Banks overhaul core systems for AI-driven services

By early 2026, it became clear that banks' longstanding focus on enhancing digital front ends while neglecting their legacy back-end core systems was unsustainable in the face of rising demands for real-time services and automation. The Temenos report underscored that without modernizing core banking infrastructures to incorporate cleaner data structures, stronger governance, and flexible platforms, banks risked losing competitiveness as these technology choices have long-term impacts on innovation and the ability to monetize AI-driven digital services.

Legacy core banking systems, predominantly monolithic platforms originally built on mainframes, have become a fundamental bottleneck for banks aiming to engage in real-time digital commerce and fintech innovation. These systems, designed for batch processing and operational stability rather than agility, now consume up to 70% of IT budgets just for maintenance, exacerbated by a shrinking pool of COBOL and mainframe experts. Middleware solutions have only temporarily masked these limitations, leaving banks with siloed data, fragmented customer experiences, and operational inefficiencies that hinder seamless integration and automation.

The market’s accelerating expectations for instant gratification—real-time information, payments, and AI-powered capabilities—have exposed the inadequacies of legacy monolithic cores, which struggle with scalability, cost efficiency, and agility. As Prashant Shah noted, customers now demand seamless digital experiences that unify payments, lending, fraud controls, and servicing into a single environment, yet fragmented legacy systems with separate data repositories impede banks from delivering these unified journeys. This has driven a shift toward organizing technology around customer journeys rather than isolated product silos to meet real-time digital demands effectively.

In response to these challenges, a new architectural paradigm is emerging with the Bank Operating System, which decouples the system of record (the legacy core) from the system of growth. This composable, real-time growth layer enables banks to launch revenue-generating products rapidly and operate independently of core downtime, addressing the inflexibility and operational risks inherent in monolithic legacy systems. This evolution represents a critical step in overcoming the compounded complexity banks have historically avoided due to cost, security, and scalability concerns.

By mid-2026, core banking modernization had evolved into a strategic imperative at the highest executive levels, transcending its traditional IT confines. Leonardo Collado of Pismo highlighted that modernization is now about delivering unique, AI-driven customer experiences rather than mere technical upgrades. Yet, this transformation is constrained by legacy systems consuming roughly 70% of IT budgets, forcing banks to adopt phased modernization strategies that balance innovation with operational continuity to avoid overwhelming complexity and risk.

Strategic approaches to modernization increasingly favor building comprehensive, end-to-end platforms rather than piecemeal upgrades. As one interviewee noted, truly transforming a bank requires addressing all components simultaneously to fundamentally change operations, rather than tinkering with isolated parts. This approach simplifies innovation and product delivery by reducing the proliferation of interdependent projects, as exemplified by Constantinople’s effort to consolidate 50 projects into a unified banking and technology capability.

Modernization must prioritize strengthening control and operational dependability over simply adding new features. Jayanth M. Prasanna emphasized that these initiatives are better framed as workflow migrations rather than technology swaps, given the complex embedded business logic and regulatory requirements banks face. The true test of modernization lies in maintaining data accuracy, exception handling, and traceability throughout transaction lifecycles, especially in international payments where downstream failures often occur despite technical connectivity.

Phased, incremental modernization remains the safer and more resilient path, particularly in complex domains like WealthTech. Tieto Banktech advocates starting modernization efforts where the greatest benefit can be realized, rather than simply targeting the oldest technology first, and embedding governance into every API and workflow to ensure trusted data flow and operational continuity. Selecting a modernization platform is thus as much a partnership decision as a technological one, requiring providers who understand the full retail and institutional lifecycle and can sequence migration around operational readiness.

Valley National Bank exemplifies a disciplined, infrastructure-first modernization approach under COO Russell Barrett, who since 2021 has driven a cloud-first migration of over 80% of the bank’s data center capacity and established a centralized enterprise data hub. This robust foundation not only yielded nearly $100 million in cost savings and $20 million in new recurring revenue but also enabled the deployment of AI applications like the AML agent 'Tara,' which improved fraud detection metrics significantly. CEO Ira Robbins highlights that such foundational data quality is crucial for responsible AI use, framing AI as a 'connectivity solution' that enhances customer engagement rather than a standalone buzzword.

