Protect AI

1. Products/Services & Features

• Main Offerings:

  • Guardian - AI Model Security Scanner
  • Recon - Automated Red Teaming
  • Layer - Runtime Security

• Feature Breakdown: Guardian: Scans 35+ model formats (PyTorch, TensorFlow, ONNX, Keras, GGUF, Safetensors); detects deserialization attacks, backdoors, runtime threats. Recon: Automated red teaming for AI applications. Layer: Runtime monitoring and threat detection. LLM Guard: Prompt/response security and sanitization. (Departments: Security, Engineering, Product, Marketing)

• Business Industry Gearing: Enterprise, Finance, Healthcare, Government, Technology

2. Security & Compliance

• Certifications: Not publicly documented, MLSecOps Certification program available

• Vendors/Tools: Integrated with Hugging Face, MLOps pipelines, DevOps workflows

• Risk Profile:

  • Breaches: Deserialization attacks, architectural backdoors, runtime vulnerabilities, prompt injection, model poisoning
  • Features: Zero Trust model scanning, comprehensive threat research (17,000+ security researchers), continuous Hugging Face monitoring, multi-format support

3. User Feedback & Adoption

• Aggregated Reviews: No public reviews found on G2 or Capterra

  • Pros: Comprehensive model scanning across 35+ formats, seamless MLOps integration, Zero Trust approach, continuous threat research, easy adoption for AI teams
  • Cons: Limited public user feedback, no transparent pricing, technical expertise may be required

• Adoption Insights:

  • Adoption Ease: High - Designed for seamless integration into existing MLOps workflows without disrupting innovation
  • Adoption Cultural Fit: Strong fit for organizations prioritizing AI security and MLSecOps practices; requires security-first mindset

• Metrics: Not publicly available

• Barriers: Custom pricing requires direct vendor engagement; no public pricing transparency; limited case studies available

4. Monetization & Business Model

• Revenue Model: SaaS subscription with contract-based enterprise pricing

• Pricing: Custom enterprise tiers based on contract terms, usage volumes, and feature requirements; no public pricing available (Sources: AWS Marketplace listing; direct vendor negotiation for enterprise deals)

• Market Context:

  • TAM: AI/ML Security market estimated at $10B+ by 2030
  • Growth Stage: Early growth; rapidly expanding as enterprises adopt AI

5. Leadership & Recent Developments

• Key Metrics Update:

  • Funding: $60 million Series B (August 2024) led by Evolution Equity Partners, Boldstart Ventures, Pelion Venture Partners
  • Employee Growth: Scaled from startup to 51-100 employees in 3 years (2022-2025)

• News/Trends:

  • News Launch: Guardian product launch (January 2024); Prisma AIRS integration (July 2025)
  • News Partnerships: Hugging Face integration; partnerships with Leidos, TELUS Digital; Palo Alto Networks acquisition
  • News Funding: Series B $60M (August 2024); Acquisition by Palo Alto Networks (July 2025)
  • News Challenges: Transition to Palo Alto Networks integration; market education on MLSecOps importance

6. Target Audience & Use Cases

• Target Market: Enterprise organizations deploying AI/ML systems; regulated industries (finance, healthcare, government)

• Target Users & Personas: Security teams, ML engineers, DevOps engineers, data scientists, CISOs

• User Experience Level: Intermediate to advanced; requires understanding of ML systems and security practices

• Key Use Cases:

  • Open-source model security - Scanning models from Hugging Face and other repositories for malicious code
  • Enterprise model governance - Inventory and policy enforcement for thousands of AI models
  • Compliance and vulnerability management - Ensuring AI systems meet regulatory requirements

7. Impact & Recommendations

• Measurable Outcomes:

  • Workflow Improvements: Automated model scanning reduces manual security review time; Zero Trust approach eliminates unsafe model deployment; seamless pipeline integration maintains development velocity
  • ROI Examples: Reduced security incidents from model compromise; faster model deployment with built-in security; compliance audit readiness

• Fit Assessment: Excellent fit for enterprises with significant AI/ML deployments requiring security governance and compliance

• Custom Rec Flags:

  • Priority ICP: Enterprise organizations in regulated industries (finance, healthcare, government) with 500+ employees and active AI initiatives
  • Short Term Goals: Expand Prisma AIRS platform adoption; establish MLSecOps as industry standard; grow enterprise customer base

8. Data Sourcing Notes

• Other sources: https://protectai.com/guardian; https://protectai.com/newsroom; https://www.paloaltonetworks.com/prisma/prisma-ai-runtime-security