Terms and conditions of BMD's AI services

Information on data protection, data processing and areas of application in the context of the AI services provided by BMD Systemhaus GesmbH

This document provides information on data processing and other aspects relating to the artificial intelligence (AI) tools offered by BMD Systemhaus GesmbH.

Last updated: July 7, 2026

1.1. General information on data processing

Depending on the AI tool and its configuration in BMD NTCS, the AI solutions we offer may use different AI models and different technical infrastructures. The specific technical setup of the AI tools—particularly with regard to the providers used, processing locations and potential data transfers—depends on the respective configuration in BMD NTCS.

If you use your own AI resources or configurations, you are responsible for ensuring that data processing complies with your own requirements.

The overview below provides general information about the providers and subproviders used and the associated legal framework for data protection.

Overview of the providers and subproviders used

The technical integration of individual AI models may be provided through the service provider Cortecs GmbH (1090 Vienna). According to Cortecs GmbH, the transmitted data are not used to train AI models. For further details, please refer to the information provided by Cortecs GmbH at https://cortecs.ai/dpa.

1.2. Data processing outside the EU/EEA

The AI tools we provide by default are generally operated within the EU or the EEA.

However, depending on the respective configuration in BMD NTCS, AI models or providers that process data outside the EU or the EEA may also be used. Any processing of data outside the EU or the EEA is based on appropriate safeguards in accordance with Chapter V of the GDPR (e.g. the EU Standard Contractual Clauses or adequacy decisions, where applicable).

1.3. Processing of inputs and outputs

The data entered into the AI tools are processed by BMD solely for the purpose of providing the respective functionality, in particular to perform AI inference. We do not use the input data to train AI models.

Please note that the cloud or model providers used may retain prompts and generated outputs for a limited period of time. Such retention is typically carried out solely for the purposes of abuse monitoring or auditing to detect unauthorised or abusive use and may include automated and, where necessary, manual reviews.

Where a provider has enabled “Zero Data Retention”, such retention does not occur. Zero Data Retention refers to a configuration in which input data (prompts) and generated outputs are not retained for logging, monitoring, auditing or training purposes after processing.

The scope, duration and specific nature of such processing depend on the provider used and the respective configuration in BMD NTCS.

1.4. Processing of metadata

Regardless of the respective AI tool, BMD stores certain metadata (such as customer number, user ID, timestamps and consumed tokens) for billing purposes on Microsoft Azure servers within the European Union.

1.5. Legal notices

The following provisions specify key legal and usage-related aspects in connection with the use of the AI tools and their outputs.

1.5.1. Responsibility for data processing

The decision as to which data you actively enter and process within the AI tools is made by you, taking into account your internal policies as well as the applicable legal and contractual requirements.

Where AI functions are run in the background within the software and data is processed automatically (e.g. to support functionalities or generate outputs), the categories of data processed in this context are defined by BMD as part of the system configuration and design. In such cases, you remain responsible for ensuring that the use of the respective functions and the underlying data complies with your internal policies and legal requirements.

In particular, you are responsible for ensuring compliance with the GDPR, the EU AI Act and other statutory, regulatory or contractual obligations.

BMD does not provide legal advice in this context.

1.5.2. Reliability and verification of AI-generated content

AI tools are based on large language models, which may produce incorrect, incomplete or biased outputs. Since this is a technical limitation of these models, BMD cannot entirely prevent it. Accordingly, the reliability of the generated content cannot be guaranteed. AI-generated content must always be reviewed for accuracy, completeness and suitability before use.

1.5.3. Liability for AI-generated content

BMD is not liable for any inaccuracies, omissions or outdated information provided by AI tools.

2. BMD’s AI tools

The following descriptions contain service-specific information on the individual AI tools, in particular regarding the types of data processed in each case and any possible deviations from the general conditions described above. The general conditions set out in Section 1 apply to all AI tools by BMD unless otherwise specified.

2.1. BMD Fox

BMD Fox is a chatbot that allows our customers to ask questions concerning the functionality and use of the BMD applications. The responses are based on the Help system of our software.

