Cisco AI Defense
Cisco AI Defense helps enterprises develop, deploy, use, and govern AI securely. Its end-to-end approach combines AI asset discovery, model and application validation, runtime guardrails, access control, and AI supply-chain risk management for AI applications, agents, models, and connected tools.
At a Glance
- Large enterprises and security teams
- CISOs, CIOs, AI governance, risk, and compliance teams
- AI/ML developers and application engineering teams
- Organizations operating LLM applications, agents, MCP servers, and RAG systems
- +2 more
AI Tools by Cisco AI Defense
(1)
Project Quarry
Agent Conversation Threat Detection
Discussions
No discussions yet
Be the first to start a discussion about Cisco AI Defense
Latest News
Project Quarry open-source repository updated with the initial OSS release and alpha framework for detecting threats and behavior patterns in agent conversations.
Cisco announced DefenseClaw, an open-source governance tool for OpenClaw agents, and related agent-runtime security initiatives.
Cisco AI Defense introduced the Adversarial Hubness Detector for RAG systems and vector-database security.
Cisco published its State of AI Security 2026 report and analysis of the expanding AI threat landscape.
Products & Services
Enterprise AI-security platform spanning AI asset discovery, model and application validation, runtime protection, AI cloud visibility, AI access, and AI supply-chain risk management.
Algorithmic AI red teaming that evaluates models and applications across broad safety and security categories and produces an assessment report; Explorer Edition lets builders try red teaming before adopting an enterprise deployment.
Real-time guardrails for AI applications and agents, including defenses against prompt injection, sensitive-data leakage, toxic content, and agent tool exploitation.
AI Bill of Materials capability and open-source CLI that scans source code or repositories to inventory models, agents, tools, prompts, frameworks, dependencies, and provenance.
Market Position
Cisco positions AI Defense as an end-to-end enterprise AI-security platform differentiated by network-level visibility and enforcement, Cisco Talos threat intelligence, multi-model and multi-cloud coverage, and a lifecycle approach spanning discovery, validation, supply-chain security, access, and runtime protection. Its competitive set includes AI security and governance vendors such as Lakera, Protect AI, HiddenLayer, CalypsoAI, Robust Intelligence technology before its acquisition, and cloud-provider guardrail/security offerings.
Leadership
Executive Team
DJ Sampath
Senior Vice President and General Manager, Cisco AI Software and Platform
Founded and led Armorblox, joined Cisco through its 2023 acquisition; previously VP of Engineering and founding team member at StackRox, which was acquired by Red Hat, and held roles at Deutsche Telekom, Juniper Networks, PARC, and BBN Technologies.
Akshay Bhargava
Vice President, Product Management, AI Software & Platform
Leads development of AI Canvas, AI Defense, and AI Studio. Previously Group VP of Product Management at New Relic, Chief Product Officer and GM at 1Password, and held senior roles at Malwarebytes, Oracle Cloud, FireEye, and Rainfinity.
Founding Story
Cisco announced AI Defense on January 15, 2025 in response to the security and safety risks created by rapid enterprise AI adoption, including prompt injection, data loss, model vulnerabilities, and unsanctioned AI use. Cisco positioned it as a purpose-built, network-integrated solution so organizations could move quickly on AI without sacrificing safety or security; Robust Intelligence's AI-security technology, acquired by Cisco in 2024, became foundational to the offering.
Business Model
Revenue Model
Enterprise software and security-platform licensing, sold through Cisco and Cisco Security Cloud Control; AWS Marketplace describes pricing as based on customer requirements and eligibility through a private offer. Cisco also offers an Explorer Edition for AI red teaming and open-source tools such as AI BOM and Project Quarry.
Pricing Tiers
Self-service AI red teaming for builders to scan AI applications and receive an assessment report before moving to an enterprise solution.
Enterprise deployment pricing is based on customer requirements and eligibility; available through Cisco and AWS Marketplace.
Target Markets
- Large enterprises and security teams
- CISOs, CIOs, AI governance, risk, and compliance teams
- AI/ML developers and application engineering teams
- Organizations operating LLM applications, agents, MCP servers, and RAG systems
- Cloud, hybrid, on-premises, and AI-infrastructure operators
- Regulated industries requiring AI risk management and standards alignment
- Enterprise governance of employee use of third-party generative-AI applications
- Secure development and pre-deployment testing of LLM applications and AI agents
- Runtime protection of production AI applications and agent workflows
- MCP server, agent skill, plugin, and A2A security
- AI supply-chain inventory, dependency analysis, and model provenance
- AI red teaming and security review preparation
- World Wide Technology
- Splunk customers
- Sompo