AI changes the security situation in two directions: your own AI use needs protection and rules, and attackers use AI for better phishing and faster attacks. With our vendors' solutions we cover both.
| Challenge | What helps | Vendors (selection) |
|---|---|---|
| Shadow AI: who uses which AI services? | Detect and assess AI applications on the network and block unapproved ones | Fortinet, Microsoft, WatchGuard, F5 |
| Company data leaks through prompts | Detect and stop sensitive content in AI inputs (DLP) | Microsoft Purview, Fortinet, ESET, Eye Security |
| AI-supported phishing and prompt injection by email | AI-supported email security, checking links and attachments | Hornetsecurity, NoSpamProxy, Eye Security |
| Protecting your own AI apps, LLM interfaces and agents | Gateways and guardrails against prompt injection, data leakage and uncontrolled agent actions | F5, Fortinet, ESET, Microsoft |
| Detecting and responding to attacks faster | AI-supported detection and analysis in XDR, SIEM and MDR | ESET, WatchGuard, Microsoft, Enginsight, Eye Security, Logsign |
| If something does happen | Immutable backups with anomaly and malware detection | NovaStor, Hornetsecurity |
AI assistant for analysis in ESET PROTECT and the MDR services. New in the portfolio: protection for AI agents, behavioural monitoring of agents and checks of AI conversations for sensitive data. Frankfurt data centre available.
FortiGate detects and controls generative AI applications via application control, web filtering and DLP. FortiAIGate protects your own AI applications as a gateway with guardrails against prompt injection and data leakage.
Security Copilot supports analysis and response. Purview DSPM for AI detects sensitive data in AI use, Defender for Cloud Apps assesses and blocks AI apps, Entra reveals shadow AI and gives AI agents their own identities.
AI Gateway and AI Guardrails protect your own AI applications, models and MCP connections against prompt injection, jailbreaks and data leakage. Can also be run on premises.
Firebox makes SaaS and AI use visible and assesses application risks. The AI assistant Rai supports detection, investigation and response.
AI-supported email security analyst for reported mails, link checks in Microsoft Teams and features that reveal permission risks before Copilot is rolled out.
The AI-based 32Guards service assesses attachments and links based on email metadata from German-speaking countries. Developed in Germany.
Prevents data from being pasted into private AI accounts and enforces data protection rules at prompt level. Plus managed detection and response operated in Europe.
SIEM with machine-learning-based anomaly detection and AI-supported actions. Developed in Germany and available on premises and via IONOS.
All information according to vendor websites and documentation, October 2026. Features and availability depend on edition and licence.
Not every security solution needs AI. These building blocks ensure that only authorised people and devices access your AI applications and data, and that you can recover in an emergency.
Zero trust network access as a VPN replacement: identity-based access with device posture checks. Developed and operated in Germany.
Network access control: only known devices get network access. Made in Germany.
The German cloud for your AI platform, with the AI Model Hub and data centres in Germany.
Immutable backups with anomaly and malware detection, developed in Hamburg.
Email and collaboration with a selectable EU data centre. If you want to use AI features in the mailbox, we define together which data may go there.
SIEM, SOAR and UEBA in one platform, with machine learning for behavioural analytics.
Together we check which AI services are used, where data can leak and which of your existing solutions already help.
An IT problem you would like taken off your plate?
30-min IT consultation