What we know now about generative AI for software development

Last year, I wrote about the 10 ways generative AI would transform software development, including early use cases in code generation, code validation, and other improvements in the software development process. Over the past year, I’ve also covered how genAI impacts low-code development, using genAI for quality assurance in continuous testing, and using AI and machine learning for dataops.

Now just over a year into the adoption of AI copilots and other genAI capabilities, it’s a good time to review how developers and devops teams are using generative AI. I will also discuss the impact of genAI-enabled tools and processes and the risks that must be addressed.

How developers and devops teams are using genAI today

“GenAI is redefining how developers work by introducing a new era of collaborative programming, acting as a dynamic partner, streamlining workflows, augmenting developer skills, and enhancing creative problem solving,” says Bharat Sandhu, SVP of the SAP Business Technology Platform. “By embedding genAI into development processes, organizations unlock new possibilities and foster a more agile, collaborative approach to building the future.”

Recent data suggests that enterprises seek proprietary genAI capabilities, and most devops teams and developers use it for various tasks in the software development lifecycle.

The 2024 AI in software development report found AI application development in high demand, with 79% of enterprises having up to 300 use cases for generative AI in their backlogs. According to the report, genAI’s most popular emerging use cases in software development included devops optimizations, code generation, documentation, and user-interface design.

Of all the genAI use cases, code generation continues to receive the most significant attention. AI copilots help skilled developers save time while advancing new skills in prompting and code reviewing.

“GenAI is used primarily for code, unit test, and functional test generation, and its accuracy depends on providing proper context and prompts,” says David Brooks, SVP of evangelism at Copado. “Skilled developers can see 80% accuracy, but not on the first response. With all of the back and forth, time savings are in the 20% range now but should approach 50% in the near future.”

AI coding assistants also help junior developers learn coding skills, automate test cases, and address code-level technical debt.

“Organizations can leverage AI coding assistants to bridge skill gaps and respond to market demands more efficiently,” says Angel Montesdeoca, program director of product management for IBM’s watsonx code assistant. “Rather than replacing developers, AI assistants empower them to focus on creativity and complex problem-solving. With AI assistance, developers can translate between programming languages, generate tests, modernize legacy systems, and deliver higher-quality software faster.”

AI copilots take on advanced coding challenges

Next, I asked experts about how AI coding assistants can help with more advanced coding challenges.

“GenAI is currently easiest to apply to application prototyping because it can write the project scaffolding from scratch, which overcomes the ‘blank sheet of paper’ problem where it can be difficult to get started from nothing,” says Matt Makai, VP of developer relations and experience at LaunchDarkly. “It’s also exceptional for integrating web RESTful APIs into existing projects because the amount of code that needs to be generated is not typically too much to fit into an LLM’s context window. Finally, genAI is great for creating unit tests either as part of a test-driven development workflow or just to check assumptions about blocks of code.”

One promising use case is helping developers review code they didn’t create to fix issues, modernize, or migrate to other platforms.

“AI coding assistants can now understand large codebases, fix bugs, and automate routine tasks like generating documentation or running simple code reviews,” says Andrew Filev, Founder & CEO of Zencoder. “In 2025, more than 25% of new production code could be AI-generated, enabling teams to tackle tech debt and improve software quality faster than ever.”

Devops teams leverage genAI for code quality and reliability

Quality assurance practices, including test automation and code reviews, are another area where genAI provides value to devops teams. In the 2024 State of Software Quality report, 58% of respondents said that time constraints were their most significant challenge when performing code reviews. According to the report, more than 50% of respondents were using AI in some aspects of code reviews.

The State of Software Quality Report 2024 focused on quality engineering and test automation. According to this report, 45% of respondents cited lack of time and skills as a primary obstacle to test automation. However, they categorized AI adoption as low, with only 45% using AI-augmented tools for quality engineering.

“Using genAI for testing and streamlining the QA part of the SDLC brings an improved final product and user experience,” says Noah Borts, co-founder and COO of HappyPath. “Software that goes through genAI testing processes is checked for errors outside of scenarios the human mind alone might foresee.”

