Visa AI: Continuous Security Testing for Software Flaws

Visa has launched a new AI-powered system to find and fix software vulnerabilities. This development signals a big shift in how security testing is done. The payments company recently worked with Anthropic's Project Glasswing, using advanced AI to spot software weaknesses. While many focus on AI finding problems, Visa believes the real challenge is what comes next. AI is rapidly identifying software flaws. This puts pressure on organizations to validate and fix these issues faster than traditional methods allow. Rajat Taneja, Visa's President of Technology, explains that this AI can build "attack chains," not just find single bugs. It can connect less severe issues to create a bigger problem. This capability has major implications. Security testing has often been separate from quality assurance. Now, AI-powered discovery could force a fundamental change. Taneja says the emphasis will shift to validating, prioritizing, and fixing issues with speed and automation. This is similar to how software testing evolved with automated frameworks like Selenium. Visa's new system, called the Visa Vulnerability Agentic Harness, aims to validate findings, suggest fixes, and automatically check if those fixes work. This could bring continuous security testing to the forefront for all organizations.