How Gen AI Is Transforming Software Development

Gen AI is far more than a tool for writing business emails or conducting research. In software development, tools like GitHub Copilot are already transforming the way developers approach programming tasks. As it stands, Gen AI acts much like a very smart auto-suggest tool that enables developers to write and debug code much faster — twice as fast, according to a study by McKinsey. But this is just the beginning. We will see even greater productivity as the technology improves and new use cases arise.

Developers are embracing Gen AI

While some developers are still working with less modern text editors like Vi or Emacs that poorly support integrations, Gen AI usage is soaring among those working with more agile IDEs like VSCode or IntelliJ. Applause’s 2024 Gen AI Survey found that 42% of respondents have used Gen AI to build apps, up from just 4% in 2023. The majority are using Copilot (41%) followed by Codex by Open AI (24%). Others have experimented with apps like TabNine, AskCodi, CodiumAI, Kite, Amazon CodeWhisperer, AI21 Labs Studio and DeepCode.

Research suggests there are differences in how developers leverage Gen AI depending on their level of experience. For junior developers, Gen AI seems to be most helpful for suggesting code, saving time they would usually spend looking up solutions online. For senior developers, Gen AI is most helpful in the design and structuring of systems. As Gen AI output often largely depends on the quality of prompts, senior engineers also tend to get more value because they are better able to describe the problem they need to solve.

Use cases are in their infancy, but expanding

Gen AI does more than simply predict the next line of code in the way Siri suggests the next word users are going to type in a text. It suggests whole blocks of code based on both recent blocks and the context of the wider project, taking into account language syntax, coding styles and standards, etc. This is a game changer for developer productivity. However, this is just one use case for Gen AI in software development. Other common use cases include:

• Writing requirements
• Writing test instructions
• Writing test cases
• Writing code (code/function completion)
• Checking code
• Fixing code
• Writing unit tests
• Writing functional tests
• Writing test cases
• Writing automated tests

Gen AI is especially helpful for quality assurance efforts, from suggesting improvements to test cases and detecting potential issues early. The most common use cases cited by respondents to Applause’s Gen AI survey were: test case generation (19%), text generation for test data (17%), test reporting (16%) and chatbot testing (15%). 

While Gen AI is currently used for these more repetitive, tedious coding tasks, it is still a nascent space and its value for developers will only increase further. In the future, Gen AI could be used for anything from creating HTML, CSS or Javascript from a screenshot to pointing out programs and carrying out automatic refactoring. It is going to completely change the industry.

Putting guardrails in place is essential

Just like code written by developers, AI-generated code needs to be thoroughly checked. Issues can run the gamut from choosing the wrong variable names to creating code that looks convincingly real but does not run at all. Gen AI tools often need spoon feeding to tackle a challenge. They can generate erroneous code and sometimes make assumptions in order to complete a task. Developers must understand that they still own the end result and go through code or test cases line by line. If you plan to integrate the technology into your workflows, it is worth investing in lint tools and unit tests that can catch errors introduced by Gen AI. 

Managers also need to plan how to roll out Gen AI to their teams. Before advocating for Gen AI use in the corporate environment, managers need to put guidelines in place to avoid data privacy breaches and other legal concerns. They should also make sure developers have selected the correct controls around copyright to attribute open-source licenses properly.

Gen AI skills are important for employers

Understanding how to leverage Gen AI is not just nice to have. It is already influencing the job market. Developers with AI expertise are likely to make for more attractive candidates and could earn greater salaries. A report from Amazon Web Services found that employers are willing to pay an average of 47% more for IT workers with AI skills. 

When I am hiring developers for my team, I look for the A-players. That doesn’t necessarily only mean the people with the best qualifications or even skillset. Today, it also means those who are experimenting with new technologies like Gen AI. Ideal candidates will have already integrated Gen AI into their workflows in a way that increases productivity without sacrificing quality, understand the technology’s limitations and take organizational context into account when reviewing off-the-shelf code.

The industry is changing

To be clear: Gen AI is not here to replace developers, it is here to enhance their work. Developers using Gen AI don’t just get the chance to accelerate code development — they also get to witness the evolution of a technology that is going to transform their industry. It’s a great time to be working in software development.

E-Books