Banks do not struggle with AI itself. They struggle with scaling AI on top of fragmented data and technology landscapes. While AI adoption continues to accelerate across the banking sector, many organizations still face challenges in making data consistently available, accessible and usable across the enterprise. As a result, successful pilots often remain isolated within individual teams or functions, limiting their ability to create value at scale. Across the Dutch banking sector, the focus is now shifting from experimentation towards production. However, legacy technology, fragmented data environments and limited standardization continue to slow the rollout of AI beyond individual use cases. Banks are therefore increasingly investing in modern data platforms and AI-ready foundations to support enterprise-wide adoption.
From AI experimentation to AI at scale
State of AI in Banking
Building AI-ready data foundations
Scaling AI requires more than access to data. Banks need data that is reliable, accessible and enriched with the context required for effective decision-making. As organizations move beyond traditional reporting and analytics, data quality and the expansion and optimization of context management capabilities become increasingly important for enabling AI at scale. AI applications and agents require more than access to data alone. They need the business context that explains how data relates to customers, products, processes and decisions across the organization.
For many banks, this challenge extends beyond structured data. Unstructured information, often spread across documents, emails and legacy systems, represents a significant source of untapped value. Making this information AI-ready requires capabilities such as classification, labeling and enrichment, enabling AI applications to operate with greater context, reliability and business relevance. At the same time, AI is not only dependent on high-quality data, but can also help improve it through classification, enrichment, quality monitoring and the identification of inconsistencies. This creates a dynamic relationship in which data enables AI, and AI improves the usability and reliability of data across the organization.
Modernizing technology for AI adoption
Many banks continue to operate on technology landscapes that were not designed for AI-driven decision-making. Modernizing infrastructure and making both structured and unstructured data available for business use cases are critical steps in becoming AI-ready.
Across the banking industry, these investments are increasingly linked to AI initiatives in areas such as KYC, AML, customer service, software development and operational processes. While these use cases can already deliver tangible value, many organizations discover that scaling them requires a stronger underlying technology foundation than initially anticipated.
This challenge becomes more pronounced with the rise of AI agents, which operate across multiple systems, datasets and workflows. Unlike traditional models, agents continuously interact with data and each other, influencing decision-making in real time. Preparing for agentic AI therefore requires more than managing individual models.
Organizations need visibility across data, agents and decision flows, enabling decisions to remain explainable, auditable and aligned with organizational policies as AI capabilities become increasingly interconnected. The shift from isolated models to interconnected AI systems fundamentally changes how risk, accountability and performance are managed.
The evolution of the CDO
Digital sovereignty in European banking ecosystems
As financial institutions increasingly depend on global cloud providers, AI supply chains and cross-border data infrastructures, digital sovereignty has become a strategic concern rather than a purely technical one. Banks need to understand not only where data is stored, but also under which jurisdictions it falls and how external dependencies influence their ability to operate independently. This introduces risks related to regulatory exposure, concentration and external control, particularly in the European context where supervisory expectations and geopolitical factors play an increasing role.
Ensuring digital sovereignty therefore goes beyond transparency. It requires organizations to actively design their Data and AI platforms in a way that preserves control over data, models and execution, while embedding compliance, resilience and scalability into how AI is applied. In this way, innovation can be developed within clear regulatory boundaries, maintaining operational stability while reducing dependency on external ecosystems. Without this, increasing reliance on external providers can limit a bank’s ability to meet regulatory expectations, sustain resilient operations and maintain strategic autonomy.
From pilots to enterprise-scale adoption
Banks across the Netherlands are increasingly shifting from AI experimentation towards organization-wide production and scale. Market insights shows that three enablers consistently emerge as critical for a successful firm-wide rollout: data, people and technology. Banks need reliable and contextualized data, a workforce that understands how to adopt AI effectively, and modern technology platforms that support secure, governed and scalable deployment of AI capabilities. Many organizations recognize that scaling AI requires more than introducing additional use cases. The challenge lies in creating the foundations that allow successful pilots to be adopted across the enterprise.
Within Intelligent Banking, the transition from pilots to enterprise-scale adoption is enabled by three key capabilities:
Many organizations attempt to scale individual AI use cases, while the real challenge lies in orchestrating the underlying technology, data, people, governance and operating capabilities that support them. By combining these elements, banks can move beyond isolated pilots and embed AI into core processes such as onboarding, transaction monitoring and lending, enabling AI to operate as a consistent enterprise capability rather than a collection of independent initiatives. Banks that make this transition are better positioned to operationalize AI at scale, accelerate value creation and build towards an AI-first operating model.
AI Jumpstart Workshop
From AI pilot to scalable adoption
Looking to get started with AI, or struggling to scale beyond initial pilots? KPMG’s AI Jumpstart Workshop helps organizations in the financial services sector gain practical control over their AI ambitions in a short period of time.
Together with our experts, you will explore relevant use cases, assess opportunities and risks, and work towards a concrete action plan for responsible and scalable AI adoption. The workshop is complimentary and takes place at the Insights Center.
Featured
Contact
