Last updated: August 2026
1.Chapter I: Introduction and Scope
Article (1) Purpose
This Policy explains how the ZynDesk Platform uses artificial intelligence technologies within the optional "AI Module," the controls governing data processing and the outputs of that Module, and supplements the Privacy Policy and the Data Processing Agreement without replacing them.
Article (2) Feature Scope
The AI Module includes, by way of example and not limitation: automated classification of documents and expenses, smart suggestions for following up on client tax deadlines, and predictive analytics on firm performance.
2.Chapter II: Nature of Outputs and User Responsibility
Article (3) Decision Support Outputs Only
All outputs of the AI Module are indicative decision-support tools only and do not constitute tax, accounting, or legal advice, nor do they relieve the User of the obligation to review and verify their accuracy before relying on them in any tax filing, financial decision, or client communication. The User remains — in accordance with Article (9) of the Acceptable Use Policy — solely responsible for the accuracy of any tax or financial data submitted in reliance, in whole or in part, on such outputs.
Article (4) Accuracy Limitations
AI Module outputs may contain errors or biases arising from the nature of the models used or the quality of the data entered. The Company does not warrant that outputs will be entirely free of errors.
3.Chapter III: Use of Data in Model Training
Article (5) No Use of Customer Data to Train General Models
The Company shall not use accounting firm client data (CRM data, financial and tax records) to train general artificial intelligence models used to serve other users, except with the prior explicit written consent of the Data Controller (the User).
Article (6) AI Model Providers
Certain functions of the AI Module may rely on models or services provided by specialist third-party providers, rather than models developed entirely in-house. Where the AI Module relies on models provided by a third party, that third party shall be treated as a sub-processor under the Data Processing Agreement and shall be listed in the Sub-processor Schedule upon actual activation of the relevant feature. The obligation set out in Article (5) above — not to use accounting firm client data to train general artificial intelligence models except with prior explicit written consent — remains in full force regardless of the identity of the model provider used, and the Company undertakes not to engage any provider that does not offer equivalent contractual guarantees in this respect.
4.Chapter IV: Human Oversight and Transparency
Article (7) Mandatory Human Oversight
Features that affect financial or tax decisions shall be designed to require explicit human confirmation before any final action is executed, such as sending a client notification or finalising an expense classification.
Article (8) Disclosure of AI Use
Outputs generated by the AI Module shall be clearly labelled within the Platform interface to distinguish them from manual User inputs.
5.Chapter V: Final Provisions
Article (9) Relationship to Other Documents and Policy Updates
This Policy shall be read together with the Privacy Policy and the Data Processing Agreement, and shall be updated periodically to reflect the evolution of AI features on the Platform and relevant regulatory requirements. Users shall be notified of any material change at least 15 days before it takes effect.