Building the Right Foundation: Why Banks Can't Leapfrog to AI

"The future has not been written. There is no fate but what we make for ourselves." - Terminator

Key Insights

• Generative AI (Gen AI) could add $200 billion to $340 billion annually to global banking, primarily through productivity gains (McKinsey, 2024).
• EY survey reveals that 99% of financial services leaders are deploying AI in some capacity, with all respondents either using or planning to use generative AI (EY, 2023).
• Successful AI implementation requires a three-stage evolution: legacy tech debt resolution, data fluency, and AI integration
• Banks must build proper infrastructure before AI adoption, similar to needing maintenance capabilities before operating a high-performance vehicle
• 87% of financial institutions are prioritizing compliance and ethics in their AI initiatives (PwC Survey, 2023)
• Real-time data analytics serve as the crucial bridge between legacy systems and AI implementation

As we enter Q4 2024, banks and credit unions are under pressure to implement artificial intelligence across their operations. The allure is clear: generative AI alone could add between $200 billion and $340 billion in annual value to the banking sector by enhancing productivity and operational efficiencies plus the global market for conversational AI alone is projected to grow from $4.2 billion in 2021 to over $15.7 billion by 2027 (MarketsandMarkets, 2021), promising transformative capabilities for successfully implementing it.

Yet a sobering reality remains - While nearly all financial institutions have adopted some form of AI, transitioning to scalable Gen AI use demands new capabilities in governance, infrastructure, and regulatory alignment.

Figure 1: Growth of the Conversational AI Market (2021-2027)

The Current State of AI in Banking

The banking sector finds itself in a unique position. Despite being among technologically advanced industries, banks are struggling to keep pace in AI adoption. While major tech companies have invested billions—like Microsoft’s $13 billion commitment to OpenAI and integration of ChatGPT into Bing—financial institutions are hindered by fundamental implementation challenges.

Nvidia’s CEO, Jensen Huang, captures the broader significance, stating, “Generative AI is the largest TAM -total addressable market- expansion of software and hardware that we’ve seen in several decades.” These investments underscore the growing emphasis on AI development across industries, highlighting the gap between tech companies and financial institutions in AI readiness.

The analogy is clear: moving from a base model to a supercar requires the infrastructure to maintain, support, and drive it effectively. Most banks, however, still lack the foundational infrastructure for successful AI integration. Core requirements like complex on-premise deployment, rigorous security testing (such as penetration testing), and real-time disaster recovery capabilities reveal gaps that prevent banks from fully embracing advanced AI technologies.

McKinsey’s research indicates that a centralized operating model for AI is particularly effective for banking. This model allows financial institutions to efficiently allocate talent, rapidly adapt to the latest AI developments, and maintain cohesive governance. Banks lacking this centralized approach may find it challenging to scale Gen AI initiatives beyond the pilot stage.

Regulatory Pressures Intensify

In parallel, regulatory requirements around AI have also escalated. The year 2023 marked a pivotal shift, with significant regulatory moves such as the Biden administration’s executive order on AI and the European Union’s AI Act. These frameworks stress that bypassing foundational steps isn't just impractical - it's potentially dangerous from a compliance perspective.

Banks face several core barriers that make immediate AI adoption challenging:

• Siloed data sets limit AI’s effectiveness
• Regulatory compliance hurdles increase the complexity of integration
• Fear of failure dampens innovation initiatives
• Unclear internal ownership creates ambiguity around AI and emerging technologies

The Three-Stage Evolution: A Necessary Journey

The path to successful AI implementation isn't a leap - it's an evolution through three distinct stages. Much like crypto adoption, it's flashy on the surface but requires building a solid and reliable infrastructure to deliver real value. As industry research shows, financial institutions achieving the highest ROI follow this evolutionary path.

Stage 1: Legacy Tech Debt Resolution

This foundational stage addresses fundamental infrastructure challenges that plague many financial institutions. Banks must confront:

• Outdated systems challenges
• Integration requirements
• Modernization needs

Stage 2: Data Fluency

The critical middle stage is achieving true data fluency. Some European banks are already demonstrating this in practice, utilizing real-time analytics for customer communications. This stage requires developed capabilities in:

• Real-time data access
• Customer engagement frameworks
• Analytics-driven capabilities

The market validates this approach: in 2024, the estimated market for multimodal AI applications is expected to exceed $12.4 billion (Gartner, 2021), driven by institutions that have successfully built these foundational capabilities.

Stage 3: AI Integration

Only after establishing data fluency can banks meaningfully integrate AI. This progression highlights the importance of complementary technologies like blockchain. While blockchain spending in banking is expected to reach $16.2 billion by 2025 (up from $2.9 billion in 2021), its primary role is to enhance data integrity, security, and transparency.

These qualities create a robust data foundation that can support advanced AI applications, but blockchain and AI investments remain largely independent, each contributing uniquely to banking's digital transformation.

Practical Implementations in Action:

Real-world baseline applications demonstrate what's possible with proper infrastructure. Banks and Credit unions implementing robust data analytics frameworks have successfully deployed:

• Predictive credit line analysis using real-time customer data for holiday season planning
• Automated flagging of personal transactions that deviate from normal amounts
• Proactive customer alerts based on account activity

These capabilities represent the essential foundation for future innovation. According to PwC, 87% of financial institutions prioritize compliance and ethics in their AI initiatives, recognizing that proper infrastructure must precede advanced AI implementation.

Moving Forward

As Mohit Kansal emphasized at the 2024 MIT Sloan Fintech Conference, "Fintechs will not be successful if they don't pay attention to regulation and figure out how to make regulation work in their favor." This insight applies directly to AI adoption, where regulatory compliance and infrastructure readiness go hand in hand.

CIOs developing technology roadmaps are unequivocal: organizations are deluding themselves when they want to go straight to AI without keeping their eye on the ball of fundamental infrastructure. The desire to leap directly to AI implementation often reveals a critical misunderstanding of what makes AI valuable in banking: the quality of underlying data and "state of the art" systems.

Conclusion

The path to AI in banking requires proper foundations and necessary evolutionary steps. Financial institutions that build their infrastructure today - resolving legacy tech debt and achieving data fluency - will be positioned to implement AI effectively tomorrow. In an industry where security, compliance, and reliability are paramount, evolution beats revolution every time.

*****************

Maria Echeverria has written this article. Maria’s mission-driven purpose is to help banks in the US, Canada, and abroad deliver world-class customer experiences and streamline operational efficiency by leveraging real-time transactional data. With over 15 years of experience in sales, finance, market research, and consultancy, Maria specializes in transforming key KPIs such as retention, satisfaction, and costs. As part of Latinia, she is committed to turning real-time analytics into meaningful connections that enhance customer experience and banking outcomes.

Want to learn more about how we can help your institution navigate these changes? Please schedule a consultation with us to discuss your specific needs and how our solutions can drive one of the challenges in banking digital transformation.