Artificial Intelligence Terms of Use
This is an English translation provided for convenience. The Greek version is the legally binding text and prevails in case of any discrepancy.
Version 2.0 · Effective from 6 November 2026
These terms supplement the Command Center Terms of Use and apply to every artificial intelligence ("AI") feature of the service. Such features include, among others, the assistant that suggests actions, the command bar, the categorisation of expenses and receipts, and draft texts. Terms not defined here have the meaning given to them in the Terms of Use. For AI Features, in the event of conflict these terms prevail; for personal data, the Data Processing Agreement prevails.
1. Definitions
1.1. "AI Features": the artificial intelligence or machine learning components of the service, together with the Models on which they are based.
1.2. "Model": a large language model or other artificial intelligence or machine learning model.
1.3. "Input Data": the data, information, prompts or material that you or your users submit to the AI Features, or that are drawn from Customer Data in order to serve a request.
1.4. "Output": the suggestions, answers, categorisations, drafts and any other material generated by the AI Features on the basis of the Input Data.
1.5. "Training": the use of data to create or improve a Model.
2. You always know it is AI
2.1. AI Features are labelled within the service. Every suggestion, answer or draft prepared by AI states so, in accordance with Article 50 of Regulation (EU) 2024/1689 on artificial intelligence.
2.2. AI Features are not a human being and do not replace human oversight.
3. Nothing is executed without your "Yes"
3.1. The AI assistant suggests. Every action, such as sending an email to a customer, assigning a task or creating a record, is executed only after the explicit approval of one of your users, with that user's permissions.
3.2. Before approval you see what will happen and you can correct it, for example the text of an email or the person to whom a task is assigned.
3.3. The AI does not take decisions with legal or similarly significant effects for individuals without a human decision (Article 22 GDPR). AI Features are not used for the evaluation of employees.
4. Who owns what
4.1. As between us, you retain all rights in the Input Data and the Output belongs to you. To the extent permitted by law, PR GANG LTD assigns to you any right it may acquire in the Output.
4.2. We process Input Data only to the extent necessary to provide the AI Features, in accordance with these terms and the Data Processing Agreement.
4.3. You represent that you and your users have the necessary rights and a lawful basis to submit the Input Data.
4.4. These terms do not limit our obligations under personal data protection legislation.
5. Your data and AI
5.1. Model provider. For each request, only the extract of data that is needed is sent to the Model provider (currently Anthropic, PBC). The provider acts as a sub-processor in accordance with the Data Processing Agreement, and its terms prohibit the use of your data for Training its models. For a change of provider, the notification procedure of the Data Processing Agreement applies.
5.2. No training. We do not train Models on the Input Data, the Output or any other data of yours, nor do we allow third parties to do so.
5.3. The assistant's memory. The assistant's "memory" for your company, i.e. what it knows about how your company operates, resides only in your company's database. It is visible on the "What the assistant knows" page and you can correct or delete it. It is never shared with other companies.
5.4. Improvement without training. To improve the AI Features we use only aggregated, anonymous statistics on how useful each type of suggestion is. For example: "payment reminders are accepted in 70% of cases". No information about you, your customers or any individual can be derived from them, and they are not used for Training.
6. What PR GANG LTD undertakes
6.1. To label every Output as a product of AI (section 2).
6.2. Not to execute any action without the explicit approval of one of your users (section 3).
6.3. To send to the Model provider only the necessary data and not to train Models on your data (section 5).
6.4. To keep the assistant's memory visible, correctable and isolated in your company's database.
6.5. To maintain the operation of the service when the usage limit is exhausted or when the AI Features are deactivated (sections 8 and 9).
7. What you undertake
7.1. Review. Output may be inaccurate, incomplete or unsuitable for your case. You review it before approving or using it.
7.2. Human oversight. In matters governed by specific legislation or professional rules, such as accounting, tax or employment matters, you do not base decisions on Output without appropriate human review.
7.3. Responsibility for approval. From the moment one of your users approves an action, it is deemed to be your action. The same applies to any text that is sent.
7.4. Honesty towards third parties. You do not falsely present Output as the work of a human where the law requires the use of AI to be disclosed.
8. Usage limits
AI Features are subject to a monthly usage limit depending on your plan. When the limit is reached, the core functions continue with simpler rules, without AI, until the next period or until the limit is increased.
9. Deactivation
The account owner may request the deactivation of the AI Features for their company. The remaining functions of Command Center continue as normal.
10. Prohibited uses
In addition to what is set out in the Terms of Use and the Acceptable Use Policy, the AI Features may not be used:
- for deception or for the creation of unlawful or defamatory content;
- for bulk sending of messages without a lawful basis;
- to circumvent the controls and approvals of the service;
- to infringe intellectual property rights or other rights of third parties;
- for the evaluation of employees or for decisions falling within Article 22 GDPR.
11. Disclaimers
11.1. Nature of AI. Due to the nature of artificial intelligence, Output may be incorrect or inaccurate. Output may also not be protected as intellectual property.
11.2. Similarity of Output. Output may resemble material generated for other users. We do not warrant that Output will not contain or reflect third-party material, that it will not infringe the rights of third parties, or that it will not be reproduced in the same or a similar manner for another user.
11.3. Not advice. Output does not constitute accounting, tax, legal or investment advice.
11.4. Liability. For the liability of the parties in relation to the AI Features, section 12 of the Terms of Use applies, including its exclusions.
Source and licence
This text is based on the Common Paper AI Addendum (Standard Terms, version 1.0) — https://commonpaper.com/standards/ai-addendum/ — which is made available under the CC BY 4.0 licence (https://creativecommons.org/licenses/by/4.0/). It was translated into Greek and adapted to EU law and to the service by PR GANG LTD; the changes are our own and Common Paper is not involved in and does not endorse this document.