The pressure to modernise legacy software is a global challenge, with ageing codebases hindering agility, driving up maintenance costs, and exposing businesses to security risks. In response, Artificial Intelligence (AI) has emerged as the most disruptive force in software engineering, moving from a coding assistant to a full-fledged modernisation partner.
Globally, this transformation is delivering measurable economic value. Analysts predict that AI could add $13 trillion to global economic activity by 2030 and is already driving substantial efficiency gains:
- Productivity Surge: Employees who actively use AI tools report an average productivity improvement of 66% on daily tasks, which is akin to 47 years of natural productivity growth condensed.
- Widespread Adoption: Approximately 78% of global companies report using AI in their business, with over 92% planning to increase their investment in the technology over the next three years.
- Time Savings: Data suggests AI can save an employee 2.5 hours per day on average by automating routine work.
By embracing this global trend, we are positioning ourselves at the forefront of the technological shift, leveraging AI to tackle massive, decades-old codebases – a process that was once slow, manual, and prohibitively expensive.
The Global Blueprint: AI Accelerates Modernisation Worldwide
The AI Code Tools market is experiencing exponential growth, with companies across the financial, government, and technology sectors reporting massive productivity gains. The key applications driving this global trend include:
- Legacy Code Analysis and Translation: LLMs excel at parsing obsolete languages (like COBOL or legacy C++) and converting them into modern equivalents (like Java, Python, or .NET Core). For example, major financial institutions have used internal LLM tools to translate millions of lines of legacy code, saving hundreds of thousands of developer hours.
- Automated Refactoring and Debt Reduction: AI tools, such as GitHub Copilot and commercial platforms, can scan entire repositories to flag ‘code smells’, detect security vulnerabilities, and recommend modern design patterns—tasks that used to consume significant senior engineer time. McKinsey estimates that AI can lead to a 40–50% acceleration in tech modernisation timelines.
- Documentation and Testing: AI automatically generates comprehensive documentation from previously undocumented functions and scaffolds unit and regression tests for old code. This is crucial for knowledge transfer and ensuring the modernised code remains functionally correct.
This global shift is characterised by a “human-in-the-loop” approach, where AI handles the repetitive, large-scale automation (estimated to be 75% of the heavy lifting), allowing engineers to focus on architectural decisions and preserving critical business logic.
Our Strategy: Three Pillars of AI-Powered Modernisation
Following this global trend of adopting AI to drive efficiency, our company has strategically integrated AI across our entire product development lifecycle. By combining purpose-built internal solutions with best-in-class commercial tools, we have established three pillars for accelerating our product development and modernisation efforts.
1. Software Product Modernisation: LLM and RAG for Codebase Intelligence
At the heart of our modernisation effort is an in-house solution that uses open-source Large Language Models (LLMs) combined with Retrieval-Augmented Generation (RAG). This system is our foundational tool for modernising our code by:
- Contextual Refactoring: The RAG component allows the LLM to access and reference our entire proprietary, internal codebase and architecture documentation. This ensures that the code it refactors and the suggestions it makes are tailored and compliant with our specific standards, avoiding the simple translation of legacy debt.
- Automatic Documentation: It maintains live documentation that automatically updates as the code evolves. This preserves institutional knowledge and significantly reduces technical debt.
2. Accelerated Development with Coding Agents
To boost the daily productivity of our developers and speed up feature delivery, we have adopted powerful AI Coding Agents.
- Speed and Efficiency: Our primary agent, Claude Code, works directly with developers to handle complex coding tasks right from the command line. It assists in everything from scaffolding new features to debugging intricate, cross-file issues, significantly shortening the cycle from concept to working code.
- Human-Augmentation: The agents serve as powerful co-pilots, allowing our development team to be ten times more productive on routine tasks, freeing them to focus their human expertise on complex problem-solving and unique architectural design.
3. AI-Powered Design Iteration
The application of AI extends to the user experience, driving faster and more informed design choices.
- Rapid UI Generation: Our design team leverages Figma’s AI capabilities to quickly generate a multitude of initial user interface concepts and layouts.
- Focus on UX: By automating the initial mockup phase, our designers can accelerate the iteration cycle and spend more critical time refining user experiences, conducting testing, and applying the crucial human touch that creates intuitive and delightful products.
The Path Forward: Augmenting Human Innovation
By integrating these internal and commercial AI solutions, we are not just keeping pace with global trends; we are leveraging them to create a high-velocity development environment. The result is a faster time-to-market, a cleaner, more scalable codebase, and a highly engaged team empowered to dedicate their creativity and expertise to the complex challenges that truly require human insight and innovation.