In December, The Economist flagged ‘agentic AI’ as one of the most highly anticipated breakthroughs of 2025. For many retail banks, this represents just one of the many opportunities to turn GenAI from a promising proof-of-concept into a scalable, secure production tool. But the complexity of GenAI deployments requires a strategic approach, reliable infrastructure, and expert guidance.
Every day, new business possibilities appear around GenAI. Driven by the speed of its development and a dizzying array of new use cases, organisations are constantly pushing the boundaries of what’s possible. In retail banking, GenAI breakthroughs promise to solve challenges with compliance, personalisation, efficiency, remediation, and legacy system modernisation – but that takes careful planning.
Without the right foundations, projects can stall, held back by infrastructure constraints, regulatory and compliance challenges, or strategic misalignment. Building these foundations means partnering with those who can provide both the retail banking domain knowledge and the expertise and tools to accelerate success.
Moving Beyond Experimentation to ROI
Going by the headline of a recent IDC blog - “Experimenting Forever Isn’t an Option” – it’s clear why a majority of retail banks have already invested heavily in GenAI. There are a number of revenue-generating or cost-cutting measures that could move the dial, including:
- Data remediation: automating Anti Money Laundering (AML) checks with techniques like fuzzy matching to eliminate backlogs, reducing manual intervention.
- Hyper-personalisation: analysing transaction patterns to deliver tailored product packages, boosting engagement and revenue.
- Consumer Duty compliance: detecting financial understanding gaps and generating communications that align with regulatory standards.
- Code compliance: continuously monitoring for developments in regulatory adherence and flagging issues early.
- Modernising legacy code: reinterpreting and enhancing mainframe code for performance and scalability on newer platforms.
However, delivering ROI on these use cases has proven more difficult than expected.
According to research by NTT DATA, 83% of businesses report that their GenAI strategies aren’t aligned with business plans. Meanwhile, 90% of organisations said that outdated infrastructure is hindering GenAI adoption.
In the rush to meet shareholder expectations, it can be difficult to find time for the essentials. This combination of legacy infrastructure and strategic misalignment can stop a project cold before it has a chance to begin. That’s why, to move forward, banks must prioritise strategic use cases and infrastructure that supports their goals.
Legacy Infrastructure is Holding You Back
Stakeholders for GenAI projects at retail banks have a mountain to climb to overcome internal governance concerns, model selection errors, budget constraints, complex training, and legacy code. Typically, as these projects scale, the issues only increase.
As Bent Flyvbjerg, Emeritus Professor at the University of Oxford, wrote in his critique of ‘megaprojects’: “large scale ventures… almost always cost far more than ever anticipated and as often under-deliver by a number of key metrics” and especially “revenues”. In retail banking, this is largely due to legacy infrastructure that wasn’t built to support enterprise AI initiatives.
On the other hand, cloud-based solutions, such as those offered by Google Cloud, provide the scalability, efficiency, and security necessary to support GenAI at scale. Google Cloud offers a wide range of tools to accelerate and simplify GenAI adoption, including Vertex AI, a platform that helps you build, deploy, and manage your ML models and applications at scale; Vertex AI Studio, which provides a simple interface for prototyping and experimenting with generative AI models; and BigQuery, a fully managed, serverless data warehouse that enables scalable analysis over petabytes of data.
Its global infrastructure ensures high reliability and low latency, while its integrated data analytics tools, like Looker and BigQuery, help organisations derive actionable insights from vast datasets.
To accelerate GenAI deployment, banks need to prioritise:
- Moving to cloud environments that support high-performance GenAI workloads.
- Ensuring data is clean, structured, and accessible.
- Managing cloud spend with expert frameworks (such as NTT DATA’s full-stack FinOps).
NTT DATA complements these capabilities with extensive experience in cloud transformation and financial optimisation. Our full-stack FinOps framework helps manage cloud costs effectively – reducing expenses by up to 50% – while ensuring optimal resource allocation. Additionally, our expertise in infrastructure modernisation ensures legacy systems can integrate smoothly with cloud environments, accelerating deployment timelines.
Strategic Partnerships Make the Difference
Having the right cloud infrastructure is one part of the puzzle. To put the picture together, you need partners who understand how to align these capabilities with your business strategy, mitigate risks, and accelerate deployment.
That’s why partnering with NTT DATA and Google Cloud offers a complete solution. NTT DATA’s strategic expertise ensures your GenAI projects are tailored to your business needs, while Google Cloud’s infrastructure and enterprise-ready tools, like Vertex AI Studio, provide the materials required to bring those projects to life. Together, we take on challenges like compliance, legacy modernisation, and cost efficiency to ensure you have a seamless end-to-end GenAI journey.
Together, Google and NTT DATA offer:
- Access to over 130 foundational models through Google’s Model Garden.
- Everything from infrastructure modernisation to compliance guardrails, ensuring AI solutions are secure and responsible.
- Tools like Dolffia for document processing and Unikix for mainframe modernisation to speed up AI-readiness.
Building the Right Foundations for ROI
To ensure GenAI projects deliver ROI, retail banks must focus on these key pillars:
- Transitioning to cloud-based platforms for scalability and performance.
- Choosing the right foundational models to avoid costly delays and retraining efforts.
- Leveraging financial management frameworks to optimise cloud spending.
- Collaborating with trusted partners like Google Cloud and NTT DATA to navigate challenges and accelerate deployment.
Beginning Your Journey
It’s difficult to downplay the impact that GenAI can have on retail banking. Without the right foundations, though, deliering returns on the investment can be a slow, drawn-out process.
With Google Cloud’s capabilities and NTT DATA’s expertise, banks have the tools they need to bring GenAI to life delivering real value, faster. Now is the time to lay the groundwork and accelerate your GenAI journey.
Ready to take GenAI from proof-of-concept to production? Schedule a 45-minute consultation to see how NTT DATA and Google Cloud can help you lay the right foundations and accelerate success.
