AWS found bugs in 60% of software requirements. Its fix isn’t more AI — it’s a 50-year-old logic engine.
According to a report by The New Stack, AWS discovered that 60% of software requirements contain bugs. To address this, AWS is using Kiro, a requirements analysis tool based on automated reasoning—a logic engine that dates back 50 years. Kiro applies formal methods to detect inconsistencies, ambiguities, and missing constraints in natural language requirements. This approach contrasts with the current trend of using AI for code generation. By catching errors at the requirements stage, AWS aims to reduce the cost and effort of fixing bugs later in the development cycle. The tool is already in use internally and has shown significant improvements in requirement quality. AWS plans to make Kiro available to customers, though no specific pricing or release date has been announced. The use of automated reasoning over AI highlights a shift toward more rigorous verification methods in software engineering.
Catching bugs in requirements before coding saves time and money.