What you can use BMD Fox for:

• Questions about the functionalities of the BMD software
• Questions about how to use the BMD software
• Specific questions about how to implement particular business cases in the BMD software
   

What you should not use BMD Fox for (not supported or outside its intended purpose):

• Questions, instructions and processing that are unrelated to the BMD software, its functionalities or its use
• As a basis for decision-making
• Legal or tax-related matters
• Questions about how to correctly post business cases in accordance with the statutory regulations
• Entering personal data
• Attempting to bypass built-in security measures or using the tool for unlawful purposes

Details on data processing

When an input is sent to BMD Fox, the data necessary to provide the functionality are processed.

This includes in particular:

• Customer number (the user ID is not transmitted)
  • Storing your customer number is required to assign specific questions to a chat and to analyse and improve BMD Fox.
• Country of operation
  • This is required in order to select the appropriate BMD Help system for your country.
• Entered question or text
  • It is used to identify the relevant contexts within the BMD Help. Your query, along with the relevant elements of the Help system, is then transmitted to a large language model (LLM) for inference (i.e. to generate a response).

Different large language models may be used for inference based on the input and the elements of the Help system (which means the customer number and the country of operation are not transmitted). These models may be operated on servers located outside the EU. The cloud providers used may also retain your input and the generated output, either within or outside the EU, in connection with the use of the AI tools. Details regarding the scope, duration and nature of such retention are provided in Section 1.3 .

2.1.1. Feedback on BMD Fox

When you interact with BMD Fox, you have the option to give feedback to BMD on its responses by clicking on the “thumbs up” or “thumbs down” icon. We use this feedback internally at BMD to improve BMD Fox. You can choose to send the feedback anonymously, with your BMD customer number or with your name and customer number. Please note that you will not receive a reply to your feedback. This function is not intended for contacting our support team.

2.2. BMD Fox Text

BMD Fox Text can be used to process various kinds of text and essentially comprises the following text processing functions: summarising, proofreading, generating and translating texts as well as interactively editing documents or web pages using BMD Doc Chat. In addition, there are several specific ways how Fox Text can be integrated into the BMD software.

What you can use BMD Fox Text for:

• Summarising emails and documents
• Proofreading promotional materials
• Generating invitations and letters
• Integration features (such as creating subjects for tasks, translating text or interacting with documents using Doc Chat)

What you should not use BMD Fox Text for (not supported or outside its intended purpose):

• Processing of texts or documents that violate legal regulations and any processing used for unlawful purposes
• Processing of texts or documents for which you do not hold the necessary rights (e.g. copyright)
• Use cases that are considered prohibited or high-risk under the AI Act
• Summarising or assessing job applications, CVs and other documents related to recruiting (we cannot guarantee that the responses are always fair and free of bias)

Details on data processing

In particular, the following data are processed when using BMD Fox Text:

• Customer number (used for billing)
• User ID (used for billing)
• Country of operation (used for statistics)
• Entered or selected text and the associated document for processing (used for ensuring functionality)
  • When you use Fox Text in combination with Doc Chat to chat with an AI model about PDF documents or to ask questions about their content, the text is extracted from the documents and used for processing. The document itself is not made available to the AI model.
• Additional instructions (used for ensuring functionality)
• Fox Text function in use (e.g. summarising), including a style guide, if available, and the AI model to be used (used for ensuring functionality)

2.3. Fox Filter

Fox Filter can be used to filter BMD NTCS tables comfortably with natural language. For this purpose, the AI creates a filter based on your natural language input.

Details on data processing

It is not possible to change the models used for Fox Filter.

In particular, the following data are processed when using Fox Filter:

• Customer number (used for billing)

• User ID (used for billing)
• Country of operation (used for statistics)
• The entered text or description of the filter to be created in natural language (used for ensuring functionality); only the metadata of the table (e.g. the field names) are used for processing. The content displayed on the screen itself is not transmitted to the AI.

2.4. Fox CV

Fox CV allows you to extract data from CVs and to automatically import them into BMD NTCS. The AI methods used simply extract information from CVs in a structured format; they do not evaluate, analyse, select, filter or rank candidates.

