How AI is reshaping the software behind everyday banking
By Martin Reeves, Global Head of Engineering Platforms & Practice
Imagine it’s Friday evening. You’re heading out to meet friends for dinner and check your phone on the way. Within seconds, your route is mapped, your taxi ride is booked, and payment is processed. By the time you look up, you’re already on your way.
What most people don’t see is the complexity behind that moment. To ensure that you can book a journey without as much as a second thought, software developers have spent hours writing thousands of lines of code, testing and perfecting it. The software must work flawlessly day and night for thousands of customers, across, different locations and networks. Even a minor mistake could affect you when it matters most.
Why maintaining banking software is hard
When we talk about engineering in a large, complex organisation like ours, success is rarely defined by a single technology choice or delivery milestone. More often, it is about carefully improving what already exists, without introducing risk. What really matters is how decisions are made - how teams balance speed and quality, and how accountability is taken to ensure business outcomes last well beyond the initial delivery.
This is where we see AI beginning to make a measurable difference.
How AI is supporting engineers and improving client outcomes
In recent years, software developers have begun using AI-powered coding assistants to help write code, identify mistakes, generate test cases or interpret unfamiliar code written in legacy languages. These coding assistants don’t replace human judgment but reduce the burden of repetitive work, allowing engineers to focus on higher-value tasks.
At Deutsche Bank, our software developers have access to coding assistants such as GitHub Copilot and Google’s Gemini Code Assist. We are already seeing tangible benefits. Over the past 12 months developers have progressed from reporting time savings of 1.5 to 2.5 hours per week to experiencing material 10x gains on a number of specific tasks. At scale these gains translate into faster development cycles and improved productivity.
These AI tools also play an important role in accelerating learning. By helping developers understand complex and unfamiliar code, they accelerate knowledge transfer across teams.
Freed from repetitive manual tasks, developers can focus on improving customer-facing features and resolving issues faster. That translates into quicker updates, more stable applications and fewer disruptions for clients.
Balancing speed with responsibility
In banking, innovation must always be balanced with control. At Deutsche Bank, we are deliberate in how we deploy AI. We define clear boundaries and never allow developers to deploy AI-generated code without review. Developers remain accountable for every line of code that is implemented. To begin with the AI-generated suggestions were only selectively integrated into code. More recently however, the more powerful large language models enabled through these tools have demonstrated themselves extremely capable of solving holistic design and coding challenges, as well as materially helping in the code review process. This highlights both the value of AI and the ongoing importance of human oversight.
Moving from assistance to transformation
We are still at an early stage of this journey. The next step is to embed AI coding tools more deeply into our development standards and focus on quality engagement rather than widespread, shallow usage. At the same time, we are preparing for advanced agentic AI capabilities, that can support complex development tasks and help enable real innovation, always within robust guardrails.
What matters most is not just the technology itself, but how we use it. The greatest benefits are already emerging among developers who actively engage with these tools and integrate them into their daily workflows.
In conclusion, AI isn’t replacing engineers. It is amplifying their impact. By combining human judgment with AI-driven efficiency, we can build systems that are faster, more resilient, and more reliable. For clients, that means banking that works, securely, seamlessly, every time.
Home
By Martin Reeves, Global Head of Engineering Platforms & Practice
Imagine it’s Friday evening. You’re heading out to meet friends for dinner and check your phone on the way. Within seconds, your route is mapped, your taxi ride is booked, and payment is processed. By the time you look up, you’re already on your way.
What most people don’t see is the complexity behind that moment. To ensure that you can book a journey without as much as a second thought, software developers have spent hours writing thousands of lines of code, testing and perfecting it. The software must work flawlessly day and night for thousands of customers, across, different locations and networks. Even a minor mistake could affect you when it matters most.
Why maintaining banking software is hard
When we talk about engineering in a large, complex organisation like ours, success is rarely defined by a single technology choice or delivery milestone. More often, it is about carefully improving what already exists, without introducing risk. What really matters is how decisions are made - how teams balance speed and quality, and how accountability is taken to ensure business outcomes last well beyond the initial delivery.
This is where we see AI beginning to make a measurable difference.
How AI is supporting engineers and improving client outcomes
In recent years, software developers have begun using AI-powered coding assistants to help write code, identify mistakes, generate test cases or interpret unfamiliar code written in legacy languages. These coding assistants don’t replace human judgment but reduce the burden of repetitive work, allowing engineers to focus on higher-value tasks.
At Deutsche Bank, our software developers have access to coding assistants such as GitHub Copilot and Google’s Gemini Code Assist. We are already seeing tangible benefits. Over the past 12 months developers have progressed from reporting time savings of 1.5 to 2.5 hours per week to experiencing material 10x gains on a number of specific tasks. At scale these gains translate into faster development cycles and improved productivity.
These AI tools also play an important role in accelerating learning. By helping developers understand complex and unfamiliar code, they accelerate knowledge transfer across teams.
Freed from repetitive manual tasks, developers can focus on improving customer-facing features and resolving issues faster. That translates into quicker updates, more stable applications and fewer disruptions for clients.
Balancing speed with responsibility
In banking, innovation must always be balanced with control. At Deutsche Bank, we are deliberate in how we deploy AI. We define clear boundaries and never allow developers to deploy AI-generated code without review. Developers remain accountable for every line of code that is implemented. To begin with the AI-generated suggestions were only selectively integrated into code. More recently however, the more powerful large language models enabled through these tools have demonstrated themselves extremely capable of solving holistic design and coding challenges, as well as materially helping in the code review process. This highlights both the value of AI and the ongoing importance of human oversight.
Moving from assistance to transformation
We are still at an early stage of this journey. The next step is to embed AI coding tools more deeply into our development standards and focus on quality engagement rather than widespread, shallow usage. At the same time, we are preparing for advanced agentic AI capabilities, that can support complex development tasks and help enable real innovation, always within robust guardrails.
What matters most is not just the technology itself, but how we use it. The greatest benefits are already emerging among developers who actively engage with these tools and integrate them into their daily workflows.
In conclusion, AI isn’t replacing engineers. It is amplifying their impact. By combining human judgment with AI-driven efficiency, we can build systems that are faster, more resilient, and more reliable. For clients, that means banking that works, securely, seamlessly, every time.
How helpful was this article?
Click on the stars to send a rating