Skip to main content

Author: dsa_admin

Humans and AI Safety Testing

July 11, 2025, by Rebecca Balebako, CEO Brandworth.AI

As organizations integrate advanced AI systems into their business, ensuring their safety, reliability, and robustness becomes a critical priority.  However, AI testing remains a complex and evolving field. Key challenges that businesses face when considering how to run AI testing in practice are:

  • What expertise is required to ensure secure application?
  • How much testing is enough and how much will it cost?  
  • Testing and security frameworks often rely on the LLMs or AI themselves to judge the safety.  Are the AI’s guarding themselves correctly? 
  • How can businesses decide between existing testing frameworks, human testers or both?

We gathered 20 experts on an April morning in 2025 to discuss these questions.  The workshop combined interactive attempts to “hack” an AI judge, and smaller group discussions centered on practical themes that Swiss companies will face when testing their AI systems.

The goals of the workshop were to explore three challenges:

  1. AI safety discussions are fragmented by varying terminology and approaches across security, legal, and operational domains.  We convened a multidisciplinary group of experts in AI security and safety to foster a cross-domain discussion on AI testing.
  1. The challenge lies in validating AI’s self-judgment and understandingwhen AI judgement aligns with human intelligence.  We evaluated two AI-as-judge testing frameworks against human expectations by using a real-world use case and a live jailbreaking challenge.
  1. For businesses, the new field of AI safety testing presents a challenge in determining what to test and who should conduct it.  We discussed practical AI testing challenges for Swiss businesses through breakout groups focused on metrics, expertise, and opportunities.

We addressed each of these goals through an interactive group activity, followed by breaking out into three smaller groups. 

Workshop Outcomes and Decisions

Multidisciplinary Collaboration

Experts from various fields, including security engineers, lawyers, project managers, and model builders, engaged in open and productive discussions on AI testing. The diversity of perspectives was particularly valuable in highlighting different aspects of AI security and safety.  Attendees were largely industry professionals who are engaging in technical safety and security from various angles and for different companies.  Startups, medium-sized businesses, and larger B2C companies were represented.

AI-as-Judge Evaluation:

Through the use case of a cosmetic surgeon chatbot, the workshop demonstrated that existing AI-as-judge frameworks may not always align with human judgment. Multiple instances were found where chatbots gave responses that humans considered “bad” but were not flagged by AI evaluation systems. This highlights the need for supplementing AI evaluations with human feedback.

Breakout Group Findings and Decisions:

  • Metrics for AI Risk Assessment: Participants emphasized the importance of aligning AI behavior with “what a reasonable person would expect.” This requires teams to carefully examine the AI’s data, analyze target users, and clearly define the scope of testing. Standards can serve as a benchmark for these expectations.
    • Opportunities and Strategic Considerations: The group highlighted the need for a human-centered approach and multi-stakeholder input in defining AI project requirements. They also identified the need for practical enforcement mechanisms, beyond just regulations, to ensure AI safety.
    • Expertise and Skillsets for AI Safety: It was recognized that the required expertise for AI safety is context dependent. Key considerations include human rights, interdisciplinary “red teams,” and protecting human testers from harmful content.

These discussions provided valuable insights that should inform future AI safety and testing practices within Swiss businesses.

Summary

This workshop proved invaluable in addressing the complex and evolving landscape of AI testing and security. By bringing together a diverse group of experts, we gained critical insights into aligning AI behavior with human expectations and identifying practical steps for ensuring AI safety in Swiss businesses.

The collective knowledge shared, and the collaborative discussions fostered will significantly contribute to shaping future AI safety practices. We extend our sincere gratitude to our generous sponsors, Innovation Booster Artificial Intelligence, Cyberfy, and Brandworthy.AI, whose support made this important workshop possible.

