The geography of AI in finance: Why North American teams are pulling ahead

Forrester data from 1,279 finance leaders reveals a regional gap in AI execution, and the causes run deeper than budget.

North American finance teams are more likely to be running AI across multi-step workflows, from automated reconciliation to real-time cash flow forecasting. European teams, it seems, have a lot of catching up to do.

We recently commissioned Forrester Consulting to survey 1,279 finance decision-makers across EMEA, APAC, and North America (NA), and some of the results have been eye-opening. 28% of EMEA finance teams report zero AI execution in their workflows. In North America, that figure drops to 17%. And the gap persists despite similar spending intent, with only five percentage points separating the two regions on planned AI investment increases.

So what's driving it? The divergence is driven by a set of structural factors that compound across data, regulation, talent, and infrastructure maturity. For finance leaders operating across multiple regions, understanding these factors is the first step toward closing the distance in your own organisation.

Unless stated, all statistics referenced below come from "Building An AI-Ready Finance Function", a commissioned study conducted by Forrester Consulting on behalf of Airwallex, June 2026.

How three regions stack up on AI in finance

Forrester breaks AI execution into four tiers, from no execution through singular actions and multi-step preconfigured workflows to fully autonomous execution with minimal human input.

North America leads on multi-step execution

While autonomous execution clusters between 11% and 12% across major regions, North America separates itself one tier below, with 37% of finance teams already using AI for multi-step execution. These are workflows where AI coordinates actions across multiple systems, such as reading invoices, matching them against ERP records, validating FX rates, and triggering payments with with humans providing final-step approval. The region’s lead suggests finance teams are moving beyond experimentation and using AI to orchestrate end-to-end operational processes at scale. Stronger investment in enterprise infrastructure, greater pressure to improve efficiency, and faster adoption of integrated finance platforms are likely accelerating that shift.

APAC is not a single maturity curve

APAC reports 23% with no AI execution. That regional average, though, conceals sharp internal variation. 

• Singapore, Hong Kong and Australia are the most advanced execution environments in the region, with a third of finance leaders in each region (36% HK, 35% AU, 31% SG) already using AI to execute multi-step workflows. This is largely enabled by more consolidated financial infrastructure and stronger incentives to automate complex, multi-entity finance operations. Singapore stands out in particular, with 18% of respondents reporting AI operating autonomously with minimal human input – the highest rate in APAC. Hong Kong also shows broad-based adoption, with 84% of respondents indicating that AI is already embedded in finance workflows in some form. Australia follows a similar trajectory, with adoption concentrated in multi-step execution rather than isolated task automation. By contrast, other parts of the region remain earlier in the adoption curve, where fragmented systems and uneven data integration continue to limit AI to narrower, task-level use cases rather than end-to-end workflow orchestration.

• China, reflects a different set of structural conditions. While it is a global leader in AI development, enterprise finance automation is progressing more gradually, shaped by different operating economics and lower pressure to automate back-office finance functions at scale.

Across APAC, the divergence is less about awareness of AI and more about whether finance systems are sufficiently connected to support orchestration across workflows, entities, and currencies.

EMEA presents a more cautious picture

EMEA teams are more likely to remain at the single-task stage or not use AI at all.

The UK is emerging as the region’s most advanced market, with 17% saying AI already executes autonomously with minimal human input, on par with leading APAC markets. Yet that progress is offset by an equally large cohort still reporting no AI execution at all, highlighting a widening maturity gap even within advanced economies. Continental Europe remains earlier in the journey. The Netherlands recorded the highest rate of no AI execution across all surveyed markets at 35%, while France and Israel also lag, with roughly one-third reporting no AI use and only 6% reaching autonomous execution. This is reflective of a combination of stricter regulatory environments, fragmented financial systems, and a more cautious approach to operational risk and governance in finance functions.

Before you invest in AI, fix what's underneath it

The most advanced teams are not just adopting AI, they are operating on connected finance infrastructure where payments, FX, spend, and reconciliation already sit in the same system.

Your AI can only see what your data lets it

The data underneath the models determines everything. Forrester's findings make that uncomfortably clear. 58% of EMEA finance teams report siloed or inconsistent data, compared with 47% in NA. Across the full survey, 65% of decision-makers cite scattered data as the primary barrier to scaling AI.

