Protect AI develops software for securing AI and machine learning systems across the full lifecycle of enterprise AI applications. Its unified platform is built to help organizations evaluate model risk, test AI behavior before release, enforce security controls, monitor production usage, and respond to vulnerabilities across models, applications, and supporting supply chains.
The portfolio spans core platform products, open source tooling, and AI-specific vulnerability intelligence for predictive, generative, and agentic AI use cases. Protect AI is positioned for organizations that need dedicated AI security controls alongside existing cloud and security operations, and its website indicates the company is now part of Palo Alto Networks.
Offerings, Capabilities, and Integrations
Protect AI centers its offering on AI security posture management across the build, buy, and run phases of AI adoption. Its capabilities cover model inspection, policy enforcement, application red teaming, runtime threat detection, supply chain vulnerability intelligence, governance, and remediation so teams can move AI systems into production with stronger control and visibility.
The platform is designed to fit into existing enterprise environments rather than replace them. Protect AI supports local and pipeline-based deployment patterns, connects with model repositories and cloud AI services, and integrates with common observability and response tooling. Its coverage is also informed by ongoing threat research from the huntr community and aligned to widely used security frameworks and operational practices.
Products and Services
- Guardian: AI model security product for scanning first- and third-party models, enforcing model security policies, and governing model intake across enterprise AI workflows.
- Recon: AI red teaming product that tests AI applications against a large attack library, generates relevant attacks, and helps teams assess risks before deployment.
- Layer: Runtime security product that monitors AI applications in production, detects threats across prompts, tool use, workflows, and metadata, and supports policy-based control.
- ModelScan: Open source machine learning model scanner that checks models for unsafe code and model serialization attacks across multiple model formats.
- LLM Guard: Open source LLM security toolkit for detecting, redacting, and sanitizing prompts and responses to reduce risks such as prompt injection, harmful content, and data leakage.
- Sightline: AI and ML supply chain vulnerability database that provides security advisories, remediation guidance, vulnerability scanners, and early warning on emerging AI risks.
- MLSecOps Foundations: Training and certification program that teaches organizations how to build security into AI and machine learning processes using an MLSecOps framework.
Target Customers
Protect AI targets organizations moving AI into production and needing controls that traditional cybersecurity tools do not fully address. Its buyers typically include application security, information security, ML engineering, platform engineering, and DevSecOps or MLSecOps teams responsible for model governance, testing, deployment, and runtime oversight.
The company is suited to large enterprises and public sector organizations running traditional ML, LLM, RAG, and agentic AI workloads. Its messaging and partnerships point especially to regulated and AI-intensive sectors such as financial services, healthcare, life sciences, government, energy, manufacturing, automotive, and e-commerce, as well as Fortune 500 and security-sensitive environments.
Cloud Integrations and Marketplace
- AWS Marketplace: Protect AI is available as a SaaS offering in AWS Marketplace, where customers can procure the Protect AI platform for AI and ML security use cases.
- Azure Marketplace: Protect AI is listed in Azure Marketplace as a SaaS offering focused on helping organizations see, know, and manage AI security risks.
- Amazon Bedrock: Protect AI provides an official integration for Guardian and Recon with Amazon Bedrock, and its AWS alliance also highlights coverage for Amazon SageMaker workflows.
Key People
- Ian Swanson: CEO
- Daryan Dehghanpisheh: Co-Founder and President
- Badar Ahmed: Chief Technology Officer
- Diana Kelley: Chief Information Security Officer
- Ralph Pisani: Chief Revenue Officer
- Zoe Hillenmeyer: Chief Marketing Officer
- Chris King: Head of Product
- Sean Morgan: Chief Architect
Key Facts
- Headquarters: Seattle, Washington, United States
- Employees: 51-200
- Annual Revenue: $5M
- Parent Company: Palo Alto Networks
- Subsidiaries: huntr, Laiyer AI, SydeLabs, and Rebuff
- Publicly Listed: Not separately listed; parent company Palo Alto Networks is listed on NASDAQ under PANW
