Giskard provides the trust layer to run AI agents safely. Our product automates testing of AI systems to detect and control risks, during development and continuously after deployment. Architected for critical agentic AI and proven through our work with global companies including BNP Paribas, Michelin, and Google DeepMind, our platform helps enterprises deploy AI agents that are not only powerful but actively defended against evolving risks.
Building: scale from {c1} to {c2}. Growing: momentum from 33 to 67. See the map.
No. 19 in this week's Top 40 Movers, down 1 from last week. Highest place 18, 4 weeks on the chart. See the chart.
Giskard's AAA uses sandboxed agents to map tools, model risks, and attack in adaptive waves with reproducible findings.
Giskard · 5 Nov 2026 · 1 source · Vendor sources onlyGiskard releases version 1.0.4 of giskard-checks, an open-source library for evaluating and testing large language model agents.
Giskard · 15 Sep 2026 · 1 source · Also reported independentlyEverything we recorded in the last 180 days.
Send us a story about Securing AI that is not here, tell us your company belongs on the list. We read everything; what goes on the page is our call.
Everything here is open to read. If you want the data in your own tools, the intelligence feed starts at $139 a month, or ask about data access below.
Choosing a product in this market? Advisory helps buyers build a shortlist, test vendor claims and plan a proof of concept.
