🔍💡What “AI Security” Means in R82.20

[JonathanR]

Need a quick overview of R82.20's new features of AI Security?

If your company is adopting AI in a much faster pace than you can keep up, which might be the norm, you can leverage these 3 new features in R82.20.

New Features in R82.20

Capability	Protected subject	Primary risk addressed	Policy / operating plane
Workforce AI Security	Employees using supported public GenAI applications	Sensitive information submitted in prompts or uploaded files; ungoverned AI use	Access Control + Content Awareness, Application Control, URL Filtering, HTTPS Inspection
AI Agent Security	Internally developed AI applications using supported external LLM APIs	Prompt injection and unsafe/non-compliant prompt content; release positioning also references MCP tool-response protection	Threat Prevention profile + AI Guardrails project/API key; HTTPS Inspection
AI Auditor (formerly Policy Auditor)	Access Control policy and segmentation intent	Rules that violate documented allow/block guidelines or create inconsistent policy outcomes	SmartConsole management analysis; segment matrix and guideline calculation

How R82.20 Elevates an Existing Check Point Environment

Existing strength	Incremental R82.20 value
Firewall and Application Control	Recognizes supported AI destinations and adds AI-specific policy intent instead of treating GenAI as ordinary web traffic.
Content Awareness / DLP concepts	Applies sensitive-data classification to AI prompts and uploads, with AI-oriented data types and user messaging.
Threat Prevention profiles	Adds prompt-injection protection for supported application-to-LLM API flows and ties enforcement to familiar profile criteria.
HTTPS Inspection	Uses established decryption placement to inspect encrypted AI interactions, while increasing the importance of certificate governance and bypass review.
SmartConsole policy management	Adds a segmentation-intent matrix, violation workflow, scheduled recalculation, Global Policy support, and improved audit trail.
Logging and governance	Adds AI-specific logs, Portal analytics, dashboards, and auditable decisions for exceptions and policy changes.

Licensing, Cost, and Blade Alignment

Capability	Technical alignment	Cost conclusion
Workforce AI Security	Part of the Content Awareness blade; also requires URL Filtering, Application Control, HTTPS Inspection, DLP-Next cloud analysis, and Portal activation	Requires Workforce AI Security license
AI Agent Security	Threat Prevention blade/profile plus Check Point Portal AI Guardrails, Guard API key, Project ID, HTTPS Inspection	Requires AI Guardrails license
AI Auditor	Security Management / SmartConsole policy-analysis function	Requires the AI Auditor license

I've attached an infographic that shows these 3 features in more detail. Enjoy🍴

 

I hope this helps!