Another function that can benefit from AI is site reliability engineering (SRE), which is tasked with improving service-level objectives and reducing production errors. The SRE Report 2025 found toil, or the burden of operational tasks, on the rise for the first time in the survey’s history.

You might expect site reliability engineers to seek AI benefits, but 37% of respondents said they were investing in or implementing AI technologies with caution, and only 24% said they were eager to use AI technologies.

On the other hand, SREs using AI capabilities report their benefits and objectives.

“We expect genAI-driven efficiencies to help our engineers and SREs to be more innovative, productive, and improve job satisfaction,” says Debo Dutta, VP of engineering at Nutanix. “Engineers can use AI to search for code better, explain existing code, and help write unit tests to improve software quality, while SREs can use genAI tools to provide quicker responses for customer support cases. Our goal is to achieve more than 25% improvement in productivity for the business with genAI.”

Measuring genAI’s productivity benefits

Released in January 2025, the IBM-sponsored Enterprise AI Development: Obstacles & Opportunities report found that AI-assisted coding tools were helping 64% of developers save over an hour per day. However, tool proliferation may impact the full business benefit, as 72% reported using between five and 15 tools to create an AI enterprise application, according to the report.

The JetBrains State of Developer Ecosystem Report 2024 took a more conservative view, with 59% of developers saying they saved less than four hours weekly using AI tools for development-related activities. The top benefits cited by at least 50% of respondents included less time spent searching for information, faster coding, and faster completion of repetitive tasks.

“With genAI autocomplete, developers can save around 20% of the time previously spent looking up code and library documentation or the right pseudocode to achieve a task of small complexity,” says Rob Skillington, CTO and co-founder of Chronosphere. “Autocomplete serves as a reasonable speed-up; plus, even if incorrect, it can help outline precisely what ambiguities need looking up.”

Brooks of Copado adds, “Developers spend this extra time tackling technical debt, researching new technologies, and learning new skills. The impact on business is a small increase in throughput and higher quality. UX will improve now that AI can generate complex UI quickly and teams can iterate multiple times in a sprint.”

Small time savings during the agile development sprints can yield larger benefits when aggregated across functional release cycles. When developers reduce time on coding tasks, they can spend more effort improving user experiences and developing robust architectures.

“In the hands of a senior developer, generative AI is a true time saver, quickly writing boilerplate code, drafting complex algorithms, and expanding unit tests,” says Mike Rinehart, VP of AI at Securiti. “Ultimately, generative AI allows senior developers to spend more mental effort focusing on architecture and higher-level code structure, which in turn improves the long-term health of the product.”

Borts of HappyPath adds, “As genAI tooling is adopted, developers have more time for innovation and improving existing products. There is more focus on strategic tasks like solving complex architectural challenges, mentoring junior team members, or deepening expertise in emerging technologies.”

Evaluating the risks of AI coding assistance

Despite strong adoption and business benefits, some leaders highlight the risks of AI code assistance. Organizations adopting AI for devops and software development should define non-negotiables, train teams on safe utilization, identify practices to validate the quality of AI results, and capture metrics that reveal AI-delivered business value.

“AI poses risks to code quality and security that can’t be ignored, making code reviews and analysis still a critical part of the development process,” says Andrea Malagodi, CIO of Sonar. “Without proper checks and reviews, AI-generated code may lead to poor software quality and increased tech debt. To maximize AI’s productivity benefits, developers must have accountability for code quality and adopt a ‘trust and verify’ approach, ensuring all code—AI-generated or human-written—meets quality and security requirements, and the user experience is not disrupted.”

Bogdan Raduta, head of AI at FlowX.AI, raises questions about quality and innovation when businesses rely too heavily on generic user experiences and AI defaults to patterns and conventions. “While faster development reduces costs, businesses may deliver functional but uninspired products, opening opportunities for competitors to stand out with bespoke, human-driven designs,” Raduta says.

Developers should continue to explore AI capabilities for building software and developing experiences, especially because these capabilities are evolving quickly. While experimentation is needed, devops teams and IT departments should create target goals and metrics for AI benefits while seeking benchmarks for where other organizations are delivering value.