Case Study: Zest AI secures its AI lending assistant with Cobalt

Zest AI advances LuLu from beta to market with Cobalt after finding 1 critical prompt injection flaw

Zest AI, a leader in credit risk assessment, faced a new security challenge with the development of LuLu, their LLM-powered lending intelligence assistant. They needed specialized pentesting to assess its unique vulnerabilities and turned to their trusted vendor, Cobalt, to utilize its LLM application pentesting service.

Cobalt followed its proprietary methodology based on the OWASP Top 10 for LLMs. The pentest uncovered and helped remediate a critical indirect prompt injection vulnerability. The detailed findings provided valuable insights that led to broader application improvements and contractual updates. This allowed Zest AI to confidently advance LuLu from beta to market, reinforcing their commitment to secure innovation.

Zest AI