Giskard

Giskard is an AI red teaming and LLM security platform founded in Paris in 2021. It helps enterprises detect vulnerabilities in AI agents before deployment by providing automated assessments that uncover issues like prompt injection, data disclosure, sycophancy attacks, hallucinations, and inappropriate denials. Trusted by major European regulated enterprises including AXA, BNP Paribas, L'Oreal, Societe Generale, and Decathlon, Giskard delivers go/no-go deployment recommendations signed by AI security experts, full vulnerability reports ranked by severity, and end-to-end remediation guidance. The platform offers EU and US data residency, SOC 2 Type II and HIPAA compliance, and a 0-training policy to protect customer IP. Backed by co-founders of HuggingFace and Mistral, Giskard also maintains an open-source Python testing library used by AI developers globally.