5 Expert Tips for Harnessing AI in Software Development

Last week, our colleague Matt Belcher, Principal Software Craftsperson at Codurance spoke to an audience of 350+ as part of TechUK’s panel discussion series Exploring the Future of Work. 

The panel aimed to address the myriad of questions around how AI might change the job of a software engineer in the future, whether AI tools will make the role redundant and if low/no code tools will become the norm in the future since there is an ongoing skills gap in the industry.

Over the past year, fueled by the hype around ChatGPT in particular, AI is now everywhere. Many businesses are looking to build AI into their products and funding for startups is often only successful if the product has an AI element. Tech businesses must ensure their software engineers are at least evaluating how AI development such as Github Copilot, Codewhisperer and the like can help them move faster and release new features faster and into the hands of customers. 

One recent survey found that 92% of coders are already using AI tools, with 70% saying these tools are helping them do better work. But many of those in professional software engineering roles, just under half, are anxious or somewhat uneasy about how AI will change their jobs, with a similar number concerned that the skills they currently use as a software engineer will rapidly become obsolete.

The panel, consisting of industry experts and led by TechUK, provided some interesting points that we have summarised below in our Top 5 Takeaways:

1. Understand where the tools can excel and work out what parts of our jobs as software engineers we can offload to the AI to make us more productive. Remember to think about security, governance, ethical and privacy issues - AI is still relatively new and unregulated so must be used with caution.
2. Whilst mundane tasks can be handed off to a GenerativeAI tool, software engineers are freed to deliver value-add activities. A new offshoot of the role called “Prompt Engineering” may become a new practice - where software engineers interact with the tools to get the right outputs. 
3. Its likely that the software engineer role will evolve and human skills such as understanding the wider context when designing software architecture will become more important as AI covers the basics. 
4. Many businesses are very wary of inexperienced junior developers using AI coding tools since there are many negative potential pitfalls and outcomes. These companies are instead starting their use of AI in QA to automate processes and become more efficient. 
5. AI could well become just another tool in the toolkit to drive efficiency and productivity; rather than being a replacement, instead see AI as a way to enrich developer experience and invest in the soft skills needed to build better products.

Matt’s final recommendation to get started was to simply start using the tools - whether you are a software engineer or an Engineering Leader, start by doing and learning. AI has to be part of the plan going forwards, there is plenty to be optimistic about but proceed with caution.  We are still in the early days of AI, many AI jobs are yet to be invented so it's best to stay ahead of the curve, get involved and see where AI tools can make you more valuable to your organisation or the wider tech industry.

Want to learn more? Download our free ebook “Is AI about to Revolutionise Software Development? By Matt Belcher & Danish Javed. Read more Data and AI/ML thought leadership at https://www.codurance.com/data-and-ai-ml

Join our events at https://www.codurance.com/events/ 

Codurance is a member of TechUK, a UK trade organisation that champions technology's role in preparing and empowering the UK for what comes next, delivering a better future for people, society, the economy and the planet.