UNYX

Legal

Contractual documents: only the French version is authoritative.

AI Charter: transparency and use of artificial intelligence

Last updated: September 2026

Preamble

UNYX, a single-member private limited company (EURL) with a share capital of EUR 1, registered with the RNE under SIREN number 100 489 004, whose registered office is at 60 rue François Ier, 75008 Paris, France, represented by its Manager Michaël Haim Chiche (hereinafter "UNYX"). Contact: [email protected].

UNYX uses artificial intelligence in its products and in its services. This charter explains in plain language what UNYX does with AI, what it refuses to do, and what you are entitled to. It supplements the Legal notice, the Privacy policy and the General terms and conditions; on any point specific to an offer, the special conditions of that offer prevail.

It implements Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence (the "AI Act"), whose obligations apply in stages: prohibited practices and AI literacy since 2 February 2025, rules applying to general-purpose AI models since 2 August 2025, and general obligations including the transparency duties of Article 50 since 2 August 2026.

1. UNYX's roles under the Regulation

1.1. UNYX is a provider of the AI systems it develops and makes available under its own name: digital audit, Léo (Google Ads agent), expert AI agents, Leest, UNYX Archétypes, LiveMonitoring, the AI agent by email and the assistance built into the back office.

1.2. UNYX is a deployer of the general-purpose AI models published by third parties, which it uses under its own authority to run those systems. UNYX trains no model and places no general-purpose AI model on the market.

1.3. You become a deployer in turn when you operate a system delivered by UNYX towards your own users: a generated site featuring a conversational agent, an AI agent connected to your mailbox, a campaign run from your account. The obligations that then fall to you are set out in Article 11.

2. The AI systems offered by UNYX

Digital audit. Automated analysis of a publicly accessible website (technical health, visibility, conversion, compliance) and production of a scored report. It decides nothing on your behalf and never modifies the site being analysed.

Léo, the Google Ads agent. Reads the Google Ads accounts entrusted to UNYX, analyses campaigns, search terms and budgets, writes recommendations and prepares actions. Sensitive actions are not applied automatically: they are prepared, submitted for human validation, logged and reversible (Article 6).

Expert AI agents. Conversational agents specialised by profession, producing content and deliverables on request. They are decision aids: you check them before use.

AI agent by email and Léo on WhatsApp. Handling of requests received by messaging. Every exchange states that it comes from an automated agent.

Leest. Augmented to-do list: rephrasing, prioritising and breaking down tasks, with routing across several models.

UNYX Archétypes. Website generation: copy, images, layout. The content produced is flagged as AI-generated and reviewed before publication.

LiveMonitoring. Continuous monitoring of a website and qualification of the incidents detected, so you are warned without delay.

Assistance and contextual help in the back office, image generation for illustrations, and synthetic voice for the audio presentations of the offers.

None of these systems analyses emotions, recognises faces, or takes any decision producing legal effects concerning you.

3. The models used, and what happens to your content

3.1. Model providers. UNYX relies on general-purpose AI models published by Anthropic (Claude models) and OpenAI (GPT models and image generation models), used through their professional interfaces. This list is kept up to date here and in the Privacy policy.

3.2. No training on your content. Content sent to these models as part of the UNYX services is not used to train or improve those models: the professional terms subscribed by UNYX with these providers exclude it.

3.3. Transfers outside the European Union. These processing operations may involve a transfer of data to third countries, framed by the safeguards provided for in the GDPR (standard contractual clauses in particular), under the conditions described in Article 6 of the Privacy policy.

3.4. Logging. Exchanges with the AI systems are logged for security, traceability and the handling of complaints, for the retention periods stated in the Privacy policy.

4. You always know when you are talking to an AI

Under Article 50(1) of the Regulation, every exchange with a UNYX conversational system (AI agent chat, agent by email, WhatsApp agent, back-office assistance) is flagged as such from the first message.

A human contact remains reachable at any time: +33 9 60 05 73 66 or [email protected]. You may ask at any point for a person to take over your request.

5. AI-generated content is labelled

5.1. Marking. Content produced or assisted by AI and distributed by UNYX (text, images, pages of generated websites, synthetic voice) is marked in a way that is legible to the person viewing it, and machine-readable wherever the format allows (Article 50(2) and 50(4)).

