Protect AI
Protect AI was a cybersecurity company focused on securing AI and machine-learning systems, models and applications. Its stated mission was to help customers build a safer AI-powered world by addressing AI/ML supply-chain and deployment risks.
At a Glance
- Fortune 500 enterprises
- Financial services and fintech
- Healthcare and life sciences
- Energy
- +5 more
AI Tools by Protect AI
(1)
LLM Guard
Python LLM Security Scanner Toolkit
Discussions
No discussions yet
Be the first to start a discussion about Protect AI
Latest News
Palo Alto Networks completed its acquisition of Protect AI; the team and technology were integrated into Prisma AIRS.
Palo Alto Networks announced its definitive agreement to acquire Protect AI and outlined Prisma AIRS lifecycle-security capabilities enabled by the deal.
Protect AI and Leidos announced a partnership to secure AI across U.S. government systems and the AI supply chain.
Protect AI and Palo Alto Networks showcased Prisma AIRS and AI-security innovations at Black Hat USA 2025.
Products & Services
A Jupyter Notebook security tool that scans notebooks for credentials, personally identifiable information, risky code and nonpermissive open-source licenses, and provides remediation guidance.
A visibility and discovery tool for AI/ML environments that maps model components, training data, test datasets and code into a machine-learning bill of materials (MLBOM), and identifies vulnerabilities, threats, data leakage and model-poisoning risk.
Open-source scanner for detecting malicious or unsafe code in serialized machine-learning model files; distributed as a Python package and licensed under Apache-2.0.
A secure model-security gateway and control plane that scans first- and third-party models and enforces policies to prevent malicious code and vulnerable models from entering an AI environment.
Market Position
Protect AI positioned itself as an early specialist in MLSecOps and end-to-end AI/ML security, spanning model supply-chain scanning, posture management, red teaming and runtime controls rather than only generic application security. Reported competitors and adjacent players included HiddenLayer, Robust Intelligence, Resistant AI, CalypsoAI and Troj.ai; after the acquisition, its capabilities were integrated into Palo Alto Networks' Prisma AIRS platform.
Leadership
Founders
Ian Swanson
Co-founder and CEO; previously worldwide leader of AI customer solutions at Amazon Web Services and VP of AI and machine learning at Oracle. He previously founded Sometrics, acquired by American Express, and DataScience.com, acquired by Oracle.
Daryan Dehghanpisheh
Co-founder and former president; previously a global leader for machine-learning solution architects at AWS and worked with Swanson at AWS and Oracle/DataScience.com.
Badar Ahmed
Co-founder and CTO/chief technologist; previously an engineering leader at Oracle and DataScience.com, where he worked with Swanson.
Executive Team
Ian Swanson
Co-founder and CEO of Protect AI; VP, Product, Prisma AIRS at Palo Alto Networks after acquisition
Former AWS worldwide AI/ML leader and Oracle VP of AI/ML; serial founder of Sometrics and DataScience.com.
Daryan Dehghanpisheh
Co-founder and former President; go-to-market lead for AI security in North America at Palo Alto Networks after acquisition
Former AWS leader for machine-learning solution architects and former Oracle/DataScience.com executive.
Board of Directors
Founding Story
The founders drew on experience deploying large AI/ML systems at AWS and Oracle and saw that AI adoption was creating security blind spots that conventional security tools did not address. They started Protect AI to create purpose-built defenses for AI/ML pipelines, models and applications and to help organizations build a safer AI-powered world.
Business Model
Revenue Model
Enterprise AI/ML security software and services, including paid platform capabilities for model security, posture management, testing/red teaming and runtime protection; open-source projects such as ModelScan and LLM Guard provided free community tooling and an adoption channel.
Pricing Tiers
ModelScan and LLM Guard were distributed as open-source tools; Protect AI's early NB Defense offering was available free to start with paid options planned.
Target Markets
- Fortune 500 enterprises
- Financial services and fintech
- Healthcare and life sciences
- Energy
- Government and national-security organizations
- Manufacturing
- Securing AI/ML development and software supply chains
- Scanning and governing third-party and internally developed models
- Protecting Jupyter-based data-science environments
- Testing and red-teaming LLM and AI applications before deployment
- Runtime protection for production AI applications and agents
- AI governance in finance, healthcare, life sciences, energy, government, manufacturing and retail
- Leidos
- TELUS Digital
- Fortune 500 companies in finance, healthcare, life sciences and energy