Praxis
AI in Practice: SK Stiftung Kultur Köln works with the neuland.ai HUB
AI in Practice: SK Stiftung Kultur Köln works with the neuland.ai HUB
Karl Heinz Land
·
6
Min. Lesezeit
aufsatz
Bild: KI generiert mit neuland.ai HUB
The SK Stiftung Kultur of Sparkasse KölnBonn is deploying the neuland.ai HUB in two of its most prominent institutions: the Photographic Collection and the German Dance Archive Cologne.
Project Report: Semantic Indexing and Time-Based Annotation in Cologne
Initial Situation and Objective
The Photographic Collection and the German Dance Archive Cologne exemplarily illustrate the challenges of large, heterogeneous collections: historical photographs, correspondence, programmes, reviews, videos and audio recordings. The objective of the SK Stiftung Kultur is to make these holdings quicker to index, richer in context to search and permanently more accessible – with clear professional responsibility and traceable processes.
Solution: the neuland.ai HUB
The AI management and orchestration platform supports the institutions in two complementary areas:
Photographic Collection: automatic annotations for images, suggestions for motives, locations and time periods, semantic and visual search, curated approval processes with a roles and rights concept.
German Dance Archive Cologne: time-based annotations for video and audio material; segmentation of scenes and sequences; linking performances with programmes, reviews, photographs and biographies.
The neuland.ai HUB combines ontology and knowledge-graph functions with modern search, quality assurance and governance. All changes go through documented workflows; decisions remain auditable, and sovereignty lies with the specialist teams. Operation is autonomous – on-premises or in a data protection-compliant cloud – and can be integrated into existing specialist IT systems.
Impact in Practice
Faster initial description: Teams verify qualified AI suggestions instead of starting from scratch.
More accurate research: semantic and visual search reveals hidden connections.
Deeper context: AV holdings are indexed with high precision; sources can be linked via timestamps.
Lasting quality: consistent metadata, clearly documented releases, compatibility with authority data.
"The aim of the SK Stiftung Kultur with an internal AI is to make the collections of both areas, the Photographic Collection and the German Dance Archive Cologne, visible as a whole. The visual and semantic search makes series, motif sequences and unexpected relationships visible. Investigating this would have previously required an enormous amount of time. Even at this stage, however, we are also getting a feel for the opportunities that AI will create for us in the future: motives and locations are to be recognised, database information enriched. For the first time, we can draw cross-connections between both collections, which remained invisible in separate systems. It is a different way of looking at what we have."
— Norbert Minwegen, Managing Director of the SK Stiftung Kultur
Outlook
The modular architecture allows step-by-step expansion from pilot to scale – for example, for additional work groups, exhibition and educational projects, teaching and research collaborations. Governance, data sovereignty and interoperability remain the guardrails for sustainable use.
Karl‑Heinz Land: "AI pre-organises – humans curate"
Archives, museums and collections are places where time speaks. Their work thrives on diligence, context and an eye for connection. In every era, there have been tools that made this categorisation easier – from the scriptorium to the card index to the database. Artificial intelligence is the next step: a cultural technique that takes care of routine tasks and makes connections visible.
What does this mean in practice? AI recognises patterns in images and texts, suggests plausible motives, locations and time periods, structures audio and video along scenes and sequences, and helps to identify duplicates as well as standardise terms. The division of labour is key: the machine makes suggestions, the human decides. In this structure, responsibility remains where it belongs – with the specialist teams who verify, weight and approve.
Perhaps the most powerful impact of AI is felt in research. Semantic and visual search query holdings by meaning, not just by keywords. This opens up spaces in which images and documents find each other that were previously separated: series emerge, cross-connections become visible and stories become more complete. Speed turns into diligence because the path to a reliable initial description is shorter, leaving more time for curatorial and research work.
Reputable AI in the cultural sector requires sovereignty and transparency. That means: roles and rights concepts, audit trails and documented approvals, on-premises operation or a strictly data protection-compliant cloud. When a system makes a suggestion, it must be comprehensible why. This is the only way to build trust – both internally and externally.
What is the vision? An archive, a museum, a collection that are as accessible as a good exhibition: clearly structured, deeply researched, intuitive to explore. The technology works quietly in the background, the human remains at the centre. AI pre-organises – humans curate. Exactly therein lies the power to preserve cultural heritage and at the same time retell it in a lively way.
— Karl‑Heinz Land, Founder and CEO of neuland.ai AG
Interview with Norbert Minwegen, Managing Director of the SK Stiftung Kultur
Mr Minwegen, why did the SK Stiftung Kultur decide on AI-supported indexing?
