AI Use & Output Notice

Last updated: 10 May 2026

This Notice describes how Accwisely uses automated and machine-learning technologies in the operation of the Service, and the position you should adopt in respect of any output produced by such technologies.

1. Use of automated processing

The Service uses automated processing, including rule-based mapping and, where applicable, machine-learning techniques, to translate trial balance input into a draft financial statements format. Specific processing steps include:

• Mapping of uploaded chart-of-accounts entries to standard financial statement line items based on a master mapping table maintained by us.
• Selection of disclosure notes based on the presence of specific account categories in the input.
• Population of templated wording, including accounting policies and disclosure notes, with values derived from the input.
• Arithmetic computation of totals, subtotals, and reconciliation figures.

2. Limitations of automated processing

Automated processing is, by its nature, mechanical. The Service does not exercise professional judgement. In particular:

• Where account names in your input deviate from common naming conventions, the mapping may be incorrect or incomplete.
• Where your input contains errors, those errors will be carried into the Generated Output without correction.
• Judgement-driven figures (useful lives, expected credit losses, fair values, going-concern assumptions) are populated based on user input or default assumptions and are not independently assessed by the Service.
• The Service does not detect fraud, irregularities, or material misstatements in the input.
• Standards, regulations, and disclosure conventions evolve over time. While we update our templates and mappings periodically, the Generated Output may not reflect the most current position as at the date of generation.

3. Your obligation to review

You must review all Generated Output before issuing, signing, filing, or otherwise relying on it. Review by a qualified accountant or tax agent is strongly recommended. The Generated Output is a draft to support, not replace, professional review.

4. Use of input data for model improvement

We do not use your User Content to train external machine-learning models. We may use de-identified, aggregated patterns derived from usage of the Service to improve our internal mappings and templates. This processing does not involve disclosure of identifiable User Content to any third party.

5. Right to refuse output

We reserve the right to refuse to generate Output where, in our reasonable judgement, the input data is incomplete, malformed, or appears intended to produce misleading financial statements.