4 insights into AI, compliance, and modern infrastructure in financial services
Financial services organizations today are navigating a perfect storm of regulatory pressure, rapid technological disruption, and the rising expectations of both customers and regulators. Against this backdrop, the adoption of AI and the modernization of digital infrastructure have become more than innovation goals; they are survival imperatives.
In a recent Software Plaza webinar, Transformational CTO Sarah Poland highlights the real-world complexities financial institutions face as they modernize amid AI advancements.
Drivers of transformation in financial services
Every major transformation starts with a catalyst. In financial services, that catalyst is increasingly driven by regulation and the need to handle tighter compliance rules, like the EU’s Digital Operational Resilience Act (DORA). These regulations set clear deadlines, explicit expectations, and impose very strict and costly penalties.
Unlike older regulatory systems, modern compliance laws set fines based on global revenue, quickly increasing pressure on multinational banks and fintech companies. As Sarah pointed out, this has become a major factor pushing organizations to modernize, not gradually, but urgently.
Another factor is the ongoing evolution of core infrastructure. Whether through cloud migrations, Kubernetes adoption, end-of-life technologies, or recent industry shakeups like the Broadcom–VMware acquisition, organizations often find themselves needing to rethink architectures sooner than expected. Even when teams hope for simple lift-and-shift approaches, reality rarely cooperates. Legacy systems remain highly stateful; containerized environments require a fundamentally different design approach.
All of this is compounded by the expectation that modernization should lower costs, enhance performance, and not disrupt business continuity, a balancing act leaders face daily.
Modernization is not only technical, it’s deeply cultural
Modernization efforts usually fail or succeed based on culture and leadership rather than on technology alone. Modern technology leaders cannot realistically be experts in every tool, platform, and pattern within their scope. Instead, their value increasingly lies in their ability to ask sharp, revealing questions, connect the dots between teams, and create environments where people can surface risks and propose ideas without fear.
This requires a broader view of stakeholders, not just the immediate engineering team, but also sales, business units, compliance, operations, and external partners. It also demands a willingness to lean into difficult conversations rather than avoid them.
When resistance arises, whether from a team, stakeholder, or function, effective leaders pause to identify the source of friction instead of pushing forward blindly. Often, resistance indicates missing context, legitimate constraints, or real misalignment that could cause costly mistakes later. Sarah compared this to parenting: when a child refuses something, simply insisting rarely works, but asking why often reveals a logical reason.
In the same way, treating people with curiosity and respect allows organizations to harness diverse perspectives. When there is trust, psychological safety, and genuine openness to feedback, modernization becomes a shared mission rather than a top-down directive.
The reality of AI in financial services
AI is one of the most discussed trends in financial services, but most organizations are still in the early stages of testing. Nearly 95% of AI pilots fail, not because AI is without value, but because companies lack the infrastructure, guardrails, and architectural maturity needed for safe deployment.
Many teams try to implement AI without the compute capacity, observability, data governance, or secure APIs required. Others expose LLMs without contextual boundaries, creating risks like prompt injection, data leakage, and soaring costs. Early AI projects also tend to be monolithic, making them hard to scale or secure, while developers often lack the time or safe environments needed to experiment effectively. These failures, however, are valuable. Organizations that learn quickly will ultimately lead in AI adoption.
Focus on first principles before investing in AI
Many organizations rushed into AI without strengthening foundational engineering practices. While AI is powerful, it presents major challenges, including high compute costs, new attack vectors, strict data privacy requirements, and the need for modular architectures. To address this, teams must focus on fundamentals.
Modern systems should be modular and decoupled, since monolithic AI stacks are fragile and expensive. User input should pass through a “static frontend” that sanitizes data, enforces access controls, filters unsafe instructions, and reduces unnecessary token usage. As organizations explore agentic AI, they must implement strict governance to manage the expanded attack surface.
Finally, a strong developer experience is crucial, allowing teams to learn, experiment, and build safely at scale.
What Leaders Should Expect in the Coming Year
Looking ahead, financial services will face tighter regulations, increased consolidation in AI tooling, and a stronger emphasis on sustainable AI architectures. Regulations like DORA will heighten scrutiny around operational resilience, third-party risk, and technology governance, prompting institutions to develop better-documented, audit-ready infrastructure and AI practices. As the AI tool landscape matures, organizations will move toward interoperable, secure, and scalable stacks.
Architecturally, AI will increasingly be seen as one component within a larger system, integrated through defined interfaces and protected by strict controls. Internally, engineering cultures will shift to prioritize focus, learning, and thoughtful experimentation. Ultimately, the winners will invest in strong foundations and deliberate decision-making rather than chasing every trend.
Conclusion
The financial services sector sits at a pivotal moment. Modernization is no longer optional; AI is no longer experimental; and compliance is no longer forgiving. Yet the organizations that approach these challenges with curiosity, strong leadership, architectural discipline, and a willingness to learn will be the ones that unlock the true competitive advantages of AI and modern infrastructure. The journey is complex, but the opportunity is extraordinary.
This blog is based on the webinar Scaling Smart: Sarah Polan on AI, Compliance & Modern Infrastructure in Financial Services with Transformational CTO Sarah Poland. You can watch the video here.
