By Seppo Ikäheimo, Ph.D., Olli-Pekka Lumijärvi, Ph.D., and Jukka Sihvonen, Ph.D.
When it comes to leveraging AI in business, the finance function’s own distinctive requirements demand an equally distinctive approach.
Drawing on in-depth interviews with finance executives at various Finnish organizations, Seppo Ikäheimo and Jukka Sihvonen (of Aalto University School of Business) and Olli-Pekka Lumijärvi (MD of Profit Isle) reveal how finance functions are approaching artificial intelligence. Their study highlights a distinctive bottom-up adoption pattern and the key factors shaping implementation in the finance function.
Artificial intelligence (AI) is widely seen as a game-changer for finance functions, promising greater automation, predictive insights, and strategic decision support. However, real-world implementation remains cautious and uneven. Our study of nine Finnish firms—spanning manufacturing, retail, energy, consumer goods, and financial services—provides rare empirical evidence on how organizations are adopting AI in their finance functions.
Conducted with CFOs, finance transformation leads, and senior finance executives, the research captures experiences from companies ranging from mid-sized (€100m+ revenue) to multi-billion-euro giants. Finland offers a relevant context: firms have the highest adoption rate of AI within the European Union (EIB, 2025), a digitally mature EU economy subject to the General Data Protection Regulation (GDPR), global competition, and diverse industry pressures. The findings point to both universal challenges and finance-specific nuances that matter for leaders across Europe.
The Bottom-Up Adoption Journey: The Cumulative AI Adoption Model
A standout insight is that AI adoption in the finance function rarely follows the classic top-down enterprise IT playbook. Instead, it advances cumulatively from the ground up through four interconnected levels. The earlier stages continue even as organizations progress to more advanced ones.
- Level 1: Individual. Adoption typically begins with accessible tools such as Microsoft Copilot or similar generative-AI assistants. Finance professionals use them for personal productivity—drafting emails, retrieving information, basic analysis, and simple automation. Success depends heavily on individual initiative, supported by firm-provided training, the encouragement of experimentation, and knowledge-sharing between peers. The benefits are immediate but primarily personal.
- Level 2: Functional. At this stage, AI and related robotic process automation (RPA) target routine processes within finance: invoice processing, accounts payable and receivable, expense reporting, duplicate payment detection, and basic reporting. Several firms achieved large-scale automation in shared service centers, delivering measurable cost savings and quality gains. Large language models (LLMs) are emerging here, often integrated via enterprise resource planning (ERP) system updates. Most of the firms were at this level of AI adoption, but their plans were clearly for adoption at a more ambitious level.
- Level 3: Operational. This involves deeper process redesign. AI supports forecasting, anomaly detection, real-time analytics, and scenario modeling. Finance roles shift toward higher-value business partnering. Only a few firms in the study had moved partially into this phase, frequently tied to ERP modernization. One large manufacturing company had already begun reorganizing parts of its finance function around these capabilities.
- Level 4: Strategic. AI fundamentally transforms the finance function’s role in the broader business, enabling real-time calculation, ecosystem-wide data platforms, and strategic scenario planning. This level remains largely aspirational, with only limited piloting observed.
As a summary, the Cumulative AI Adoption Model (CAIAM) shows that progress is incremental, and evidence-driven (figure 1).
Quick wins at individual and functional levels build trust, demonstrate return on investment (ROI), and create momentum for broader transformation. The finance function’s core requirements around accuracy, auditability, and risk management explain this deliberate, bottom-up approach.
What Shapes Adoption in Finance?
Several interconnected factors influence how quickly and how far organizations advance. These factors cluster around technological readiness, organizational capabilities, and external pressures (Tornatzky & Fleischer, 1990).
Technological Foundations. Data infrastructure and quality consistently emerge as the biggest practical constraint. Finance data is often fragmented across ERP systems, planning tools, business intelligence platforms, and other sources, with varying levels of quality. Inconsistencies in metric definitions (for example, sales margin calculated differently in retail systems versus accounting) can lead to flawed insights. Executives described multi-year efforts in data cleansing, standardization, and governance as prerequisites for reliable AI use.
One leader explained: “We are making a significant work investment in data cleansing… . Only then can we support the business properly.” Legacy systems and heavy ongoing reliance on Excel add further complexity, particularly in larger and older organizations. Public companies face extra restrictions: external AI tools are often unsuitable for final financial figures due to audit and compliance needs.
There was a clear tendency to avoid “first-mover disadvantage” and wait for solutions to mature.
Vendor solution maturity also shapes decisions. Despite marketing hype, fully production-ready, finance-specific AI solutions remain limited. Many organizations adopt measured strategies, waiting for proven options or partnering closely with major platform providers (such as SAP or Microsoft) to influence development and reduce integration risks. There was a clear tendency to avoid “first-mover disadvantage” and wait for solutions to mature.
On the positive side, the rapid evolution and increasing accessibility of AI tools act as powerful enablers. Modern interfaces allow non-technical finance staff to leverage AI without deep programming skills, shortening project timelines and broadening participation. Security and governance requirements, while sometimes slowing rollout, provide necessary guardrails. Cybersecurity policies typically favor internal or tightly controlled solutions to protect sensitive financial and customer data.
Organizational Capabilities. Internal factors often prove most decisive. Visible leadership commitment from the C-suite (CEOs and CFOs) stands out as a critical accelerator. When executives actively champion AI—through steering groups, dedicated resources, personal involvement, and clear communication—adoption gains real momentum. Without it, initiatives remain fragmented and limited to individual efforts.
