Seed · Cybersecurity · Paris, France · Founded 2021 · 11–50 employees
Giskard AI is a Paris-based open-core platform specialising in automated red teaming, LLM security testing, hallucination detection, and prompt injection prevention for enterprise AI agents. Founded in 2021 by veterans from Dataiku and Thales, it operates an open-source Python library alongside a paid enterprise hub (Giskard Hub), and in May 2026 launched Giskard Guards — billed as Europe's first independent sovereign guardrail platform for enterprise AI agents. The company has raised a €1.5M seed round led by Elaia with Bessemer Venture Partners, plus a €3M strategic grant from Bpifrance and the European Commission's EIC Accelerator, and its open-source library has accumulated over 5,000 GitHub stars. Giskard is trusted by regulated-sector enterprises including AXA, BNP Paribas, Michelin, Google DeepMind, and Doctolib, and is positioned as a key enabler of EU AI Act compliance.