corporate treasury teams are actively evaluating AI solutions for cash forecasting, 47% have no immediate plans to implement them [2][3]. This represents a narrow 1% difference between those actively exploring the technology and those keeping it at arm’s length for the time being [1][2]. This hesitation is further underscored by the fact that only 4% of treasury departments have successfully transitioned AI processes into active production, while 7% are not considering the technology at all [2][3].
The Medium-Term Outlook
Despite the low rate of current deployment, treasury leaders recognise that artificial intelligence will play a significant role in the medium term. According to the survey, 16% of respondents project that AI will become “mission-critical” or “business-critical” for their operations within the next 24 months, leading up to 27 July 2028 [1][3]. Meanwhile, 43% of those surveyed view the technology as “important but complimentary” to their existing systems [1][2]. Combined, a substantial 59% of treasury leaders believe AI will hold a valuable position in their departments in the near future, even if they are currently holding back on implementation [1][2].
The Data Quality Conundrum and Legacy Roadblocks
The primary obstacles preventing corporate treasurers from taking the leap into AI adoption are deeply rooted in structural and systemic challenges. Chief among these are poor data foundations, legacy system integration difficulties, and a fundamental distrust of AI-generated outputs [1][3]. Indeed, over 50% of the surveyed treasurers identified data quality as the single largest impediment to achieving forecast accuracy [1]. Without clean, reliable, and consolidated data, the sophisticated algorithms underpinning machine learning models cannot function effectively, leading to fears of flawed financial forecasts [GPT].
Exposing Foundational Weaknesses
This perspective is echoed by technology providers who interact closely with corporate finance departments. Charles Bennett, the Chief Product Officer at TIS, points out that AI should be viewed as a powerful lens that exposes existing weaknesses in data quality and forecasting assumptions—weaknesses that traditional models often quietly tolerate [1][2]. Instead of allowing organisations to work around these foundational flaws, Bennett notes that AI surfaces these issues much faster and more clearly, forcing corporate finance teams to confront and resolve the root causes of their data deficiencies [1][3].
Real-World Applications and Volatility Lessons
While many treasurers remain cautious, some major multinational organisations have already begun pioneering these technologies. For instance, Siemens Energy has successfully implemented machine learning and AI specifically for cash forecasting [1][2]. However, Robert McAnally, the Senior Vice President and Head of Treasury and Corporate Finance at Siemens Energy, warns that treasurers cannot afford to ignore the underlying data and assumptions that feed these AI-driven models [1][2]. The technology is only as good as the inputs it receives, meaning human oversight remains indispensable [GPT].
Navigating Geopolitical Disruptions
The dangers of relying on outdated forecasting assumptions became particularly evident at Abu Dhabi Ports, which utilised AI as a decision-support tool to manage liquidity buffers and cash flows amidst intense geopolitical and macroeconomic volatility [1]. Khaloud Alhammadi, the Director of Treasury at Abu Dhabi Ports, explained that traditional forecasting models relied heavily on historical data, stable customer behaviour, and predictable payment terms [1][2]. When macroeconomic and geopolitical disruptions invalidated those assumptions, forecasting accuracy dropped rapidly, regardless of how advanced the underlying technology was [1][2]. Alhammadi emphasised that leveraging AI was “not a silver bullet,” but it served as an invaluable tool for identifying deficiencies in data, corporate governance, and forecasting assumptions [1].
The Path Forward: Augmentation, Not Replacement
Ultimately, the successful integration of AI into corporate treasury will require a shift in how the technology is perceived. Rather than viewing machine learning as a fully automated replacement for human analysts, experts advocate for a collaborative approach. Bennett asserts that AI should augment a treasurer’s expertise rather than replace it, noting that the treasury leaders who will succeed in the coming years are those who treat AI as a catalyst for better decision-making, rather than merely a tool for reducing manual workloads [1][2].
A Trillion-Dollar Market Opportunity
The scale of modern treasury operations underscores the necessity of solving these data quality issues. For context, TIS supports over 300 B2B clients and 39,000 active users across 150 countries, managing a massive $2.9 trillion in annual payment volume with 11,000 bank connection options [1][2]. Managing liquidity at this scale requires absolute precision [GPT]. For Benelux fintech founders and venture capital investors, this widespread hesitation among corporate treasurers presents a clear market opportunity: the demand for highly secure, explainable, and enterprise-grade AI solutions that can seamlessly integrate with legacy systems while actively correcting data deficiencies [GPT].