Overview

Netcore's AI capabilities are designed to help marketers work faster, create smarter campaigns, and make better decisions across the customer lifecycle. AI is built into the Netcore experience to improve speed, relevance, and efficiency across content generation, campaign performance analysis, segmentation, and planning assistance.

To support enterprise use, Netcore follows a security-first, responsible AI approach. AI-powered workflows are built with safeguards for:

• Data privacy
• Tenant isolation
• Prompt and output safety
• Controlled access to business context
• Auditability and monitoring

Core Security Commitments

Every customer should understand the following non-negotiable principles that govern how Netcore handles AI.

AI Is Used in Netcore

Netcore uses generative AI to power the following capabilities within the platform:

• Campaign content generation
• Performance insights and analytics
• Audience segmentation recommendations
• Campaign planning assistance

These features are built as governed workflows, not open-ended AI tools. Every AI interaction is tied to a specific task, operates within defined boundaries, and follows the same access controls as the rest of the platform.

AI Models in Use

Netcore uses externally hosted foundation models from approved providers, accessed through secure cloud APIs.

📘

Note

Model selection may vary by feature and deployment environment. The list of models used is subject to change as the AI landscape evolves and new models are evaluated.

Netcore does not fine-tune any of these models using customer data. Business context is provided at the time of each request through a controlled retrieval process, keeping the model itself free of any customer-specific information.

Data Handling and Privacy

Netcore is deliberate about what information flows into its AI pipeline. There are two categories of data that may be used:

Publicly Available Reference Information

Channel best practices sourced from platform providers such as Meta, Google, and Yahoo are stored in Netcore's retrieval layer to help the AI produce relevant, platform-compliant recommendations.

• It is refreshed periodically.
• It does not represent customer data.
• It is not subject to per-customer data retention policies.

Tenant-Specific Campaign Context

Aggregated performance summaries, such as campaign delivery rates and engagement metrics, scoped exclusively to the requesting customer's account. This is operational metadata, not individual-level data.

• This context is refreshed once every 24 hours to ensure AI responses remain current.
• Short-term conversational context is retained within a session to support follow-up questions.
• This is cleared when the session ends and is never shared across users or customers.

Data Excluded from AI Processing

The following are explicitly excluded from any AI interaction:

• Individual-level personal data like, email addresses, mobile numbers, names, national identity numbers, or any other personal identifiers tied to end customers.
• Raw contact lists, CRM records, or unmasked behavioural event data.
• Payment card data or financial information.
• Credentials, API keys, or infrastructure information.

👍

Good to Know

Technical controls are in place to prevent this data from reaching the AI model even if it is inadvertently included in a user's prompt.

AI Model Training and Customer Data

Netcore does not contribute any customer data, including prompts, campaign information, or audience context to train or update any foundation model.

The API agreements with Netcore's model providers enforce zero data retention: prompts and model responses are transient by design and are not stored or used for model training by the provider.

In the case of Amazon Bedrock specifically, standard inference API calls are fully transient, with no persistence of input or output data

Data Retention Schedule

Refer to the table below to understand how different categories of AI-related data are retained, managed, and automatically purged across Netcore’s governed AI workflows.

📘

Important Point to Remember

Customers with specific data retention requirements can discuss custom schedules as part of their enterprise agreement.

Customer Data Isolation

Netcore is a multi-tenant platform. Ensuring that one customer's data cannot be accessed by or exposed to another customer is a foundational requirement for AI as much as for any other part of the platform.

• Every record in Netcore's databases is partitioned by a unique customer identifier.
• Every AI-related query, including retrieval operations that provide context to the AI model, must include this identifier.
• A query that does not carry a valid customer identifier is rejected before execution.

This makes cross-customer data access structurally impossible, not merely policy-restricted. The AI models themselves are stateless between requests; they retain no memory across calls, which provides an additional layer of isolation at the model level.

Trust Layer

All AI activity in Netcore passes through a centralised control layer, the TRUST Layer that sits between the platform and the external AI model. Passing through the TRUST Layer is mandatory for every AI request and response. It enforces privacy, safety, and governance at every step.

Continuous Security Testing

Netcore tests its AI systems on an ongoing basis using a structured security testing framework. Rather than waiting for issues to surface in production, Netcore proactively evaluates AI agents against scenarios designed to expose weaknesses before they can be exploited. The current test coverage spans eight categories of adversarial scenarios:

• Prompt injection attempts
• Personal data extraction attempts
• Cross-customer data access attempts
• Out-of-scope usage
• Role and persona manipulation
• Harmful content generation
• Infrastructure probing
• Scope boundary enforcement

Each test is evaluated on a pass/fail basis for security scenarios, with detailed reporting available for enterprise governance reviews. Test suites are regularly updated as new adversarial techniques emerge.

Responsible AI Principles

Security is one part of how Netcore thinks about responsible AI. More broadly, Netcore is committed to building AI that is transparent, fair, and governed.

Every AI request in Netcore follows a controlled, multi-stage flow before a response is returned to the user. No AI feature in Netcore operates as an unrestricted system — each request goes through validation, access control, and policy enforcement at every stage.

Encryption in Transit

All data in transit between Netcore's application layer and the AI model infrastructure is encrypted using TLS 1.2 or higher.