Visa Expands AI Cybersecurity Tools to Cut Fix Time From Weeks to Hours
Visa's updated Vulnerability Agentic Harness and three new advisory services aim to cut the time from vulnerability discovery to resolution, with some fixes shrinking from weeks to hours.
Visa released an updated version of its open-source cybersecurity framework and launched three new advisory services, the company said on Aug. 28. The moves aim to shrink the time between discovering a vulnerability and fixing it, a metric Visa calls Mean Time to Adapt.
The latest Visa Vulnerability Agentic Harness, or VVAH, extends the original tool beyond vulnerability discovery into remediation and validation. The framework is model-agnostic and was first released after Visa participated in Anthropic's Project Glasswing, a frontier AI security initiative. Since the open-source release in June 2026, tens of thousands of developers have downloaded VVAH, according to the company.
Visa Consulting & Analytics introduced three cybersecurity advisory services informed by the company's work on VVAH and Project Glasswing. The services help clients assess risk, prioritize remediation and strengthen operational resilience.
"Across Asia Pacific, AI is accelerating both the scale of cyber threats and the speed at which vulnerabilities can be exploited," said Prateek Sanghi, head of Visa Consulting & Analytics for Asia Pacific. "In this environment, identifying vulnerabilities is no longer the differentiator. The real challenge is determining which risks matter most and remediating them before they can be leveraged by attackers."
Sanghi said the shift from counting vulnerabilities to measuring the Mean Time to Adapt has become "more critical than ever."
Visa cited CAIXA Cartões, a Brazilian credit card issuer, as a client that used the advisory services for a cybersecurity maturity assessment. Lessandro Thomaz, executive director at CAIXA Cartões, said the partnership "helped broaden our strategic perspective on cybersecurity."
The company also joined two industry initiatives: NVIDIA's Open Secure AI Alliance, to which Visa contributed VVAH as a model-agnostic framework, and the IBM and Red Hat-led Project Lightwell, which focuses on securing open-source software.
Drafted by a large language model from the source reporting linked above, then screened by automated publishing checks. It is not read by a journalist before publication. Some articles cite our Alpha Score. Verify prices and figures against the original source. Educational coverage, not personalized advice.