AI Data Policy

1) Scope & Roles

For your account/workspace data, Praisma is a data controller. For data you manage on behalf of your organization (e.g., posts, messages, analytics), you are the data controller and Praisma acts as your data processor. This AI Data Policy complements our Privacy Policy and Terms of Service.

2) Purpose of AI Features

• Generate or refine captions, hashtags, and campaign copy from prompts.
• Summarize conversations and draft replies for human review.
• Provide insights, recommendations, and scheduling assistance.
• Automate repetitive tasks (e.g., variations, tags) under your control.

3) Data Used by AI

• Content you submit (prompts, text, media metadata) and optional context you choose to include.
• Workspace data you explicitly target with an AI action (e.g., a thread to summarize, a draft to rewrite).
• Operational metadata necessary to deliver the feature (timestamps, model name, locale).

AI features do not read private messages, analytics, or external account data unless you initiate an AI action that explicitly involves those items.

4) Processing Locations & Providers

• Local / Self‑Hosted: Your workspace may use self‑hosted models (e.g., Ollama) running on your own infrastructure. Data stays within that environment.
• Cloud Inference: If enabled, we may use vetted AI API providers to process your inputs for inference only. Requests are transmitted securely and are not used to train provider models.

We use data minimization and encryption. Sub‑processors are bound by data processing agreements and may only act on our documented instructions.

5) Strict Use Restrictions

• Your data is never used to train third-party AI models.
• Prompts and outputs are not retained beyond delivering the feature.
• We do not sell or share AI inputs/outputs with third parties.