Details on data processing

When you extract information from CVs with Fox CV, the following data are transmitted to servers and processed:

• Customer number (used for billing)
• User ID (used for billing)
• Country of operation (used for statistics)
• The CV or the text of the document that the AI processes (used for ensuring functionality)

2.5. RechtGPT

Integration of RechtGPT (https://rechtgpt.at/) is available in BMD NTCS (provided that you have the required valid licence). RechtGPT is an AI-based tool provided by Prime AI GmbH, based in 2500 Baden, Austria, which is designed to assist with legal research and is directly integrated into tools such as BMD Fox. Additionally, there are further integrations in the form of additional convenience features and integration points. In all cases of using RechtGPT in the BMD software—regardless of how it is accessed—Prime AI GmbH alone is responsible for the content of the responses. Use of RechtGPT within the BMD software is also subject to Prime AI GmbH’s terms of use (https://rechtgpt.at/#terms). The tool is intended for use by legally trained professionals. Please also note that AI-generated responses may contain errors. BMD is not liable for any inaccuracies, omissions or outdated information in the results generated by or through RechtGPT.

Details on data processing

Data entered in the BMD software in connection with RechtGPT are processed by both Prime AI GmbH (operator of RechtGPT) and BMD.

For information on how Prime AI GmbH processes data and any potential subproviders involved, please refer to the RechtGPT privacy policy (https://rechtgpt.at/#privacy)).

BMD stores the questions you enter and the responses generated by RechtGPT, along with your customer number, user ID and metadata (such as timestamps), on Microsoft Azure servers located within the EU. These data are used for billing purposes and for manual quality improvements. BMD does not use your data to train AI models. Your interactions with RechtGPT are saved in the BMD Fox history for each user, but they may also be accessible to other persons within your organisation—for example, when archived in connection with a specific tax case.

2.5.1. Feedback to RechtGPT

When you interact with RechtGPT, you have the option to give feedback to RechtGPT on its responses by clicking on the “thumbs up” or “thumbs down” icon. This feedback is processed and stored by BMD, before it is transmitted to Prime AI GmbH for the purpose of enabling them to improve their system. Any feedback transmitted to Prime AI GmbH is anonymous, provided that the query and the feedback itself do not contain any personal information. Please note that you will not receive a reply to your feedback. This function is not intended for contacting any of our support teams.

2.6. Fox Analytics

BMD Fox Analytics is a tool used to analyse and evaluate various types of data. This tool is available in screen views, reports and evaluations, and—depending on the use case—provides either dedicated analyses and evaluations recommended by BMD or a general standard evaluation. You can also customise and control the analyses and evaluations yourself; the necessary instructions can be centrally stored in style guides, if required.

What you can use Fox Analytics for:

• Analysis of asset lists
• Analysis of irregularities in open item lists

What you should not use Fox Analytics for (not supported or outside its intended purpose):

• Processing of texts or data that violate legal regulations and any processing used for unlawful purposes
• Processing of texts or data for which you do not hold the necessary rights, including rights required for AI-based processing (e.g. copyright)
• Use cases that are considered prohibited or high-risk under the AI Act

Details on data processing

By default, AI models that process data within the EU or the EEA are used. The specific nature of the data processing depends on the respective settings in BMD NTCS. Depending on the configuration, additional options such as Zero Data Retention may be available.

The following data are processed when using BMD Fox Analytics:

• Customer number (used for billing)
• User ID (used for billing)
• Country of operation (used for statistics)
• Uploaded or selected data (used for ensuring functionality)
• Additional instructions (used for ensuring functionality)
• Fox Text function in use including a style guide, if available, and the AI model to be used (used for ensuring functionality)

Note: When using Fox Analytics, the underlying data are processed as described above and transmitted to the relevant AI models and providers or subproviders in order to provide the functionality. It is therefore recommended to reduce the volume of data transmitted wherever possible. For example, any unnecessary columns can be hidden temporarily in screen views prior to using Fox Analytics.

2.7. Speech-to-Text

Within BMD NTCS, you can enter text with your voice by selecting an input field and clicking on the microphone icon. You can use this function both for entering short texts and creating long transcripts.

Details on data processing:

To perform automatic transcription, the recorded speech input is transmitted to the configured AI infrastructure. By default, AI models for which, according to the provider, Zero Data Retention is applied are used.

The following data are processed when using voice typing:

• Customer number (used for billing)
• User ID (used for billing)
• Country of operation (used for statistics)
• Speech input as an audio file for transcription (used for ensuring functionality)