NEW YORK, June 12, 2025 /PRNewswire/ — Qodo, the quality-first AI coding platform, released today a comprehensive survey of over 600 developers, revealing that despite widespread adoption of AI coding tools there is still a striking confidence crisis: while 82% use AI coding assistants daily or weekly, very few fully trust the generated code. The ‘The State of AI Code Quality 2025’ report analyzes how AI tools affect software development across productivity, code quality, and developer experience.
Key findings from the report include:
AI adoption is mainstream – 82% of developers use AI coding tools daily or weekly
Productivity advances with AI – 78% of developers experience productivity improvements from AI coding tools
But relevant context is missing – 65% of developers say AI misses relevant context during critical tasks like refactoring, writing tests, or reviewing code
AI coding tool market isn’t winner takes all – 59% of developers are using three or more different AI coding tools
Job satisfaction improves – 57% of developers say AI makes their job more enjoyable or relieves pressure, with only 20% reporting increased burnout
Overall improved quality from AI – 60% of developers say AI has improved code quality, only 18% say AI has degraded it
AI code review correlates with improved quality – Teams integrating AI code review gain a significant quality edge – reporting 35% higher rates of code quality improvement than teams without automated review
The study reveals a critical trust gap undermining AI coding’s potential. While 78% of developers report productivity gains from AI tools, a striking 76% still don’t fully trust the generated code—creating a bottleneck that erodes the very speed benefits that automated code generation promises as developers shift towards manual reviewing of code. However, the research identifies automated AI-empowered code review as the key catalyst that transforms this dynamic. Teams integrating AI review into their workflow see quality improvements jump to 81% compared to just 55% for similar teams without review, creating a ‘Confidence Flywheel’ – where better quality builds trust, leading to faster shipping and better outcomes.
“Developers have clearly embraced AI tools and are seeing benefits in productivity,” said Itamar Friedman, CEO and co-founder of Qodo. “But, the most successful teams aren’t using AI just to generate more lines of code, they are integrating AI with quality practices, so they can trust the software they produce, and create faster, better software as a result.”
Story continues