In practice, this could look like your AP team in London processes invoices through one system. Your treasury team in Singapore manages FX exposure in another. Your US entity runs payroll on a third. An AI agent asked to optimise your global cash position can't see across those three systems. It's working with a partial map, and partial maps produce partial answers.

More advanced organisations have spent the last cycle consolidating onto integrated finance systems, giving AI access to a unified operational dataset. Others remain constrained by fragmented architectures, where pilots rarely progress into production because underlying data structures are inconsistent or incomplete.

North America's investment gap is wider than it looks

According to Morgan Stanley, the US accounts for roughly $109 billion in corporate AI investment, more than the rest of the world combined. North America also holds over 60% of global data centre capacity. For a CFO in New York, spinning up the processing power to train and run AI models is straightforward. The data centres are local, the costs are competitive, and the infrastructure is mature.

For teams in parts of EMEA or APAC, that same processing power is often pricier, constrained by energy caps, or tangled in cross-border data transfer agreements that add friction before a single model is trained.

Within APAC, the data integration picture fractures along similar lines. Markets like Hong Kong and Australia, where digital infrastructure is more consolidated, are progressing faster. Others remain tethered to earlier stages of the curve because the plumbing hasn't caught up.

Why regulation and hiring patterns shape AI speed

Regulation and talent don't show up in most AI readiness dashboards; the Forrester data suggests they should.

EMEA's compliance-first culture

EMEA has built the most comprehensive regulatory architecture for AI anywhere in the world. The EU AI Act and GDPR mean that before an AI model touches a live finance workflow, teams must demonstrate that its outputs can be audited, explained, and governed. In practice, that looks like months of documentation, impact assessments, and sign-off chains before a pilot goes live.

Some of that caution will prove logical foresight in high-stakes finance contexts, where a misclassified transaction or an opaque credit decision carries regulatory consequences. European regulators are building a durable framework, and the rest of the world will likely converge toward something similar over time. That timing cost shows up clearly in the Forrester data. Within EMEA, the UK is advancing faster, and a lighter regulatory approach correlates with higher AI execution maturity.

By contrast, NA operates under a more fragmented, voluntary framework that lets finance teams deploy iteratively. They can launch a "good enough" model, monitor its performance, and refine it in production rather than waiting for a fully compliant version before going live.

The talent accelerator

Morgan Stanley estimates that the US houses roughly 60% of the world's elite AI researchers. NA finance teams now hire data scientists and machine learning engineers directly into the function rather than routing requests through a centralised IT team. This cross-pollination lets them build bespoke workflows, a controller building an automated close process with a data engineer sitting three desks away, rather than submitting a ticket to IT and waiting six weeks.

Regulatory and talent pressures interact. In regions where compliance requirements add verification steps before rollout, finance teams need people who understand both the AI and the governance. That hybrid skill set, part data scientist, part compliance specialist, is scarce everywhere. The bottleneck tightens where the compliance bar is higher and the talent pool is thinner.

What the gap means for your infrastructure choices

Strip the regional analysis back to its foundation, and the data points to a single precondition for AI scale in finance: connected data flows across every system your team touches.

65% of decision-makers globally identify disconnected data as a barrier. 84% still require manual steps to complete finance workflows. Organisations have consolidated platforms, but consolidation hasn't translated into connectivity. Data sits in fewer systems, yet those systems still don't talk to each other.

This is where infrastructure design becomes the constraint or the unlock. AI requires finance systems where payments, FX, accounts, and spend data are not just stored centrally, but structurally connected across workflows.

Forrester's findings on vendor selection reinforce this direction. 66% of decision-makers prioritise orchestration across the broader finance ecosystem when evaluating platforms. Winning architecture connects systems rather than replacing them. Multi-region businesses should choose infrastructure that works across borders, currencies, and regulatory environments without spawning new data silos in every market they enter. This is the difference between system consolidation and true operational connectivity.

Source: Unless stated, all statistics referenced in this article come from "Building An AI-Ready Finance Function", a commissioned study conducted by Forrester Consulting on behalf of Airwallex, June 2026.