5.2. The Journal. The articles published in the Journal are prepared with the help of AI, then reviewed, corrected and published under the editorial responsibility of UNYX, which stands behind their content.

5.3. Customer reviews. The reviews published on the Reviews page come from real people. They are never generated, rewritten or rephrased by an AI.

5.4. Reports. Audit reports, performance reviews and other automatically produced documents state that they come from automated processing.

6. Human oversight, safeguards and reversibility

6.1. No UNYX system carries out an irreversible action on your accounts or your data without human validation. Sensitive actions are prepared, presented, and applied only once validated.

6.2. Actions carried out on your accounts are logged and traceable. You may at any time ask for automated processing to stop, withdraw an access right, or return to fully human management.

6.3. You remain the owner of your accounts, your data and your assets. UNYX keeps no access once an assignment ends.

6.4. An AI can be wrong. The output of UNYX systems consists of decision aids, to be checked by you before use. It does not constitute legal, accounting, tax or medical advice.

7. What UNYX rules out

UNYX neither develops, uses nor makes available any system falling under the practices prohibited by Article 5 of the Regulation, in particular:

  • subliminal, manipulative or deceptive techniques that distort behaviour;
  • the exploitation of vulnerabilities linked to age, disability or social situation;
  • social scoring of individuals;
  • predicting the risk of a person committing an offence on the sole basis of their profile;
  • the untargeted scraping of facial images to build facial recognition databases;
  • emotion recognition in the workplace or in education;
  • biometric categorisation revealing sensitive data;
  • remote biometric identification in publicly accessible spaces.

Nor does UNYX use AI to produce fake reviews or fake testimonials, or to imitate the identity of a real person or organisation.

8. No high-risk system

As at the date of this charter, no system provided or deployed by UNYX falls under the high-risk uses listed in Annex III to the Regulation. In particular, UNYX offers no system intended for screening job applications or managing human resources, assessing learners, granting credit or pricing insurance, accessing essential public services, biometrics, critical infrastructure, justice or migration.

Should a high-risk system ever be offered, it would be declared here, together with its documentation and its obligations, before being put into service.

9. Your personal data and AI

9.1. The processing of personal data associated with these systems is described in the Privacy policy, which sets out the purposes, legal bases, recipients and retention periods.

9.2. No fully automated decisions. In accordance with Article 22 of the GDPR, no decision producing legal effects concerning you, or significantly affecting you, is taken solely on the basis of automated processing. The scores and qualifications produced by UNYX systems (audit score, qualification of a request) are indications subject to human judgement.

9.3. Please do not send sensitive data (health, opinions, banking details) or information covered by a duty of confidentiality in exchanges with a conversational agent.

10. AI literacy

Under Article 4 of the Regulation, UNYX ensures that the people who operate these systems within the company have a sufficient level of AI literacy: capabilities, limits, risk of error and the right verification reflexes. Each module comes with a factsheet stating what it reads and what it never modifies. Training in the professional use of AI is offered by UNYX Academy.

11. Your obligations if you deploy a system delivered by UNYX

Where you operate a system delivered by UNYX towards your own users, you become its deployer and it falls to you:

  • to inform your users that they are dealing with an AI system;
  • to label the AI-generated content you publish;
  • to ensure human oversight and a sufficient level of AI literacy among the people concerned (Articles 4 and 26);
  • to use the system in accordance with the instructions provided by UNYX, and not to divert it towards a prohibited or high-risk purpose. A substantial modification of the system or of its intended purpose may make you a provider within the meaning of Article 25, with the corresponding obligations.

12. Question, incident or complaint

For any question about UNYX's use of AI, to report problematic content or to contest the result of automated processing: [email protected], subject line "AI". UNYX replies within thirty (30) days.

You may also refer the matter to the competent market surveillance authority for artificial intelligence and, as regards your personal data, to the CNIL, the French data protection authority (3 place de Fontenoy, TSA 80715, 75334 Paris Cedex 07, cnil.fr).

13. Changes to this charter

The AI Act continues to phase in until 2027 for systems embedded in already regulated products. This charter is updated whenever the applicable framework or the systems offered by UNYX change; the date at the top of the page is authoritative.