TrendAI Companion FAQ

What is TrendAI™ Companion?

TrendAI™ Companion is an AI-powered cybersecurity adviser designed to help security teams work faster and more effectively. Companion uses pre-trained large language models enhanced with TrendAI™ cybersecurity knowledge, threat intelligence, research, and expertise to provide contextual guidance for security operations.

Does Companion train AI models with user data?

No. Companion does not use customer data to train AI models. Companion uses pre-trained large language models enhanced with TrendAI™ cybersecurity knowledge and expertise rather than user-submitted data.

Which languages does Companion support?

Companion currently supports prompts in English, French, German, Japanese, Mandarin, Portuguese, and Spanish. Companion can also respond in a language different from the one used in the prompt.

Can an organization that has not enabled full functionality of TrendAI Vision One™ use Companion?

Yes. The data accessible to Companion is based on your organization’s currently enabled solutions and features. If the prompt requires access to data not available to your currently enabled solutions and features, Companion will offer the option to enable those relevant solutions and features.

How can a user submit feedback on Companion or Companion responses?

To submit feedback on Companion or Companion responses, go to Resource Center → Share your feedback.

Can Companion be integrated with an MCP server or an AI agent?

TrendAI Vision One™ has a public MCP server which enables integration between TrendAI Vision One™ and external AI tooling. For more information, see the GitHub documentation.

How does Companion reduce hallucinations or inaccurate answers?

Companion reduces hallucinations through the use of a cybersecurity-focused AI model, validation checkpoints, domain-specific prompting, and human supervision for higher-stakes decisions. Companion is also continually being improved to deliver more specific and accurate responses.

Are evidence and data sources given for Companion responses?

This feature is coming soon. In the near future, data sources and evidence will accompany Companion responses, including internal and external references where applicable. Responses will also include verification indicators, Companion confidence score, and session context to support audits and traceability.

Do users have the power to select or reject the remediation actions suggested by Companion?

Yes. Human supervision and validation is a core part of the design of Companion that helps preserve accountability and control.

Is any personally identifiable information (PII) sent to third-party LLMs?

No. TrendAI Vision One™ ensures the protection of sensitive information and PII through a combination of contractual and technical controls. Data processed by external AI components is governed by strict data processing agreements that explicitly prohibit the use of customer data for model training or any secondary purpose. These controls ensure that sensitive information remains protected throughout the AI processing pipeline and that all AI usage complies with privacy and data protection principles when interacting with external language models or AI services.

Where is Companion data stored?

Companion data is stored within your TrendAI Vision One™ instance using the same security protections applied to all other TrendAI Vision One™ data, including session history and audit logs.

For more information, see https://docs.trendmicro.com/en-us/documentation/article/trend-vision-one-data-privacy-security-compliance.

Is Companion user data protected in transit and at rest?

Yes. Companion user data is protected using the same security protections applied to all other TrendAI Vision One™ data, both in transit and at rest.

For more information, see https://docs.trendmicro.com/en-us/documentation/article/trend-vision-one-data-privacy-security-compliance.

Does user data stay in the same geographic region?

User data might be transferred outside the user's TrendAI Vision One™ region for processing in order to support data residency requirements.

For more information, see https://docs.trendmicro.com/en-us/documentation/article/trend-vision-one-data-privacy-security-compliance.

Who can access user prompts?

Access to user prompts is limited to authorized TrendAI™ personnel such as site reliability engineers, threat research and analytics teams, and customer support, with support access enabled only when explicitly allowed in the TrendAI Vision One™ console. Access is purpose-limited, recorded, and audited.

For more information, see https://docs.trendmicro.com/en-us/documentation/article/trend-vision-one-data-privacy-security-compliance.

How long are prompts stored?

Prompt session history is retained for up to 30 days. If a single session exceeds 500 prompts within 30 days, the earliest prompts beyond that threshold are deleted.

Can users review audit logs?

Yes. Master administrators can review and export audit logs. Audit logs for user actions are automatically generated and stored for 180 days.

How does TrendAI™ approach responsible AI?

The TrendAI™ responsible AI approach includes transparency, fairness, privacy, accountability, safety and security, human control, social responsibility, and continuous improvement. The approach also includes a commitment to compliance with the EU Artificial Intelligence Act, including risk-based management, transparency obligations, prohibited practice restrictions, requirements for high-risk AI, and post-market monitoring.