Article 1 – Applicability
1. These supplementary terms and conditions apply to the use of AI features within Floriday, including current and future AI Agents made available to users by Floriday.
2. These terms and conditions apply in addition to the Floriday Terms and Conditions, the Floriday Privacy Policy and any other applicable platform terms and conditions. In the event of any conflict, specific provisions relating to the use of Floriday AI Agents shall take precedence over general provisions, insofar as they relate exclusively to the use of the AI Agents.
3. By activating or using an AI Agent, the user confirms that they have read and accept these terms and conditions.
Article 2 – Description of AI Agents
1. AI Agents within Floriday are digital support features that utilise artificial intelligence. Among other things, they can help users to analyse available data, answer questions, organise information, prepare drafts, complete or suggest data fields, highlight areas for attention, or prepare platform actions.
2. AI Agents are intended as a tool to assist the user. They do not replace the user’s own judgement, checks, authority or decision-making.
3. An AI Agent does not make any binding decisions independently, does not carry out any legal acts independently, and does not perform any platform actions with legal consequences without prior verification and approval by the user.
4. Floriday may modify the operation, functionalities, models used, model versions, prompts, guardrails, data links, authorisations and restrictions of AI Agents, for various reasons including security, quality, compliance, technical availability, further development or changes to legislation and regulations.
Article 3 – AI characteristics and limitations of output
1. AI Agents use artificial intelligence. The output may be incorrect, incomplete, out of date, inappropriate or otherwise flawed.
2. The output from an AI Agent is intended to support the user and does not replace the user’s own judgement, verification and decision-making.
3. Users must verify AI output before relying on it or using it for commercial, operational, legal, financial or administrative purposes.
4. When an AI Agent prepares data, concepts, transactions, changes or other actions, the user must check all relevant data before approving or using it. Depending on the functionality, this includes, amongst other things, the other party, product, price, quantities, delivery date, available stock, terms and conditions, amounts, addresses, documents, data selections and other relevant data.
5. The user must not use AI output where they have reasonable grounds to suspect that it is incorrect, incomplete, unauthorised, misleading, inappropriate or otherwise unsuitable.
Article 4 – User’s responsibility
1. The user is responsible for the questions, prompts, files, messages, data and other input that they provide to an AI Agent or have processed via an AI Agent.
2. The user is responsible for any action they take themselves following the use of an AI Agent, including amending, approving, confirming, submitting, initiating or using data, drafts or platform actions prepared by an AI Agent.
3. An order, amendment, agreement, confirmation or other transaction shall only be concluded at the time and subject to the conditions set out in the applicable Floriday platform terms and conditions. A proposal, draft, completed field or prepared process generated by an AI Agent does not, in itself, have any binding effect unless and until the user carries out the required action themselves.
4. The user must ensure that only authorised persons within their organisation use AI Agents and that internal mandates, rules on power of attorney, limits, authorisations and any four-eyes principles are complied with.
5. The user must not use AI Agents for unlawful, misleading, fraudulent, discriminatory, harmful, unauthorised or otherwise improper purposes.
Article 5 – Permitted input and restrictions on use
1. The user may only enter, upload or have processed information which they are authorised to use within Floriday.
2. The user must not enter any information that is not relevant to the use of the AI Agent in question or the processing of which is not permitted within Floriday.
3. The user must not use any prompts, files, instructions or technical means intended to circumvent security measures, authorisations, guardrails, model limitations, logging, monitoring or other controls.
4. The user must not use AI Agents to access other users’ data or to generate output that falls outside the scope of their own authorisations, access rights or permitted uses.
5. In the event of suspected misuse, security risks, unusual patterns, a breach of these terms and conditions, or a contravention of applicable laws and regulations, Floriday may take the measures set out in Article 8.2.
Article 6 – Data, privacy and confidentiality
1. AI Agents process data to the extent necessary for the operation of the relevant AI Agent within Floriday.
2. An AI Agent only uses data to which the user has access within Floriday or which the user themselves submits for processing, unless Floriday explicitly states otherwise for a specific AI Agent.
3. Floriday takes appropriate technical and organisational measures to prevent output or data from one user becoming accessible to another user.
4. Where personal data is processed, such processing shall take place in accordance with Floriday’s Privacy Statement and the relevant agreements between Floriday and the user.
5. User data will not be used to train underlying AI or LLM models provided by model providers, unless Floriday expressly states otherwise in advance and there is a valid legal basis and contractual foundation for doing so.
6. The user must refrain from entering or uploading unnecessary, excessive or unauthorised personal data, confidential information or third-party data.
Article 7 – Monitoring, logging and quality improvement
1. Floriday can monitor and log the use of AI Agents for security, fault diagnosis, quality control, abuse prevention, incident investigation, compliance and service improvement.
2. Logging may include, amongst other things, the AI Agent used, the model, the model version, technical calls, tool usage, the time, error messages, actions, authorisations, relevant system output and actions approved by the user.
3. Floriday uses monitoring and logging to investigate errors, incidents, unusual patterns, security risks and potential abuse scenarios.
4. Users can report suspected errors, incorrect output, incidents or undesirable behaviour by an AI Agent via Floriday’s usual support channels.
Article 8 – Availability, changes and suspension
1. Floriday endeavours to ensure that AI Agents operate reliably, but does not guarantee that AI Agents will always be available, error-free or suitable for any specific purpose of the user. To the extent permitted by mandatory law, Floriday shall not be liable for any damage, loss, costs or claims arising out of or in connection with output, orders, amendments, confirmations or other platform actions prepared by an AI Agent which have been checked, approved, confirmed or used by the user, including incorrect or incomplete order, price, quantity, product, delivery, availability or other party details, even if these were processed automatically in Floriday following approval. To the extent that Floriday is liable, notwithstanding the foregoing, such liability shall be limited in accordance with the Floriday Terms and Conditions and any other applicable platform terms and conditions. The user shall indemnify Floriday against any claims by third parties arising from actions authorised or confirmed by the user via an AI Agent, insofar as these result from incorrect input, insufficient supervision or unauthorised use by or on behalf of the user.
2. Floriday may (temporarily) restrict, block, suspend or disable AI Agents, input, output or usage due to maintenance, security, incidents, legal obligations, compliance requirements, risks of misuse or technical problems or circumstances referred to in Article 5.5.
3. Floriday may amend model versions, prompts, guardrails, data links, authorisations and functionalities where this is necessary or desirable for quality, security, compliance or further development.
Article 9 – Reporting of errors and incidents
1. Users must report any suspected errors, incorrect data, data breaches, unauthorised access, misuse, security incidents or other irregularities relating to the use of AI Agents as soon as possible via Floriday Support.
2. The user must not approve or use any data, concepts or actions generated or suggested by an AI Agent if they suspect that these are incorrect, incomplete, unauthorised, fraudulent or otherwise unsuitable.
3. Floriday may request further information in order to investigate a report or incident.
Article 10 – AI Act [Regulation (EU) 2024/1689] and user information
1. Floriday informs users in an appropriate manner that they are using AI functionality and that the output may be AI-generated.
2. For each AI Agent, Floriday provides information on its purpose, core functionality, key limitations, the role of human oversight, relevant data categories and how errors or incidents can be reported.
3. The user acknowledges that they must verify AI output and that AI Agents are intended as a support feature within Floriday.
4. If additional information, instructions or usage restrictions are required for a specific AI Agent, Floriday may include these in the interface, documentation, release notes, help centre information or supplementary terms and conditions.
