AI Transparency Statement

Structify Solutions UG (haftungsbeschränkt)
Version 2.0 · valid from 1 November 2026

The German version prevails.

This statement describes where and how we use artificial intelligence (AI) in our products and on our website. It supplements our Privacy Policy and fulfils our transparency obligations under Article 50 of Regulation (EU) 2024/1689 (AI Act).

1. Our principle: AI supports, people decide

For us, AI is a tool. It takes repetitive work off professionals' hands, such as typing up voice notes or pre-sorting defects, and makes documentation more complete. It does not replace professional responsibility. Three rules always apply:

• People stay in control. All AI results are suggestions that are reviewed, edited and approved. The only exception is automatic notifications in Structify Operate that are triggered according to fixed rules; whether they are triggered is not decided by the AI (section 6).
• Data stays protected. Hosting and storage take place in Germany; AI providers may not use your content for training.
• Use of AI stays transparent. We tell you when you are working with AI, and we label AI results.

2. Where we use AI

3. Which providers we use

If we use additional or different providers, we will update this overview. For Customers, the procedure for new sub-processors under the Data Processing Agreement also applies.

4. What happens to your content

• Purpose limitation: We transfer to the providers only the content required for the respective feature, and only when you use the feature.
• No training by the providers: We have contractually agreed with all AI providers that they will not use your content to train AI models.
• No training by Structify without consent: We ourselves do not train AI models with our Customers' content unless the Customer's management expressly switches this on for its organisation. The setting is off by default; as long as we do not offer it, no training takes place.
• Retention by the providers: The providers store content only insofar as necessary for providing the service and for abuse monitoring, and only for the contractually limited period. On our side, we delete analysis photos from the AI photo analysis after 90 days.

5. Storage location and transfer

Hosting and storage in Germany (EU). For AI analyses (voice, text, images), we use US providers (OpenAI, Anthropic; for the voice assistant, Retell AI) on the basis of data processing agreements and EU standard contractual clauses.

Transmission is encrypted throughout. For each provider, we have assessed the consequences of the transfer and defined supplementary safeguards. Details can be found in the Core Section of the Privacy Policy (G7) and in Schedule 3 to the Data Processing Agreement.

6. Classification under the AI Act

• Role: Structify is the provider of the AI-supported features in its own products and a deployer insofar as we use AI for our own purposes.

• No high-risk classification: The AI features in Structify Web, Desktop and Mobile and on our website are supporting tools: their results are suggestions or information, and people decide how to use them. They do not make decisions that take effect by themselves and do not control any installations. We therefore do not assign them to the high-risk areas under Annex III to the AI Act; an EU declaration of conformity is not required for them. For Structify Operate, the next point applies.

• Structify Operate: Structify Operate is used in the operation of heat and energy supply installations. Its AI features are not intended to be used as a safety component in the operation of critical infrastructure, in particular the supply of heating, gas or electricity (Annex III No. 2 to the AI Act). We therefore do not classify them as a high-risk AI system:

  • The AI features do not issue control commands, do not intervene in the operation of the installations and have no access to control or automation systems.
  • They are not intended for safety tasks. Structify Operate is not a safety, control or alarm system. Under our Terms of Use, the operator uses it only if the protective and safety devices of its installations and the prescribed monitoring work independently of Structify Operate, so that a failure or error of Structify Operate endangers neither the installations nor persons (Terms of Use § 11(3)).
  • They document, analyse, calculate and recommend. Recommendations, for example on orders or maintenance, are implemented only after an authorised person has reviewed and approved them.
  • Automatic notifications: Structify Operate triggers some notices without prior approval, according to fixed rules and not by a decision of the AI. If a recorded measured value exceeds a set threshold, Structify Operate creates a ticket; if a fault remains unprocessed for a defined period, it marks the ticket as „eskaliert“ ("escalated"). In both cases, it notifies the persons entered for this purpose in the application and by e-mail. The AI only drafts explanations and the text of the notification. These notifications supplement the operator's monitoring; they do not replace it.

  We will review this classification again before we extend Structify Operate, in particular before any feature would control installations or take over safety tasks.

• We review this classification if features, the intended purpose or the legal situation change materially.

7. Labelling

The transparency obligations under Article 50 of the AI Act have applied since 2 August 2026. We fulfil them as follows:

• Notice before interaction: Before you speak or write to an AI assistant, we point out that it is an AI system.
• Labelling of results: AI-generated texts and suggestions bear the notice „Von KI erstellt“ ("Created by AI") in the application.
• Machine-readable labelling: AI-generated reports in PDF format receive a machine-readable indication of origin (IPTC metadata and a provenance record signed by us). We are extending the labelling to further outputs; for features that were on the market before 2 August 2026, we will complete this by 2 December 2026.
• Notice before transfer: Before any of your content is transferred to AI services for the first time, we show a one-time notice, uniformly in Structify Web, Structify Desktop (macOS and Windows), Structify Mobile and Structify Operate. By ticking a checkbox you confirm: „Ich habe die Hinweise zur Übermittlung an KI-Dienste gelesen und zur Kenntnis genommen.“ ("I have read and taken note of the information on the transfer to AI services."). You then click „Zustimmen“ ("Agree"). In doing so, you confirm that you have taken note. The legal basis for the transfer is your company's Data Processing Agreement, not consent under Article 6(1)(a) GDPR.

8. AI on our website

Images. Some of the images on structify.solutions have been generated with AI: the portraits and the joint picture of the founders Tim Rosen and Felix König (re-enacted scenes, not photographs), the industry motifs on the product and solution tiles (people on construction sites, in plant rooms and planning offices) and the large-format mood images between sections of text. With the exception of the founders, the persons shown are fictitious. The screenshots of our applications (real user interface with sample data) and the logos of the companies mentioned are not AI-generated. The alternative text of each AI-generated image begins with „KI-erzeugtes Bild:“ ("AI-generated image:").

Voice assistant "Kai". If we offer the voice assistant "Kai" on the website, we point out before the conversation begins that you are speaking to an AI system; the conversation begins only after you have given your consent. Details can be found in the Privacy Policy (M4.6).

Use by third-party AI systems. Our website may be read by third-party AI systems, such as search assistants, and used to answer questions. All information on the website is public.

9. Limits and due care

AI systems work with probabilities. Despite high quality, they may deliver incorrect or incomplete results. AI results do not replace a professional review and do not release anyone from exercising their own due care. We design our features so that results are easy to check and correct. There is no decision based solely on automated processing that produces legal or similarly significant effects (Article 22 GDPR).

10. Your rights, objection and complaints

• You can revoke your confirmation regarding the transfer to AI services in the settings, in Structify Mobile under Profil → Rechtliches & KI-Transparenz (Profile → Legal & AI Transparency); the notice will then be shown again before the next transfer.
• Your rights under the GDPR, including the right to object, are described in the Privacy Policy (G10). Please send enquiries to datenschutz@structify.solutions. If they concern a Customer's content, that Customer is your point of contact; we will forward your enquiry.
• You can address complaints concerning the AI Act to the competent market surveillance authority; in Germany, this is the Federal Network Agency (Bundesnetzagentur). The Bavarian State Office for Data Protection Supervision (Bayerisches Landesamt für Datenschutzaufsicht) is responsible for complaints under data protection law.

11. Questions

If you would like to know how a particular AI feature works or what data it processes, please write to us at datenschutz@structify.solutions.