What Data Does Agentic AI Process?
Email Data
When emails are submitted for analysis, Agentic AI processes:
| Data Category | Examples | Why We Need It |
|---|---|---|
| Email Headers | Sender address, recipient, subject line, reply-to | |
| Email Content | Message body | To analyze for threats, phishing indicators, and malicious content |
| URLs | Links in the email | To check for malicious websites and phishing pages |
| Attachment Names | File names attached to emails | To identify potentially dangerous file types |
How Does Agentic AI utilize AI?
Agentic AI uses Large Language Models (LLMs) for intelligent threat analysis. The following email data is sent to LLM providers for analysis:
- Email headers (sender, recipient, subject, etc.)
- Email body content
- URLs contained in the email
- Attachment filenames
AI Processing Activities
- Threat Classification: Categorizing emails as potential threats, spam, or legitimate
- Deep Analysis: Secondary review for cases requiring deeper analysis
- Threat Explanation: AI-generated summaries explaining why an email was flagged
Your email content is not used to train shared AI models.
Subprocessors
Agentic AI uses the following subprocessors to deliver its services. The applicable infrastructure provider is determined by the specific Trend Micro product.
| Subprocessor | Purpose | Product |
|---|---|---|
| Amazon Web Services, Inc. | Hosting and provision of the Agentic AI Service | Cloud Email Gateway Protection |
| Microsoft Corporation (Azure) | Hosting and provision of the Agentic AI Service | Cloud Email and Collaboration Protection |
| Microsoft Corporation (Azure) OpenAI | Artificial intelligence-enabled functionality for threat analysis | All Products |
Cloud Email Gateway Protection and Cloud Email and Collaboration Protection utilize Agentic AI with outlook plug-in feature.
AI Processing Activities
| What We Process | Purpose | Shared With |
|---|---|---|
| Email headers | Threat detection | Microsoft Corporation (Azure) OpenAI |
| Email body content | Security analysis | |
| URLs | Link scanning | |
| Attachment filenames | File threat detection |