The arrival of generative AI has put pressure on software development organizations to demonstrate greater efficiency. In practice, generative AI appears to save about 10% to 15% of total software engineering time. However, many companies aren’t making profitable use of these savings.
This might be the best "AI Engineer" I've tried so far.
I'm an old school developer who started 30 years ago. I feel very uncomfortable letting AI take control of my code, but for the sake of science, I spent two hours building an application that took me weeks to build a… pic.twitter.com/4EQH0Hb85t
— Santiago (@svpino) October 1, 2024
I've been reading a lot of AI stans claiming that AI will replace software developers—clear proof that whoever claims that is not a developer. First, the days of the coder—who just codes up a specification—are over. A software developer does way more than code—it's a…
— Allen Holub @allenholub.bsky.social (@allenholub) September 29, 2024
Improvements of 30% or more are possible, but they require leveraging the full potential of generative AI and adopting a broader agenda. Developers spend about half their time writing and testing code. Although they report a 30% improvement from generative AI in these activities, it translates to a net efficiency improvement of 15% overall.
Asked @simonw what the biggest productivity boost has been, throughout his 20-year software engineer career. He answered, without hesitation:
Open source.
(Clip from @Pragmatic_Eng Podcast from the episode "AI tools for software engineers, but without the hype") pic.twitter.com/pjYZILETsb
— Gergely Orosz (@GergelyOrosz) September 30, 2024
A more comprehensive approach includes not only generative AI-assisted code generation and testing but also focusing on the right work, ensuring speedy and high-quality execution, and optimizing resourcing costs. The fastest way to improve efficiency is to refocus efforts on the work that creates the most value. This involves aligning investments with strategy across products and markets, balancing resource allocation, and linking product strategy to developer priorities.
Clear prioritization can prevent developers from addressing issues that don’t support strategic goals. Improved visibility into how time is spent often reveals a mismatch between leadership’s ambitions and actual resource allocation. Generative AI is a priority today, but foundational elements like continuous delivery and modern architecture can be more effective ways to drive efficiency.
Clear roadmaps, managing tech debt, and ensuring optimal resource allocation are key to improving productivity. Leaders in generative AI adoption can achieve up to 30% efficiency from optimal deployment. For example, Intuit has made significant strides by using its proprietary generative AI operating system (GenOS) to analyze developer documentation and expedite development velocity.
Generative AI boosts development efficiency
Intuit’s broader use of generative AI has reduced integration task completion times and improved the standardization of code and documentation across various development teams. Before deploying new code, developer teams need to ensure that it won’t break anything in the live product or create security risks.
Automated testing in a virtual environment is more efficient and avoids the risks associated with manual testing. Continuous integration and delivery allow developers to assess the effects of new code deployments efficiently and address security threats promptly. Customers also appreciate the consistency and quick response to identified issues.
A modular architecture allows teams to adapt and improve products without needing to start from scratch. Continuous investment in modular design helps avoid technical debt, maintaining competitive edge by keeping up with evolving technology. Two software development organizations operating at similar speeds and quality can have different cost profiles based on their models and talent structures.
Factors such as geographical footprint, outsourcing levels, and the ratio of senior to junior engineers affect costs. An imbalanced staff composition can lead to inefficiencies and higher costs despite apparent savings. Many companies struggle to understand their baseline efficiency and measure improvements from initiatives like generative AI.
About two-thirds of leaders surveyed are not satisfied with their current measurement practices. A clear understanding of baseline efficiency and effective measurement are crucial to realize the full benefits of generative AI and other efficiency-improving initiatives. Generative AI presents significant opportunities for improving software development efficiency, but organizations must take a comprehensive approach to realize its full potential.
By focusing on high-value work, ensuring speedy and high-quality execution, and optimizing resource costs, companies can achieve substantial efficiency gains and better position themselves for future success.