AI in Robotics & Mobility

May 04, 2025, by Reik Leiterer, data innovation alliance

On March 4th, 2025, the co-creation workshop of the three innovation boosters New Mobility, Robotics and Artificial Intelligence on the overarching topic of AI & Robotics in Mobility took place at EPFL in Lausanne. Four focus topics elaborated and presented by the Swiss Federal Railways SBB were used as base for exploration of challenges related to these topics and to discuss and outline possible ideas in the specific fields. The initial focus topics were:

1) IMPROVING CUSTOMER EXPERIENCE ALONG THE TRAVEL CHAIN
How to enhance the overall travel experience by e.g. personalized, real-time travel information during disruptions – ensuring passengers are well-informed and minimizing inconvenience.

2) MAINTENANCE AND REPAIR
How to use robotics in operational activities for maintenance to save time and costs, or to increase safety, for e.g. train or infrastructure inspection, lifting heavy equipment, hazardous environments, or automating heavily redundant tasks.

3) OPTIMIZATION OF PASSENGER FLOW
How to leverage predictive analytics to optimize passenger/crowds, by e.g. analysing environmental conditions, historical data, or scheduled events.

4) AUTONOMOUS DRIVING
How to develop a comprehensive travel chain that integrates autonomous driving technologies, including the use of autonomous vehicles for in-house logistics and seamless transitions between different modes of transport.

The workshop started with the inspiring keynote by Alexandre Alahi (Associate Professor for Visual Intelligence for Transportation), not only recognized as one of the top 100 most influential scholar in Computer Vision over the past 10 years but also co-founder of multiple start-ups.

To make AI-driven systems a safe reality, he presented his lab works at the intersection of Computer Vision, Machine Learning, and Robotics: a new type of so-called socially aware AI, i.e., an AI augmented with social intelligence.

This was followed by the first breakout sessions, in which participants divided up according to their interest in the various focus topics. These breakout sessions were moderated and supported by experts Kenan Zhang, Christopher Ganz, Laszlo Etesi, and Eric Silva to ensure targeted problem exploration and ideation. Among other things, the discussion focused on key questions such as: What are the biggest challenges? Where do we see bottlenecks or unresolved pain points (e.g., perception technology, ethical AI, human-machine interaction, regulatory hurdles)? And what problems have we personally experienced or observed?

The following key challenges were identified and discussed, which were then embedded in the larger SBB context through a quick review and reflected upon regarding their potential for further exploration:

1) IMPROVING CUSTOMER EXPERIENCE ALONG THE TRAVEL CHAIN

  • How to realize mobility as a Service (MaaS), enabling a seamless and personalized travel experience from A to B, including different modes of transport (public transit, private vehicles, shared mobility) and context-aware travel recommendations?
  • How to improve pre- and post-trip experience, including the handling of passenger belongings, i.e. optimizing luggage storage and management.
  • How to optimize international travelling, considering booking and payment challenges (e.g., tickets reserving across borders) or navigating different platforms and connections (e.g., transitioning in Germany) while maintaining the high data security standards of SBB?
  • How to prevent incidents on the platforms caused by overcrowding or pushing, considering also passengers’ perception of safety.

2) MAINTENANCE AND REPAIR

  • What maintenance needs to be done on the equipment even if it’s not degrading (e.g. lubrication, cleaning) considering existing/upcoming regulations requiring maintenance activities and the general operation time of a component?
  • How to optimize the time to bring the equipment back to operation if degrading (i.e. shorter maintenance time/downtime), including management of critical spare parts?
  • How to improve fault detection by not only identifying problems but also by improving the predictive fault detection while keeping the safety & security standards?

3) OPTIMIZATION OF PASSENGER FLOW

  • How to balance comfort, service quality, and safety, mainly related to the rush hours (i.e. improving crowd load management)?
  • How to improve passenger by dealing with e.g. visibility challenges (e.g. colour blindness) or communication barriers, considering cultural differences in signalling and safety measures?

