Klipt Studio
ENFRLire en françaisStart now

AI Transparency

Effective date: 8 October 2026

Klipt Studio presents itself to Clients as a content agency; behind that experience, artificial intelligence does most of the drafting and production work, under human supervision. This page explains how, in plain terms, and sets out our disclosures under Article 50 of EU Regulation 2024/1689 (the "AI Act"), applicable since 2 August 2026.

1. Where AI is used

  • Planning: a weekly content plan (topics, formats, timing) is drafted by AI from your brand profile, past performance, and business updates you share, then made available for your review.
  • Production: captions, scripts, and media (images, video, voice-over) for each piece are produced with the assistance of AI models and media-generation providers (see Sub-processors).
  • Quality checks: automated checks run before anything reaches you (format/ratio per network, language, brand colors, banned words, caption length, media rights flags, and whether an AI-disclosure label is required).
  • Portal conversations: replies in the message bar ("Tell us what to change") are generated by AI, which interprets your intent (approval, change request, business update, question, complaint) and acts accordingly, escalating anything sensitive to a human.

2. Human supervision

A member of our team can read, override, or take over any conversation or piece of content at any time. Content that is ambiguous, high-risk (e.g., a complaint, a legal question, a cancellation request, or anything our systems are not confident about) is escalated to a human rather than answered automatically. Nothing is published without passing through the review/autopilot flow your account is configured with (see Terms of Service Section 5).

3. Disclosure in the chat (AI Act Article 50(1))

The client portal clearly labels the "Studio" conversation and every content thread as AI-assisted. We do not present our AI as a human team member; replies are signed "the Klipt Studio team" and the portal UI carries a persistent, visible indicator (e.g., a badge or inline notice in the conversation header) disclosing that messages may be generated or drafted with AI assistance, placed where a reasonable user will see it before relying on a reply — in practice, the chat shows an "AI-assisted" label next to every reply.

4. Labelling of AI-generated media (AI Act Article 50(4))

Where a piece of published media is a realistic, AI-generated or AI-manipulated image, audio, or video that could appear to depict real people, places, or events, we mark it as artificially generated or manipulated using the labelling mechanism offered by the destination platform (for example, the is_ai_generated/AI-content flag passed to our publishing provider), in addition to any disclosure the platform itself applies. This labelling obligation is assessed per piece as part of our automated quality checks before submission for your review.

5. No training on your data

We do not use your Brand Information, Client Media, content, or conversations to train our own or any third party's general-purpose AI models. Our AI and media sub-processors' commercial terms (not their free consumer terms) exclude using API/commercial inputs for model training — including for our use of Anthropic's models, which is under commercial API access, not a personal subscription; see our Sub-processors page for all providers involved and their respective policy basis.

6. Your controls

You can request a human review of any AI-drafted reply or piece of content at any time, switch between review-required and autopilot modes in Settings, and ask us questions about how a specific piece was produced through the "Why this post" note shown with each item.

7. Contact

Questions about our use of AI: hello@klipt-ai.com.