Valley Bank’s modernization extends beyond internal upgrades through its innovation ecosystem—Valley Foundry and Valley Ventures—that integrates external fintech partnerships, recognizing that continuous innovation requires collaborative technology embedding rather than isolated efforts. This ecosystem approach complements their core infrastructure investments, positioning the bank to compete effectively against both large incumbents and agile startups.

In the GCC region, banks are transitioning from initial digital transformations that enhanced customer access to a more complex phase focused on simplifying and controlling the intricate web of connections between core banking systems and numerous external providers. This evolution is driven by region-specific open finance initiatives, such as Saudi Arabia’s Open Banking Framework launched in early 2026 and the UAE’s broader open finance model with API hubs and trust frameworks, reflecting tailored regulatory and market contexts that push banks toward modular, integrated platforms.

To manage the growing complexity of multiple external service dependencies—including card processing, identity verification, and fraud control—GCC banks are adopting integration as a strategic, ongoing capability rather than a one-off project. Traditional point-to-point integrations have proven inadequate, prompting the use of middleware and orchestration technologies like Velmie's API middleware to stabilize core systems while enabling agile service delivery and ensuring reliability in a distributed technology environment.

Deutsche Bank’s Private Bank is executing a landmark decade-long, €600 million modernization initiative to consolidate 15 fragmented legacy core banking systems into two cloud-native platforms, with Thought Machine’s Vault Core selected as the foundational engine. This strategic overhaul aims to simplify the bank’s complex technology landscape, accelerate time-to-market for new products, and establish a scalable infrastructure to support future growth in Personal Banking and Wealth Management, as emphasized by executives like Yiping Li and Christian Rhino.

The bank’s phased migration approach, commencing in 2027, involves running legacy systems in parallel with the new Vault Core platform to manage operational risks and ensure service continuity. This deliberate strategy, shaped by hard lessons from the problematic Postbank integration that triggered regulatory scrutiny and internal governance reforms, reflects Deutsche Bank’s commitment to mitigating risk while progressively transitioning products to the modernized core.

Deutsche Bank’s partnership with Thought Machine and GFT Technologies exemplifies a collaborative effort to build a resilient, scalable, and innovation-friendly core banking platform. Thought Machine co-founder Paul Taylor highlights this decade-long commitment as a pioneering model for international banks aiming to modernize infrastructure, while GFT’s involvement as system integrator underscores the operational efficiency and future growth potential embedded in this transformation.

Beyond technological renewal, this modernization addresses critical operational failures experienced during the Postbank acquisition, which resulted in nearly 10,000 customer complaints and regulatory intervention by BaFin. By replacing the fragmented legacy systems that contributed to these issues, Deutsche Bank seeks not only to enhance product development agility by decoupling product logic from infrastructure but also to restore client trust and position itself as a leader in core banking innovation.

Modernization in banking transcends feature enhancement by fundamentally reinforcing control, dependability, and operational integrity under real-world conditions. Jayanth M. Prasanna highlights that true modernization demands preserving embedded business logic and workflow continuity, especially in complex areas like international payments where data quality, exception handling, and traceability throughout the transaction lifecycle are critical to avoid downstream failures. This approach ensures that modernization is not merely a technical upgrade but a rigorous migration of operational workflows that sustain trust and compliance.

As banks transition from episodic modernization projects to a perpetual state of evolution, new operational models become essential to safely enable continuous change. This shift requires clear shared priorities, defined decision rights, disciplined funding, and cross-functional accountability to prevent destabilization of core systems. Michael Araneta of AWS underscores that the ultimate litmus test for banking infrastructure now lies in its capacity to support rapid innovation and AI-driven product introductions, moving beyond traditional metrics like cloud migration percentages to focus on lead time, failure rates, and customer continuity.

Architecting for continuous deployment demands a paradigm shift toward resilience, scalability, and rapid iteration rather than mere stability and processing power. Chetan Sharma from AWS emphasizes that cloud architectures must be designed to accommodate ongoing change seamlessly. Concurrently, embedded risk, compliance, and operational controls must be integrated upstream to keep pace with compressed decision cycles, ensuring swift yet controlled innovation without overwhelming organizational resources.

During the complex migration phases where legacy and modern core banking systems coexist, deliberate data governance and coexistence strategies become paramount. Banks must architect around routing, orchestration, reconciliation, latency, and recovery to manage accounts, products, and data flowing between dual cores. This dual-operating environment elevates concerns such as customer communication and operational recovery to first-class design priorities, ensuring seamless service continuity and regulatory compliance amid ongoing transformation.