Frequently Asked Questions

Q. Does Netcore use generative AI?

A. Yes. Generative AI is used across content generation, campaign analytics, segmentation recommendations, and planning assistance. All AI capabilities operate within defined security boundaries and governed workflows.

Q. What AI models does Netcore use?

A. Netcore currently uses Anthropic Claude, Google Gemini, and Amazon Titan (for embedding tasks via Amazon Bedrock). These models are selected based on specific use-case requirements. The list is subject to change as the AI landscape evolves.

Q. Will my data be used to train the AI model?

A. No. Customer data is never used to train, fine-tune, or update any AI model. Data is used only to generate a response for the current request.

Q. Can another customer's AI response include our data?

A. No. Every AI request is bound to a specific customer account via a unique identifier. Retrieval operations are scoped exclusively to that account, and the model has no memory between calls. Cross-customer data exposure is architecturally prevented.

Q. Is the sensitive business context permanently stored inside the model?

A. No. A sensitive business context is not permanently embedded into the model. Where context is required, it is retrieved at runtime using controlled retrieval patterns and scoped access controls.

Q. Can we view a history of AI interactions from our panel?

A. Yes. Authorised administrators can view conversation history within their account panel for up to 30 days.

Q. What personal data is protected from the AI model?

A. Email addresses, mobile numbers, names, national identity numbers, payment card data, and other personal identifiers are never sent to the AI model. The TRUST Layer's redaction control removes this information from prompts and responses automatically.

Q. How long is AI-related data stored?

A. Public best-practice reference data has no formal retention period and is not customer data. Customer attribute and taxonomy context used for retrieval is retained while the account is active and deleted upon churn. Campaign context is stored for up to 12 months and automatically purged. Conversation history is accessible for 30 days. Audit logs are retained per Netcore's platform data retention policy.

Q. How do you test AI systems for security?

A. Netcore runs a continuous AI security testing programme covering eight categories of adversarial scenarios: prompt injection, PII extraction, cross-customer data access, out-of-scope requests, persona manipulation, harmful content generation, infrastructure probing, and scope boundary enforcement.

Q. How can we conduct a security review?

A. Enterprise customers can request a security review through their Netcore account representative. This may include a technical architecture discussion, review of compliance documentation, and completion of a standard TPRM questionnaire.