Our institutions work with large, historically grown collections. We want to make them quicker to index and richer in context to access – for our specialist teams, for research and for the public. AI helps us to speed up routine processes without relinquishing responsibility. This strengthens quality and creates free space for what matters most: interpreting, communicating, curating.
What changes specifically in the Photographic Collection?
The platform suggests motives, locations and time periods and complements title variations. This is a real relief in the initial description. At the same time, semantic and visual search improve research quality. An image becomes searchable not only via a keyword, but also via similarities and contexts. Our specialists verify, correct and approve – every decision is documented.
And what does AI mean for the German Dance Archive Cologne?
Movement needs timestamps. We segment video and audio material into scenes and sequences and link these with programmes, reviews, photographs or biographies. This allows performances to be indexed and researched with pinpoint accuracy. This saves time, increases reliability and creates new avenues for education and research.
You emphasised that cross-connections between collections become visible. Can you elaborate on that?
Gladly. Our objective with an internal AI is to make the collections of both areas – Photographic Collection and German Dance Archive Cologne – visible as a whole. The visual and semantic search makes series, motif sequences and unexpected relationships visible. Investigating this would have previously required an enormous amount of time. Already now we can see what opportunities are opening up: motives and locations are recognised, info enriched. For the first time, we can draw cross-connections between both collections, which remained invisible in separate systems. It is a different way of looking at what we have.
How do you ensure data sovereignty and quality?
We work with clear roles and rights concepts, documented approvals and auditable workflows. Operation is autonomous – on our own infrastructure or in a data protection-compliant cloud. AI suggestions are always traceable and are verified by specialists. This builds trust and ensures that speed does not come at the expense of diligence.
What impact do you expect in the medium term for research, education and the public?
Faster, more accurate searches and richer contexts. Researchers discover connections that were previously hidden. For education, new narrative pathways emerge, and the public benefits from clearer access – both digitally and on-site. In the end, cultural memory benefits because more material becomes visible and usable.
Your look ahead: what are the next steps?
We will gradually expand the proven workflows and connect further collections. The close involvement of the teams remains crucial – training, feedback, continuous quality assurance. Our benchmark is simple: AI assists, humans decide. This is what we are aligning our next steps with.
If you would like to learn more about the platform or gain an insight into the workflows, please write to us at info@neuland.ai. We will show you the neuland.ai HUB in action – sovereign, auditable and tailored perfectly for archives, museums and collections.
The SK Stiftung Kultur of Sparkasse KölnBonn is deploying the neuland.ai HUB in two of its most prominent institutions: the Photographic Collection and the German Dance Archive Cologne.
Project Report: Semantic Indexing and Time-Based Annotation in Cologne
Initial Situation and Objective
The Photographic Collection and the German Dance Archive Cologne exemplarily illustrate the challenges of large, heterogeneous collections: historical photographs, correspondence, programmes, reviews, videos and audio recordings. The objective of the SK Stiftung Kultur is to make these holdings quicker to index, richer in context to search and permanently more accessible – with clear professional responsibility and traceable processes.
Solution: the neuland.ai HUB
The AI management and orchestration platform supports the institutions in two complementary areas:
Photographic Collection: automatic annotations for images, suggestions for motives, locations and time periods, semantic and visual search, curated approval processes with a roles and rights concept.
German Dance Archive Cologne: time-based annotations for video and audio material; segmentation of scenes and sequences; linking performances with programmes, reviews, photographs and biographies.
The neuland.ai HUB combines ontology and knowledge-graph functions with modern search, quality assurance and governance. All changes go through documented workflows; decisions remain auditable, and sovereignty lies with the specialist teams. Operation is autonomous – on-premises or in a data protection-compliant cloud – and can be integrated into existing specialist IT systems.
Impact in Practice
Faster initial description: Teams verify qualified AI suggestions instead of starting from scratch.
More accurate research: semantic and visual search reveals hidden connections.
Deeper context: AV holdings are indexed with high precision; sources can be linked via timestamps.
Lasting quality: consistent metadata, clearly documented releases, compatibility with authority data.
"The aim of the SK Stiftung Kultur with an internal AI is to make the collections of both areas, the Photographic Collection and the German Dance Archive Cologne, visible as a whole. The visual and semantic search makes series, motif sequences and unexpected relationships visible. Investigating this would have previously required an enormous amount of time. Even at this stage, however, we are also getting a feel for the opportunities that AI will create for us in the future: motives and locations are to be recognised, database information enriched. For the first time, we can draw cross-connections between both collections, which remained invisible in separate systems. It is a different way of looking at what we have."
