About Glean Protect and Protect+
Glean Protect provides the security foundation for every Glean deployment. Your data lives in a fully isolated cloud with strictly enforced permissions and zero LLM data retention.
Glean Protect+ is an add-on to Protect. It adds continuous sensitive content scanning across all connected sources and AI security guardrails that detect and block harmful content at every stage of an LLM interaction.
Glean Protect+ is a separately licensed add-on to Protect. Contact your Glean account team to get started.
Capabilities at a glance
| Capability | Glean Protect | Glean Protect+ |
|---|---|---|
| Sensitive content detection | One-time CSV reports | Weekly and continuous policies across 100+ sources |
| Findings review | CSV export | Customizable dashboards, CSV, and API |
| Remediation | Manual via CSV uploads | Automated enforcement within AI surfaces |
| Severity analysis | Not included | Sensitive content models with classifiers Beta |
| AI security guardrails | Not included | Prompt injection, malicious code, and toxic content protection |
| Agent alignment | Not included | Alignment models that check actions before execution Beta |
| Restricted topics | Not included | Policy to monitor AI usage of sensitive topics Beta |
| Partner integrations | Not included | Open ecosystem (Palo Alto Networks, Tines, and SIEM/SOAR) |
Benefits of Protect and Protect+
Move from one-time to continuous protection
Instead of relying on one-time scans, Protect+ continuously monitors indexed content and AI interactions in near real time, automatically hiding sensitive documents and blocking unsafe AI requests as they occur.
Keep your data safe
Detect and remediate overshared or externally exposed sensitive content before it becomes a risk in AI-powered search and agents. Glean’s models start protecting your data right away with out-of-the-box policies.
Make security an enabler, not a blocker
Teams can deploy agents faster with AI security guardrails in Glean Protect+. Automated enforcement keeps Glean secure while users continue working.
Unified security across data, AI, and agents
Protect your entire AI ecosystem with a single platform, from indexed documents to AI assistant interactions to autonomous agent workflows.
Reduce alert fatigue
Sensitive content models separate real risks from noise, classifying findings by severity and applying enhanced analysis to reduce false positives while preserving recall on actual threats.
Help meet regulatory requirements
Maintain compliance with data protection regulations through automated scanning, detailed audit trails, customizable policies, and integration with your existing governance workflows.
Private-by-design platform
Every Glean deployment includes these foundational security capabilities:
- Single-tenant architecture across clouds (AWS, Azure, GCP)
- Glean-hosted or customer-hosted isolated cloud options
- Strict permissions enforcement of data sources
- Zero LLM data retention: no training of LLMs on enterprise data
- Strong single sign-on (SSO) and role-based access controls (RBAC)
- Encryption in transit and at rest
- Detailed audit logs
- Compliance certifications, including ISO 42001, SOC 2 Type II, and ISO 27001
- Regional data residency
Data security
Glean Protect and Protect+ give you layered controls over what content is indexed, scanned, and surfaced in AI responses.
Glean Protect
- Customizable inclusion and exclusion indexing rules to control what content gets indexed from each data source
- Sensitive content protection with one-time CSV reports for point-in-time audits
- Manual remediation via CSV uploads to hide sensitive content from Glean
Glean Protect+
- Weekly and continuous sensitive findings across 100+ data sources
- Out-of-the-box sensitive findings policies with support for custom configurations
- Customizable dashboards based on sensitive findings policies
- Automated remediation of overshared sensitive content within AI surfaces
- Sensitive content models via classifiers for automated severity analysis Beta
- Triage workflows to review, assign, and resolve findings Beta
- Open partner ecosystem, including Palo Alto Networks and Tines
Safeguard AI agents
Glean Protect and Protect+ control how agents are managed, what data they can access, and what guardrails apply at runtime.
Glean Protect
- Permissions for agent management: control who can create, edit, view, or share agents
- User-level permissions enforcement: agents follow user permissions, so they only see data and take actions the user already has access to
Glean Protect+
- Prevention against prompt injection, jailbreak attempts, malicious code, and toxic content
- Agent alignment models that check actions before they execute Beta
- Restricted topics policy to monitor AI usage of sensitive topics Beta
Feature availability
| Capability | Glean Protect | Glean Protect+ |
|---|---|---|
| Private-by-design platform | ||
| Single-tenant architecture across clouds (AWS, Azure, GCP) | ✓ | ✓ |
| Glean-hosted or customer-hosted isolated cloud options | ✓ | ✓ |
| Strict permissions enforcement of data sources | ✓ | ✓ |
| Zero LLM data retention; no training of LLMs on enterprise data | ✓ | ✓ |
| Strong single sign-on (SSO) and role-based access controls (RBAC) | ✓ | ✓ |
| Encryption in transit and at rest | ✓ | ✓ |
| Detailed audit logs | ✓ | ✓ |
| Compliance certifications, including ISO 42001, SOC 2 Type II, and ISO 27001 | ✓ | ✓ |
| Regional data residency | ✓ | ✓ |
| Data security | ||
| Customizable inclusion and exclusion indexing rules | ✓ | ✓ |
| Sensitive content protection with one-time CSV reports | ✓ | ✓ |
| Manual remediation via CSV uploads | ✓ | ✓ |
| Weekly and continuous sensitive findings across 100+ data sources | ✓ | |
| Out-of-the-box sensitive findings policies | ✓ | |
| Customizable dashboards based on sensitive findings policies | ✓ | |
| Automated remediation of overshared sensitive content within AI surfaces | ✓ | |
| Sensitive content models via classifiers for automated severity analysis Beta | ✓ | |
| AI security triage workflows Beta | ✓ | |
| Open partner ecosystem, including Palo Alto Networks Beta and Tines | ✓ | |
| Safeguard AI agents | ||
| Permissions for agent management | ✓ | ✓ |
| User-level permissions enforcement | ✓ | ✓ |
| Prevention against prompt injection, jailbreak attempts, malicious code, and toxic content | ✓ | |
| Agent alignment models Beta | ✓ | |
| Restricted topics policy Beta | ✓ |