Increase productivity with artificial intelligence AI and the challenges for software developers

Related Vendors

GitLab GmbH

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

Wuxi InfiMotion Technology Co., Ltd.

SIMTOS (Seoul International Machine Tool Show)

AI plays an increasingly significant role in software development and has the potential to sustainably transform it. However, as with all newly introduced tools, there are growing pains that can complicate the transition.

The advantages of AI-driven software development are faster iteration cycles, fewer vulnerabilities, and less time spent on administrative tasks. All of this enables companies to bring software to market faster than before. However, to achieve these productivity gains, companies must consider not only the introduction of AI-driven tools but also process- and culture-specific changes.

Closing the AI training gap

In the recently conducted DevSecOps Report by GitLab, 25% of the surveyed employees stated that their companies do not provide adequate training and resources for the use of AI. However, only 15% of the executives agreed with this. Thus, they and their teams perceive investments in AI training differently.

One possible explanation for this is that some companies still see AI as a potential replacement for developers rather than as a tool for creative, strategic, and human-centered work. However, to truly benefit from AI, companies must complement their investments in AI with training and development resources. This enables their teams to build momentum and valuable motivation over time.

Additionally, companies should give their teams a grace period to figure out how AI best fits into their processes. Initially, productivity may decrease as teams adapt to new workflows. But only by testing how AI best integrates into daily workflows can teams build confidence in new tools and achieve better long-term results.

Reduce sprawling toolchains

Another factor affecting developers' experiences and productivity is sprawling toolchains. The same applies when there are too many standalone solutions in software development workflows. The GitLab's DevSecOps Report found that two-thirds of DevSecOps experts want to consolidate their toolchain. Many cite the negative impact on the developer experience due to context switching between tools as the reason.

A sprawling toolchain has further disadvantages. For example, it leads to additional costs and complexity, creates information silos, and makes standardizing processes across different teams more difficult. Additionally, security concerns arise due to increasing attack surfaces and unnecessary handoff points. AI-driven point solutions exacerbate these problems further.

Leaders should rely on best practices when implementing AI, instead of trying to integrate it into cumbersome and complex toolchains. These should first be examined to see where streamlining of inconsistent tools can be made or where they can be eliminated altogether. Otherwise, there will be an additional burden from integrating AI solutions into already unnecessary areas.

Rethink outdated productivity metrics

Developer productivity is one of the main concerns of leadership. Almost all executives surveyed in the GitLab DevSecOps Report (99%) stated that measuring developer productivity could drive company growth. However, 45 percent do not measure it against business outcomes. It has always been challenging to accurately measure developer productivity, but the introduction of AI has further complicated this.

Many companies find it difficult to quantify the impact of AI-powered tools on their developers' productivity or accurately measure how output affects business results. Traditional metrics like lines of code, code commits, or task completion are no longer sufficient for this.

The best approach to modernizing measurement methods begins with combining quantitative data from across the software development cycle with insights from developers on how AI supports or hinders their daily work.

To determine the effectiveness of AI in software development, companies should assess ROI (short for: Return on Investment) based on user acceptance, time to market, revenue, and customer satisfaction metrics. Relevant business outcomes to monitor will likely vary depending on the company, department, and project.

It is evident that AI has the potential to accelerate and further develop DevSecOps practices. However, to achieve faster productivity gains, companies need to navigate potential obstacles. They can do this by proactively addressing cultural and process-oriented challenges that may arise in the early stages of AI implementation.