1. Overview
KappaX integrates AI capabilities across two core areas of the platform:
- Creative AI — a conversational AI copilot within the KappaX Editor that assists users in planning and building interactive ad creatives, including layout generation, design suggestions, interaction auto-configuration, and asset optimization
- Analytics AI — AI-powered analysis of interactive campaign data to surface performance insights, identify engagement patterns, and generate optimization recommendations
This policy governs how AI is used within the platform, how data flows through AI systems, and the responsibilities of all parties involved.
AI features are powered by third-party large language model (LLM) providers — including Anthropic, OpenAI, and Google. Hyperstate Technologies may also develop and deploy proprietary AI models, subject to the controls described below.
2. Data Handling
Creative AI: When invoked, the following data may be sent to AI providers: user prompts, project metadata (name, type), and design context (element structure, styling, text content, asset URLs).
Analytics AI: When invoked, aggregated and anonymized campaign interaction data (clicks, navigation events, quiz responses, completion rates, geographic and demographic aggregates) may be sent to AI providers. No raw personal data or individual-level identifiers are included.
AI features do not process personally identifiable information (PII) or data from other customers' workspaces.
3. Data Retention and Model Training
KappaX may retain AI conversation data for service delivery, usage metering, and model improvement.
Enterprise customers: AI conversation data is never used for model training or improvement. Enterprise data is retained only for service delivery and deleted per the enterprise agreement.
Non-enterprise customers: Anonymized AI conversation data (stripped of workspace and user identifiers) may be used to train and improve proprietary models, as disclosed in the Terms of Usage.
No personal data is used for model training under any circumstances.
4. Human Oversight and User Responsibility
All AI features require explicit user initiation. For complex operations, the AI proposes a plan that must be approved before execution. All outputs can be reviewed, modified, or undone. AI cannot publish or deploy creatives — that is always a separate user action.
The user is solely responsible for all content created or generated using AI features. This includes verifying accuracy, legal compliance, brand safety, and appropriateness before publishing. KappaX provides AI as a productivity tool and does not warrant the accuracy, completeness, or legal compliance of AI-generated outputs.
5. Transparency
Users are clearly informed when they are interacting with AI. The AI copilot operates exclusively within the Editor during the creative design phase — it does not interact with end consumers who view published ads. AI assists in creating the creative; it is not embedded in the delivered creative.
6. Governance
This policy is reviewed annually or upon introduction of new AI capabilities, changes to provider terms, or changes in applicable regulations. Owned by the Head of Product in coordination with the Information Security Officer.
