North American finance teams are ahead of peers in Europe and parts of Asia-Pacific in using artificial intelligence across multi-step workflows, according to research commissioned by Airwallex. The survey covered 1,279 finance decision-makers across EMEA, APAC and North America.
The data shows 37% of finance teams in North America use AI for multi-step execution, compared with lower levels elsewhere. In EMEA, 28% reported no AI execution in workflows, compared with 17% in North America.
The findings point to a regional gap that goes beyond planned spending. Intended increases in AI investment differed by only five percentage points between North America and EMEA, suggesting other factors are shaping adoption.
Forrester grouped AI execution in finance into four levels: no execution, single actions, multi-step preconfigured workflows, and autonomous execution with minimal human input. Autonomous execution clustered between 11% and 12% across major regions, but North America stood apart at the level below that, with more organisations already deploying AI across linked processes.
Examples include reading invoices, matching them against ERP records, validating foreign exchange rates and triggering payments, while humans retain final approval. The pattern suggests some finance teams are moving from small-scale experimentation to broader operational use.
Regional split
APAC presented a more uneven picture. The region recorded 23% with no AI execution, but that average masked substantial variation between markets.
Hong Kong, Australia and Singapore were among the region’s most advanced markets for multi-step execution. A third or more of finance leaders in those markets said AI was already being used in multi-step workflows: 36% in Hong Kong, 35% in Australia and 31% in Singapore.
Singapore also recorded APAC’s highest rate of autonomous execution with minimal human input, at 18%. In Hong Kong, 84% of respondents said AI was embedded in finance workflows in some form.
Elsewhere in APAC, progress was slower. Fragmented systems and uneven data integration were limiting AI use to narrower tasks rather than end-to-end processes. China was described as following a different path, with finance automation advancing more gradually despite the country’s broader strength in AI development.
EMEA was more cautious overall. The UK emerged as one of the region’s most advanced markets, with 17% saying AI already executes autonomously with minimal human input, but a similarly large share reported no AI execution at all.
Several continental European markets were earlier in adoption. The Netherlands recorded the highest rate of no AI execution among all surveyed markets at 35%, while France and Israel also lagged, with about one-third reporting no AI use and only 6% reaching autonomous execution.
Data barrier
The study identified fragmented data as the most common obstacle to scaling AI in finance. Across all respondents, 65% cited scattered data as the main barrier, while 58% of EMEA finance teams reported siloed or inconsistent data, compared with 47% in North America.
That matters because many finance functions still operate across separate systems for accounts payable, treasury, payroll, spend management and foreign exchange. When those systems are not connected, AI tools cannot draw on a full operational picture.
The research also found that 84% of organisations still require manual steps to complete finance workflows. At the same time, 66% of decision-makers said orchestration across the wider finance ecosystem was a priority when selecting platforms.
Infrastructure and regulation
Beyond internal systems, the study linked North America’s lead to broader infrastructure and labour market conditions. It cited Morgan Stanley data showing that the US accounts for roughly USD $109 billion in corporate AI investment and more than 60% of global data centre capacity.
That can make computing resources easier and cheaper to access than in parts of EMEA and APAC, where costs, energy constraints and cross-border data rules may add complexity. The report also said the US is home to roughly 60% of the world’s elite AI researchers.
Regulation was another dividing line. In EMEA, finance teams often face stricter obligations around auditability, explainability and governance before AI systems can be used in live workflows. In North America, the framework is more fragmented, allowing organisations to deploy and refine systems more quickly.
The study also suggested that hiring patterns are influencing adoption. Some North American companies are bringing data scientists and machine learning engineers directly into finance teams rather than relying solely on central IT departments, thereby speeding the development of workflow-specific tools.