— Norbert Minwegen, Managing Director of the SK Stiftung Kultur
Outlook
The modular architecture allows step-by-step expansion from pilot to scale – for example, for additional work groups, exhibition and educational projects, teaching and research collaborations. Governance, data sovereignty and interoperability remain the guardrails for sustainable use.
Karl‑Heinz Land: "AI pre-organises – humans curate"
Archives, museums and collections are places where time speaks. Their work thrives on diligence, context and an eye for connection. In every era, there have been tools that made this categorisation easier – from the scriptorium to the card index to the database. Artificial intelligence is the next step: a cultural technique that takes care of routine tasks and makes connections visible.
What does this mean in practice? AI recognises patterns in images and texts, suggests plausible motives, locations and time periods, structures audio and video along scenes and sequences, and helps to identify duplicates as well as standardise terms. The division of labour is key: the machine makes suggestions, the human decides. In this structure, responsibility remains where it belongs – with the specialist teams who verify, weight and approve.
Perhaps the most powerful impact of AI is felt in research. Semantic and visual search query holdings by meaning, not just by keywords. This opens up spaces in which images and documents find each other that were previously separated: series emerge, cross-connections become visible and stories become more complete. Speed turns into diligence because the path to a reliable initial description is shorter, leaving more time for curatorial and research work.
Reputable AI in the cultural sector requires sovereignty and transparency. That means: roles and rights concepts, audit trails and documented approvals, on-premises operation or a strictly data protection-compliant cloud. When a system makes a suggestion, it must be comprehensible why. This is the only way to build trust – both internally and externally.
What is the vision? An archive, a museum, a collection that are as accessible as a good exhibition: clearly structured, deeply researched, intuitive to explore. The technology works quietly in the background, the human remains at the centre. AI pre-organises – humans curate. Exactly therein lies the power to preserve cultural heritage and at the same time retell it in a lively way.
— Karl‑Heinz Land, Founder and CEO of neuland.ai AG
Interview with Norbert Minwegen, Managing Director of the SK Stiftung Kultur
Mr Minwegen, why did the SK Stiftung Kultur decide on AI-supported indexing?
Our institutions work with large, historically grown collections. We want to make them quicker to index and richer in context to access – for our specialist teams, for research and for the public. AI helps us to speed up routine processes without relinquishing responsibility. This strengthens quality and creates free space for what matters most: interpreting, communicating, curating.
What changes specifically in the Photographic Collection?
The platform suggests motives, locations and time periods and complements title variations. This is a real relief in the initial description. At the same time, semantic and visual search improve research quality. An image becomes searchable not only via a keyword, but also via similarities and contexts. Our specialists verify, correct and approve – every decision is documented.
And what does AI mean for the German Dance Archive Cologne?
Movement needs timestamps. We segment video and audio material into scenes and sequences and link these with programmes, reviews, photographs or biographies. This allows performances to be indexed and researched with pinpoint accuracy. This saves time, increases reliability and creates new avenues for education and research.
You emphasised that cross-connections between collections become visible. Can you elaborate on that?
Gladly. Our objective with an internal AI is to make the collections of both areas – Photographic Collection and German Dance Archive Cologne – visible as a whole. The visual and semantic search makes series, motif sequences and unexpected relationships visible. Investigating this would have previously required an enormous amount of time. Already now we can see what opportunities are opening up: motives and locations are recognised, info enriched. For the first time, we can draw cross-connections between both collections, which remained invisible in separate systems. It is a different way of looking at what we have.
How do you ensure data sovereignty and quality?
We work with clear roles and rights concepts, documented approvals and auditable workflows. Operation is autonomous – on our own infrastructure or in a data protection-compliant cloud. AI suggestions are always traceable and are verified by specialists. This builds trust and ensures that speed does not come at the expense of diligence.
What impact do you expect in the medium term for research, education and the public?
Faster, more accurate searches and richer contexts. Researchers discover connections that were previously hidden. For education, new narrative pathways emerge, and the public benefits from clearer access – both digitally and on-site. In the end, cultural memory benefits because more material becomes visible and usable.
Your look ahead: what are the next steps?
We will gradually expand the proven workflows and connect further collections. The close involvement of the teams remains crucial – training, feedback, continuous quality assurance. Our benchmark is simple: AI assists, humans decide. This is what we are aligning our next steps with.
If you would like to learn more about the platform or gain an insight into the workflows, please write to us at info@neuland.ai. We will show you the neuland.ai HUB in action – sovereign, auditable and tailored perfectly for archives, museums and collections.