Organizational culture and change management play equally important roles. Generational differences exist, with younger employees often driving experimentation, and, interestingly, hands-on experience frequently converts skeptics across all age groups. Leaders must address resistance stemming from time constraints or uncertainty about benefits by crafting positive narratives focused on more meaningful work rather than job displacement.
Resource constraints, particularly human capital, create common bottlenecks. Teams are often “too busy running the business” to invest time in transformative projects. Competing priorities such as mergers can temporarily stall progress. Workforce skills development is therefore essential. Leading organizations conduct skills-gap analyses, roll out mandatory AI literacy training, and encourage experimentation while working to preserve core finance domain knowledge as routine tasks become automated.
Cross-functional knowledge sharing—through internal communities, peer learning sessions, and external benchmarking—helps spread successful practices and accelerate learning.
Environmental / External Pressures. Competitive dynamics generate awareness and sometimes public rhetoric, visible in investor communications, but rarely force immediate action. Pressure is strongest in financial services due to customer-facing applications and efficiency imperatives. Industry characteristics matter: capital-intensive sectors with long investment cycles and legacy infrastructure face different timelines than more agile service-oriented businesses.
Regulatory requirements, including GDPR and financial reporting rules, shape implementation approaches rather than blocking them outright. They reinforce preferences for controlled, auditable internal solutions. Relationships with vendors and consultants provide valuable external expertise and knowledge transfer, although organizations increasingly recognize the need to build internal capabilities to avoid over-dependence.
Three Finance-Specific Considerations
While many challenges are shared across functions, three factors appear with particular intensity in the finance function:
- Information Security Requirements. Handling sensitive financial and customer data drives a strong preference for internal, controlled AI systems over external platforms. Cybersecurity policies act as the primary governance mechanism, ensuring that data remains protected.
- Data Accuracy and Standardization. Precise, consistent metric definitions are non-negotiable to prevent erroneous reporting or decisions. Inconsistencies across systems create significant hurdles that must be resolved before scaling AI.
- Output Verification Practices. Professional skepticism remains strong. Finance teams typically validate AI-generated outputs against multiple sources before relying on them. This “human-in-the-loop” approach builds trust gradually and aligns with regulatory and audit expectations.
These considerations are not exclusive to the finance function but gain heightened importance given the function’s role in compliance, reporting accuracy, and high-stakes decision-making.
Practical Roadmap for Finance Leaders
The executives’ experiences suggest several actionable steps:
- Build Strong Foundations First: Prioritize data quality, governance, integration, and modernization of core systems (especially ERP). Expect this foundational work to take years for scalable results.
- Secure Executive Sponsorship: Establish clear ownership at the highest levels. Adopt a “beta mindset”—encourage experimentation, learning from failures, and iterative progress.
- Pursue Staged Implementation: Start with low-risk pilots at individual and functional levels to generate quick, visible wins. Use demonstrated ROI to unlock resources for operational and strategic initiatives.
- Invest in People: Develop comprehensive training programs, redesign roles for analytical and business-partnering work, and foster internal communities. Focus on preserving domain expertise while embracing new capabilities.
- Choose Partners Strategically: Work with established platform providers rather than chasing every new point solution. Influence vendor roadmaps where possible.
- Balance Caution with Ambition: Leverage finance’s strengths in governance, verification, and risk management to build trustworthy AI capabilities that deliver sustainable value.
Interviewees expect significant progress in the coming years, with AI enhancing efficiency, forecasting accuracy, anomaly detection, and overall business support.
Implications for European Businesses
European companies operate under comparable regulatory and competitive conditions. The Finnish lessons apply broadly: success hinges less on the latest technology and more on patient foundational work, strong leadership, cultural adaptation, and people-centric implementation.
Professional accounting bodies should update competency frameworks to reflect shifting roles—less routine processing, more interpretation, oversight, and strategic contribution. Regulators can take confidence that well-managed compliance requirements help shape responsible adoption paths.
For business schools, it would be of great value to follow the development of AI in firms and consultancy agencies and collaborate with them to maintain relevance in business academic research, education, and the employability of their students.
Looking Ahead
AI will not replace skilled finance professionals; it will elevate them. Routine tasks and basic analysis will increasingly be handled by intelligent systems, freeing talented people to focus on higher-value work: strategic analysis, proactive business partnering, judgment-based decision-making, and interpreting complex, uncertain environments.
Yet the path forward is not uniform. The four foundations of successful AI adoption—data readiness, skills, culture, and leadership—are evolving at different speeds. The skills constraint is closing fastest. What once required rare data-science expertise can now be accessed “on prompt” through increasingly capable generative AI tools. As technical barriers fall away, the harder, distinctly human challenges come into sharper focus: organizational culture and leadership.
The Finnish finance leaders interviewed for this study convey a message of pragmatic optimism. They see substantial potential in AI for finance, from major efficiency gains and sharper forecasting to more timely, insightful decision support. However, they are equally clear that realizing the potential demands disciplined execution, incremental proof points, and a deeply human-centered approach.
In a fast-evolving technological landscape, the winners will be those organizations that build strong foundations today—robust data infrastructure, continuous skills development, adaptive culture, and committed leadership—while remaining flexible enough to seize tomorrow’s opportunities.
Seppo Ikäheimo, Ph.D. 
Olli-Pekka Lumijärvi, Ph.D. 
Jukka Sihvonen, Ph.D. 
