{"version":"0.42.0","built_at":"2026-11-01T08:34:43.408Z","license":"CC BY 4.0 © Mike Reams","canonical":"https://catalog.mikereams.com","product":{"id":"PR-790","name":"Cisco AI Defense","vendor":"Cisco Systems, Inc.","category":"AI security","description":"A security product for AI applications and agents that finds the models, agents and MCP servers in use, attacks them in automated tests before release, scans model files and AI components for tampering, and inspects prompts, responses and tool calls at run time. Runs across the major clouds and on-premises.","website":"https://www.cisco.com/site/us/en/products/security/ai-defense/index.html","license_model":"subscription","deployment":["saas","hybrid"],"aliases":["AI Defense"],"csdm_note":"One Business Application for the service. Each protected AI application keeps its own Application Service; record AI Defense as a dependency of those services, since a runtime policy change alters what every one of them returns.","sources":[{"url":"https://www.cisco.com/c/en/us/products/collateral/security/ai-defense/ai-defense-ds.html","title":"Cisco AI Defense Data Sheet - Cisco"},{"url":"https://www.cisco.com/c/en/us/products/collateral/security/ai-defense/ai-defense-so.html","title":"Cisco AI Defense Solution Overview - Cisco"},{"url":"https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m02/cisco-redefines-security-for-the-agentic-era.html","title":"Cisco Redefines Security for the Agentic Era with AI Defense Expansion and AI-Aware SASE"}],"as_of":"2026-11-01","product_capabilities":[{"id":"PC-790.10","name":"AI Inventory","description":"Finds models and agents running in cloud AI services such as Amazon Bedrock, Google Vertex and Azure AI Foundry.","business_capability_ids":["BC-770.10.50","BC-760.40.40"],"maturity":"strong","confidence":"high","source_url":"https://www.cisco.com/c/en/us/products/collateral/security/ai-defense/ai-defense-ds.html","business_capabilities":[{"id":"BC-770.10.50","name":"AI System Inventory"},{"id":"BC-760.40.40","name":"Cloud Security Posture"}]},{"id":"PC-790.20","name":"AI Model and Application Validation","description":"Automated attack runs against models and applications across a broad threat taxonomy, with findings mapped to public AI security frameworks and suggested guardrails for each model.","business_capability_ids":["BC-760.50.20","BC-770.10.30","BC-770.30.40"],"maturity":"core","confidence":"high","source_url":"https://www.cisco.com/c/en/us/products/collateral/security/ai-defense/ai-defense-ds.html","business_capabilities":[{"id":"BC-760.50.20","name":"Application Security Testing"},{"id":"BC-770.10.30","name":"AI Risk & Model Review"},{"id":"BC-770.30.40","name":"Output Evaluation & Guardrails"}]},{"id":"PC-790.30","name":"AI Runtime Protection","description":"Checks prompts and responses as they pass and stops prompt injection, leaks of sensitive data and unsafe content.","business_capability_ids":["BC-760.50.50","BC-770.30.40"],"maturity":"core","confidence":"high","source_url":"https://www.cisco.com/c/en/us/products/collateral/security/ai-defense/ai-defense-ds.html","business_capabilities":[{"id":"BC-760.50.50","name":"AI Runtime Protection"},{"id":"BC-770.30.40","name":"Output Evaluation & Guardrails"}]},{"id":"PC-790.40","name":"AI Supply Chain Risk Management","description":"Checks models, the repositories they come from and MCP servers for concealed threats before developers adopt them.","business_capability_ids":["BC-760.60.30","BC-760.50.10"],"maturity":"strong","confidence":"high","source_url":"https://www.cisco.com/c/en/us/products/collateral/security/ai-defense/ai-defense-ds.html","business_capabilities":[{"id":"BC-760.60.30","name":"Software Supply Chain Integrity"},{"id":"BC-760.50.10","name":"Secure Development Practices"}]},{"id":"PC-790.50","name":"Agent Runtime Guardrails","description":"Watches agent interactions and MCP traffic live, enforces allowed and blocked tools, and stops tool misuse, poisoned tools and attempts to gain more privilege.","business_capability_ids":["BC-760.50.50","BC-770.30.30","BC-770.50.50"],"maturity":"emerging","confidence":"medium","source_url":"https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m02/cisco-redefines-security-for-the-agentic-era.html","business_capabilities":[{"id":"BC-760.50.50","name":"AI Runtime Protection"},{"id":"BC-770.30.30","name":"Agent Orchestration"},{"id":"BC-770.50.50","name":"Model & Tool Access Brokering"}]},{"id":"PC-790.60","name":"AI BOM","description":"A single bill of materials for the AI estate that lists MCP servers and outside dependencies, so AI supply chain risk can be governed.","business_capability_ids":["BC-760.60.30","BC-770.10.50"],"maturity":"emerging","confidence":"medium","source_url":"https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m02/cisco-redefines-security-for-the-agentic-era.html","business_capabilities":[{"id":"BC-760.60.30","name":"Software Supply Chain Integrity"},{"id":"BC-770.10.50","name":"AI System Inventory"}]},{"id":"PC-790.70","name":"MCP Catalog","description":"A list of the MCP servers and registries an organization depends on, public or private, with the risk of each kept up to date.","business_capability_ids":["BC-770.10.50","BC-760.60.30"],"maturity":"emerging","confidence":"medium","source_url":"https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m02/cisco-redefines-security-for-the-agentic-era.html","business_capabilities":[{"id":"BC-770.10.50","name":"AI System Inventory"},{"id":"BC-760.60.30","name":"Software Supply Chain Integrity"}]}],"page_url":"https://catalog.mikereams.com/products/PR-790"}}