The financial services industry has been one of the fastest adopters of generative and agentic AI, yet many financial institutions limit their adoption to providing access to the models / LLMs, pushing for pilot projects and proofs of concept along the value chain, or implementing small-scale uses, deferring the broader transformation. According to data from Evident’s AI Banking Index for Q1 2026, a mere 9 out of the 50 financial institutions monitored have successfully deployed an agentic AI application, whether in live production or a pilot phase.
The main issues banks face for a deeper transformation are lack of a semantic data layer, complex internal and external integrations, and legacy core systems built for slower, paper-based work that cannot support the real-time, automated processes needed for a more transformative wave of AI. Then, the real question is, how can financial institutions modernize effectively?
This article explores four areas that banks must work on to modernize effectively and build the foundations for a deeper transformation that not only embeds AI across the value chain, but truly transforms the ways banks operate in the AI-native era.
1. Legacy & Technical Debt Modernization
During the digital era, people talked about modernization mainly to cut costs and boost efficiency. But legacy systems determine the speed at which organizations can launch new products, reach new markets, work with partners, and use AI at scale.
Across the banking sector, decades-old core systems, undocumented codebases, and accumulated technical debt continue to limit innovation and operational agility. But the challenge runs deeper than slow systems or outdated interfaces. Decades of undocumented development have left critical business logic, pricing rules, risk thresholds, and compliance checks buried inside fragile code that only a shrinking group of specialists still understands. When information and knowledge lives in a mainframe rather than in accessible, governed services, innovation stalls. Teams end up making quick fixes, custom links, and small experiments that are hard to scale across the whole bank.
Meanwhile, many existing platforms remain optimized for standardization and control rather than adaptability. The future requires something different: modular business capabilities, governed data products, and orchestration layers that allow banks to continuously evolve customer journeys and business processes without constantly impacting the core. Modernization should not only be considered as transitioning the old to newer versions, but building a platform that supports AI as well.
The challenge for banks then is not only about building the proper foundations for AI, but on using AI itself as a powerful accelerator for modernization. The technology is helping institutions analyze legacy environments, document complex dependencies, refactor aging applications, and reduce the risks traditionally associated with large-scale transformation. Banks that don’t modernize at speed risk being left out in this AI-native era.
The Bottleneck: Why Banking Modernization Is Unique
Banks face structural challenges like those of most highly regulated industries, but on a much larger scale because of high transaction volumes, real-time settlement needs, and years of combining old systems.
- M&A Architectural Debt: Years of bank mergers have left many institutions running several core systems, ledgers, and customer databases at the same time, all connected by weak middleware.
- The Mainframe Dependency: Important ledger functions often run on mainframes with old code. This means banks depend on highly specialized knowledge that is increasingly scarce, and have trouble accessing real-time data.
- The Batch Processing Lag: Core banking systems from the 1970s and 80s use overnight batch jobs to settle transactions and update balances. Without modernization, real-time, event-driven decisions are not possible.
Legacy systems do more than slow down IT. They hide important business rules in fragile code. When pricing, risk, and compliance checks are stuck in mainframes, AI agents cannot use them safely.
Modernization is an imperative in business transformation
Modernization should focus on future business needs, not just today’s technical limits, and the outcomes that matter are measurable business results that impact:
- Faster time-to-market for new products and channels. Multimodal channels with real time needs will dominate
- Better customer experiences. As AI might enable a “sea of sameness”, using it to differentiate will be key to new customer acquisition and retention
- Improved straight-through processing. Not only embedding AI to solve bottlenecks, but reimagining end-to-end processes with AI + Human as the key principle
- Greater AI readiness through unified data and exposed business capabilities. Overall transformation of the Middle and Back Office, while balancing regulation and risk management with innovation
When modernization is anchored to business value, its benefits compound over time. Rather than pursuing a disruptive “big bang” transformation, leading banks take a careful, value-driven approach that delivers measurable outcomes at every stage. By progressing through defined milestones, banks can reduce risk, maintain business continuity, and generate tangible value throughout the transformation journey while building momentum toward their long-term strategic goals.
2. Customer Experience & Financial Wellbeing
As customer expectations continue to rise, banks are moving beyond transactional interactions toward more personalized, proactive experiences. AI is enabling institutions to anticipate needs, personalize financial guidance, and help customers build healthier financial habits through contextual, real-time support. The shift is from delivering merely functional services to creating seamless, engaging experiences that strengthen loyalty and enable traditional banks to compete more effectively with digital-first (and soon AI-first) challengers such as Revolut and other fintechs.
