Please note: This is a non-binding courtesy translation. Only the German version is legally binding.
1. Our fundamental understanding of AI
For us, artificial intelligence is a tool — not a decision-maker. It takes recurring, time-consuming work off professionals and raises the quality of the data captured, but at no point does it replace professional responsibility. Three principles guide us: humans remain in control, data stays protected and the use of AI remains traceable. This statement is an expression of this third principle.
2. Where and what we use AI for
In our products, AI supports individual working steps in a targeted way. Specifically, we use AI models for:
- Speech to record — converting spoken notes from the construction site into structured, searchable text
- Structuring — arranging and categorising documentation data into the correct fields and trades
- Defect detection — suggestions for recognising and classifying defect characteristics from text and image
- Drafting and summarisation support — drafts for reports, records and inspection documentation
The aim is always to reduce manual rework and to increase the completeness of the documentation. The professional assessment and approval remain with the human (see Section 6).
3. AI services deployed and sub-processors
For the functions mentioned, we rely on specialised, established AI services. These act as processors exclusively on our instructions:
- Anthropic Claude (Anthropic PBC, USA) — text comprehension, structuring, image analysis and summaries
- OpenAI Whisper (OpenAI, USA) — speech transcription (speech-to-text)
- Groq (Groq Inc., USA) — accelerated speech transcription (Whisper) during site inspections
- OpenAI Realtime and Retell AI (USA) — voice dialogue in the support assistant
- Plaud.AI — import of recordings from external recorders, should you use this function
A website voice assistant (provider ElevenLabs, USA) is currently not integrated (Section 10); in any case it would not be part of the product processing.
We no longer use Google Gemini in our products; earlier statements to that effect are outdated.
The application itself is operated in the European Union and your data is stored there. For the AI processing, the respective necessary contents are transmitted to the providers named above; this includes a transfer to the USA. We base the transfer on the safeguards under Chapter V of the GDPR (General Data Protection Regulation). We make details as well as the respective current list of sub-processors available to our customers within the framework of the data processing agreement. Should we deploy further or different services in the future, we will update this overview.
4. Handling of your data — no training with your content
Your voice recordings, images and documents are processed exclusively to provide the contractually agreed service. It is contractually agreed with the AI providers deployed that your content is not used to train AI models. The AI services store transmitted content only to the extent and for as long as this is necessary for providing the service and for abuse control; permanent storage or any other use is contractually excluded. Purpose limitation applies — your content is not used for other purposes.
5. Data security and storage location
Your data is stored on servers within the European Union. For the AI processing, content is transmitted to the providers named in Section 3, predominantly to the USA. The transmission is encrypted throughout (TLS), and stored data is likewise stored in encrypted form. Access is limited to the necessary extent and logged. Details on retention, legal bases and your data subject rights can be found in our Privacy Policy.
6. Human oversight (human-in-the-loop)
All AI-generated content is a suggestion that is reviewed, edited and approved by human users. AI does not replace human professional expertise — it supports it. Decisions with legal or contractual effect are always made by a human. Without this approval, no AI result becomes effective as a binding document.
7. No automated decision in the individual case
No decision based solely on automated processing takes place that produces legal effects concerning you or similarly significantly affects you (Article 22 GDPR). The AI provides suggestions and drafts; the final decision always lies with a responsible human.
8. Classification under the EU AI Act
Regulation (EU) 2024/1689 (AI Act) is the governing instrument. Its transparency obligations under Article 50 have applied since 2 August 2026. Structify is the provider of the AI-supported functions in its own products and, at the same time, a deployer insofar as we use AI for our own purposes.
Our AI functions are designed as supporting tools under continuous human oversight. They do not make autonomously effective decisions and are not to be assigned to any of the high-risk fields of application listed in Annex III of the AI Act. An EU declaration of conformity under Article 47 is therefore not applicable — it is reserved for high-risk systems. We review this classification on an ongoing basis and adjust it if functions, purpose of use or the legal situation change materially.
We meet our transparency obligations by pointing out before every interaction with an AI assistant that it is an AI system, and by marking AI-generated content. We are progressively expanding the machine-readable marking of AI-generated content under Article 50(2); for systems that were on the market before 2 August 2026, the deadline for this runs until 2 December 2026.
9. Limits, accuracy and diligence
AI systems work probabilistically and, despite high quality, may deliver incorrect or incomplete results. AI outputs are therefore no substitute for a professional review and do not release you from your own duty of care. We design our functions so that results are easy to check and correct, and we point out the AI-supported nature within the products.
10. AI on this website
Voice assistant. A voice assistant ("Kai") is currently not integrated on this website. Should we reactivate it, we will inform you here in advance about the provider, data flows and the consent required.
Images on this website. Some of the images on this website were created with the help of artificial intelligence. This concerns three groups of motifs: the portraits and the joint image 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 technical control rooms and planning offices) as well as the large-format mood images between the text sections. The persons depicted are — with the exception of the founders — entirely fictitious; any resemblance to real persons would be coincidental. Not AI-generated are the screenshots of our applications (real interface with synthetic sample data) as well as the logos of the companies named.
We mark AI-generated images centrally in this statement; in addition, the alternative text of each affected image begins with the prefix "AI-generated image:". In this way we comply with the transparency obligations under Article 50 of Regulation (EU) 2024/1689 (AI Act). If the stock of images changes, we update this section. If you would like to know, for a particular image, whether and how it was created, please write to us at info@structify.solutions.
Indexing by third-party AI systems. This website may be indexed by AI systems (such as ChatGPT, Google Gemini, Claude or Grok) and drawn upon to answer user questions. All information contained here is publicly accessible and correct to the best of our knowledge. We regard this as a legitimate component of modern information provision.
11. Your rights and your objection
You can object to the processing of your data by AI components and can request access, rectification or erasure. Your further data subject rights under the GDPR are described in detail in our Privacy Policy. To exercise them, an informal email to info@structify.solutions is sufficient.
Irrespective of this, you can lodge a complaint concerning the AI Act with the competent market surveillance authority. In Germany, since the KI-Marktüberwachungs- und Innovationsförderungsgesetz (AI Market Surveillance and Innovation Promotion Act) entered into force on 29 July 2026, this is the Bundesnetzagentur (Federal Network Agency). For data protection complaints, the Bayerisches Landesamt für Datenschutzaufsicht (Bavarian State Office for Data Protection Supervision, BayLDA) is competent.
12. Contact for questions
If you have questions about how our AI components work, please write to us at info@structify.solutions. We will be glad to explain to you how a particular function works and which data is processed in the process.
Letzte Aktualisierung: September 2026