Unlocking potential and empowering software engineers for unprecedented productivity using AI-driven code generation
by Louis Liu
ChatGPT has exploded into our collective conscience and is the subject that everybody’s talking about. Within the software engineering community most of the talk has centred on the immense potential of generative AI - a ground-breaking technology that promises to revolutionise the way we approach software development by boosting engineers’ productivity and improving the cost-effectiveness of their work.
In early 2023, ANZ conducted several preliminary proof-of-concept AI experiments. These experiments involved over 150 people: software engineers, SRE engineers, DevOps engineers, data engineers and platform architects. The results were unequivocal and very promising, so much so that we’re going to adopt selective practices from the experiments into appropriate pockets of the bank. I’d like to delve into the transformative power of AI-driven code generation and the profound impact it’s having on the field of software engineering at ANZ.
Personalisation: unleashing the potential of individual ANZ engineers
We found that one of the most remarkable advantages of integrating generative AI into software engineering is its ability to personalise code generation. Traditional code-writing methods often involve extensive manual efforts and many repetitive tasks. By using AI models to analyse an ANZ engineer's coding style, preferred frameworks, and specific use cases, we’ve been able to generate code tailored to each engineer’s unique requirements.
This personalised approach empowered our engineers to work more efficiently and effectively - as they received code that aligns precisely with their needs! The productivity enhancements were especially pronounced for junior to mid-level ANZ engineers.
Other upsides we’ll realise from the ability to leverage AI for code personalisation is that our engineers will be able to focus on higher-level tasks and spend more time on creativity. Clearly this reduction of mechanical time means ANZ will deliver higher-quality software for our customers.
Enhanced productivity: streamlining development with a unified platform
ANZ’s software engineers possess expansive mindsets and far-reaching ideas. To allow our team to reach their full potential, we strive to provide engineers with a simple, unified platform that serves all end-to-end needs of the development process. The integration of AI-powered tools into our software engineering workflow is the next step.
AI-powered tools will provide our engineers with an enhanced centralised platform where they can understand requirements, write code, generate testing cases, execute tests, and handle administrative processes seamlessly. AI-powered tools also will incorporate code standards and style checks, ensuring code quality and best practices are maintained or elevated.
Practical implementation: a Visual Studio Code extension for seamless integration
At the forefront of AI-driven code generation is an extension of ChatGPT we have developed specifically for Visual Studio Code. Integrating this brings the power of AI directly into our development environment and enables a seamless and efficient workflow. Our purpose-built extension offers a wide range of functionalities - most notably code standard checking and unit test creation and execution.
We’ve seamlessly integrated our Visual Studio Code extension with JIRA to automate the registration of test case execution results. This second integration reduced manual interactions, minimised errors, and promoted a more streamlined development process. By providing our engineers with a unified and AI-powered environment we can enable them to work more productively and to deliver high-quality software faster than ever before.
Accelerated development: quantifying the benefits of AI-generated code
Preliminary experimental results have demonstrated the remarkable speed improvements achieved through AI-generated code and JIRA integration. In comparison to traditional manual coding and JIRA interaction, our initial experiments have shown an incredible four-fold increase in development speed!
The massive increase in performance will allow engineers to deliver better software solutions in significantly shorter timeframes. It’s a game changer.
The time saved through AI can then be reinvested into refining ANZ’s software, addressing complex challenges, and focusing on strategic initiatives. This substantial improvement in efficiency will revolutionise the way software is developed and delivered in ANZ, setting new standards for productivity within the software engineering landscape.
Looking towards the future: collaborating with AI for innovation
While the potential and benefits of AI-generated code are undoubtedly exciting, it’s important to note that these technologies are still evolving and need to be used in the appropriate places. The successful integration of AI into complex real-world development processes requires extra careful consideration, rigorous validation, and ongoing refinement.
As we continue to explore the possibilities of generative AI in software engineering, collaboration between human engineers and AI systems will play a vital role in unlocking its full potential. By harnessing the power of AI and nurturing a symbiotic relationship between humans and machines, we can redefine the boundaries of technological advancement and unlock a new realm of possibilities.
In conclusion, the application of AI in code generation is ushering in a new golden age of software engineering. Through personalised code generation, a unified platform for development tasks and unlocking more time to spend on high-value tasks, we’re empowering all engineers to unleash their creativity and reach their full potential. Our Visual Studio Code extension integrated with JIRA has demonstrated remarkable speed improvements and will revolutionise the way ANZ develops and delivers software.
As we look towards the future, let’s embrace the power of AI to drive innovation, enhance productivity, and shape the future of software engineering. Together, we can redefine the boundaries of what’s possible and chart a new course towards a balanced, sustainable society in which everyone can take part and build a better life.
Louis is an experienced leader in data science and software engineering, with demonstrated experience working in the banking sector. He is currently working as an Engineering AI and Data Analytics Capability Area Lead at ANZ. He is skilled in problem solving, data analytics, automation and management. He is a strong engineering professional with a Doctor of Philosophy (PhD) focused on software engineering (AI - intelligent agent system) from the University of Melbourne.
This article contains general information only – it does not take into account your personal needs, financial circumstances and objectives, it does not constitute any offer or inducement to acquire products and services or is not an endorsement of any products and services. Any opinions or views expressed in the article may not necessarily be the opinions or views of the ANZ Group, and to the maximum extent permitted by law, the ANZ Group makes no representation and gives no warranty as to the accuracy, currency or completeness of any information contained.