June 2026 Summaries

AI is accelerating finance tasks such as reconciliations, reporting, and customer-response drafting, but many organizations have not achieved faster closes, approvals, or audits because bottlenecks have shifted to disconnected data, manual handoffs, reviews, and sign-off chains. A Forrester Consulting study commissioned by Airwallex found that while one-third of finance leaders consider workflows fully digital, 84% still require manual intervention, and 65% identify fragmented or inconsistent data as the primary barrier to scaling AI. Finance leaders must also preserve accountability amid risk, compliance, audit, and trust concerns, since AI-generated outputs still require explanation and human approval. The proposed next step is not simply deploying more AI, but redesigning finance operating models around connected workflows, consistent data movement, automated routing and audit trails, and exception-based human oversight so that machine-speed work can translate into quicker decisions, stronger controls, and improved business outcomes.

A Forrester Consulting survey of 1,279 finance decision-makers, commissioned by Airwallex, finds that North American finance teams are ahead of EMEA and much of APAC in deploying AI across multi-step operational workflows, with 37% reporting such use compared with higher rates of limited or no execution in Europe. While planned investment increases are similar across regions, the gap is attributed to differences in connected data infrastructure, computing capacity, regulation, and access to AI talent rather than spending alone. EMEA has the highest reported rate of no AI execution at 28%, with siloed data affecting 58% of its teams, while stricter governance requirements under frameworks such as GDPR and the EU AI Act can lengthen deployment timelines. APAC shows substantial variation, with Hong Kong, Australia, and Singapore more advanced due to consolidated infrastructure, and Singapore reporting the region’s highest autonomous-execution rate. Across all regions, disconnected data is identified as the leading obstacle to scaling AI, and the findings suggest that finance organizations need interoperable systems linking payments, foreign exchange, spending, accounts, and reconciliation before AI can reliably automate complex, cross-border workflows.

AI agents are increasingly able to triage issues, generate code, and deploy fixes, reducing engineering capacity as the central constraint on software development while exposing organizational coordination as a new bottleneck. The passage argues that many companies still apply AI tools within traditional sequential handoffs among product, design, engineering, and QA teams, rather than redesigning workflows so people closest to customer problems can act directly. Drawing on examples from Anything and Airwallex, it describes support, finance, and other non-engineering teams using AI to investigate problems and build internal tools, with human attention shifting toward product strategy, customer insight, differentiation, and distribution. It also contends that domain expertise is becoming more valuable than coding ability alone, enabling specialists in fields such as recruiting, logistics, real estate, and finance to create software shaped by their operational knowledge. As AI capabilities advance rapidly and software becomes cheaper to produce, the passage concludes that companies’ ability to adapt their organizational structures, including for products used by AI agents, may determine their competitive speed and effectiveness.