During the digital era, most banks invested in improving their channels, but their core systems remain constrained by legacy technology and direct system connections. But the competitive advantage is no longer simply offering digital services that work. Increasingly, it is about delivering personalized customer experiences that feel intuitive, proactive, and contextually relevant. Much like premier hospitality leaders that transformed entertainment by orchestrating seamless customer journeys, leading financial institutions are using AI to create experiences that anticipate needs, simplify decisions, and actively support customers’ financial well-being.
In this environment, modernization is not just about operational efficiency or a new coat of paint on customer-facing channels. It is the foundation for a new generation of AI-driven customer experiences that strengthen loyalty, improve financial outcomes, and make customers’ lives materially easier. And that bar is higher than it sounds: money looks like a rational domain and behaves like an emotional one. Systems that only understand transactions will keep producing experiences that are technically correct, with the moments that matter buried in undifferentiated routine.
Neobanks like Revolut have already demonstrated that customers will migrate to platforms that help them manage their money more effectively, automating savings, categorizing spending, surfacing personalized recommendations, and providing real-time financial insights. Traditional banks risk losing not just customers, but the primary financial relationship altogether, as AI-powered assistants begin managing transactions and financial decisions on people’s behalf.
3. Change Management & Cultural Transformation
Technical architecture is rarely the reason modernization initiatives fail. More often, the blocker is cultural. Employees who have spent years building expertise in legacy systems, or who hold institutional power precisely because they are the only ones who understand them, might resist changes that would make that knowledge more accessible. AI, by its nature, democratizes expertise. For many inside a financial institution, that is not an opportunity. It is a threat.
This resistance rarely surfaces as outright opposition. It shows up as delayed decisions or lack of prioritization of modernization efforts, scope creep, preference for endless pilots over production rollouts, and a tendency to frame every governance requirement as a reason to slow down. The result is organizations that are technically capable of modernizing but organizationally unable to do so.
Closing this gap requires treating change management as a first-order deliverable, not an afterthought. That means redefining roles early, creating visible wins that demonstrate AI augments rather than replaces, and building leadership alignment before technical work begins. The organizations that succeed at modernization are not necessarily the ones with the best architecture. They are the ones who move their people forward alongside their systems.
4. Governance & Standardization
As AI adoption accelerates, many financial institutions risk creating fragmented solutions, duplicated efforts, and inconsistent standards across teams. A combination of platform engineering, reusable foundations, and Agentic AI helps banks scale innovation without sacrificing control. The goal is not more bureaucracy, but smarter governance: clear guardrails, shared platforms, and standardized practices that enable business teams to move quickly while ensuring security, compliance, and long-term sustainability.
Why Agentic AI requires modern foundations
While predictive AI helped banks to operate better under uncertainty, and generative AI enabled a first wave of transformation in IT and Marketing, Agentic AI goes a step further by orchestrating decisions and actions across systems, workflows, and business processes.
To operate effectively, AI agents require four foundational capabilities: First, a semantic layer of data that provides a trusted view across multiple data domains; second, real-time integration that supports continuous, event-driven decision-making; third, automation and orchestration of processes that coordinate actions across systems, business rules, APIs, and people; and finally, governance by design to ensure every decision operates within defined controls, audit requirements, and escalation paths.
These capabilities are difficult to achieve in legacy environments but are inherent to modern, API-first, cloud-native architectures. This is why API-first has become a
strategic priority across the financial services industry, although APIs alone are not enough. A truly modern setup needs a modular core, managed data products, and event-driven orchestration. Without these basics, API-first is just a cover for old complexity, not a real solution for Agentic AI.
Accelerating transformation with AI Pods
The risks of ungoverned AI adoption are not hypothetical. When organizations allow teams to independently choose their own tools, models, and workflows without a shared standard, the result is fragmentation, inconsistent outputs, incompatible data, higher costs, and processes that cannot be audited or replicated. The impacts show up later, when leadership asks for results or tries to scale a pilot, and there is no direct answer.
To help organizations accelerate modernization while avoiding this trap, Globant leverages AI Pods: Expert-supervised, enterprise-grade, high-quality AI Pods to solve the problem that generic AI implementations have – getting to production at enterprise standards.
By embedding AI directly into the delivery process, banks can accelerate architectural redesign, modernize obsolete code, and establish the foundational capabilities required for future automation, including unified data, governed APIs, and modular architectures.
A practical starting point is mapping which core business processes – onboarding, credit decisioning, compliance checks – still depend on logic buried in undocumented legacy code. Those are the highest-leverage areas for modernization, and the natural entry point for an AI Pods engagement.