4) AUTONOMOUS DRIVING

  • How to optimize “in-between” travel (e.g., train-to-gate transitions), with solutions for people with reduced mobility and/or elderly people?
  • How to enable multimodal solutions by re-using rail infrastructure for e.g., autonomous vehicles?
  • How to realize am infrastructure for autonomous driving, related to electrification and charging infrastructure (e.g., remote charging solutions & autonomous parking)

After lunch, a second round of breakout sessions was held to discuss possible solutions to the challenges that had been identified, which were then presented. These cross-ecosystem exchanges were invaluable for rethinking the challenges holistically and unlocking potential new synergies.

Based on this, the innovation teams were then able to submit their ideas within three weeks as part of the idea stage of the Innovation Booster New Mobility. The following submissions were evaluated by an external panel of experts and out of them three ideas were selected for funding:

  • Autonome Mobilität neugedacht durch persönliche digitale Chauffeure
  • Autonomous Mobile Robot with Energy Storage for Distributing Renewable Energy to Every Parking Space with Minimal Infrastructure Costs
  • Distribution Management Customer Service App
  • Forecasting crowds to reduce train platforms occupancy during rush hour
  • Inclusive PRM Services: Bridging Accessibility Gaps with Autonomous Wheelchairs
  • Leveraging Autonomous Mobility to Shift Car-Dependent Travelers Towards Public Transport: A Blueprint for a User-Centered Business Model based on Personas and Psychological Needs
  • PREACT – Predictive Analytics for Crowd Transit Optimization
  • Privacy-preserving Monitoring of Customer Flows and Intentions
  • RailWise: Empowering SBB with Staff-Driven AI Insights
  • Real-time Detection and Correlation of Infrastructure Damage Events
  • Reinventing Rail Maintenance: Real-Time IoT Meets Low-Code Automation
  • Robotic Platform for maintenance works in rail depots
  • Sensorized robotic arm for enhanced train maintenance and safety

Finally selected project ideas:

  1. PREACT – Predictive Analytics for Crowd Transit Optimization
    Project Idea – jointcreate.com
  1. Inclusive PRM Services: Bridging Accessibility Gaps with Autonomous Wheelchairs
    Project Idea – jointcreate.com
  1. Leveraging Autonomous Mobility to Shift Car-Dependent Travelers Towards Public Transport
    Project Idea – jointcreate.com

The format convinced with very good speakers, experts, and open-minded participants interested in exchange and cooperation. We would like to thank all participants for joining this workshop and for their numerous high-quality submissions! We are already looking forward to the next event and a lively and active participation.

Harnessing Hybrid Intelligence: Smarter Together

April 02, 2025, by Jürg Meierhofer, Expert Group Smart Services

At the ZHAW Service Lunch on April 2nd, 2025, PD Dr. Jochen Wulf and Prof. Dr. Frank Hannich explored the exciting potential of ‘Hybrid Intelligence for Customer Management’, sharing conceptual possibilities and concrete ZHAW project examples on using generative AI across the customer lifecycle.

Ein Bild, das Text, Diagramm, Screenshot, Design enthält.

KI-generierte Inhalte können fehlerhaft sein.

What happens when human ingenuity meets AI’s power? You get Hybrid Intelligence (HI) – systems designed for humans and AI to collaborate, achieving more than either could alone.

The goal isn’t replacement, but augmentation.

Their ZHAW research explores HI, especially in customer management and technical service. They highlighted powerful applications like personalized content, churn prediction, and augmented customer support.

A key focus is using Large Language Models (LLMs) for tasks like technical assistance, as examplified in their “Guided Maintenance Copilot” project. This involves Retrieval Augmented Generation (RAG) to ground AI responses in specific knowledge.

Developing effective RAG systems requires careful, test-based optimization across factors like search, model selection, and context size. Different models (CPU vs. GPU) and languages show varying performance.

While LLMs can automate tasks from simple translation to complex reasoning, their findings emphasize: there’s no “one-size-fits-all.” Tailored, optimized solutions are key to unlocking the true potential of human-AI collaboration. Ultimately, the session demonstrated that thoughtful design and testing are crucial for successfully implementing hybrid intelligence solutions.

asut Conference “Shaping the Future of IoT – Technologies, Business Models, Security”

March 11, 2025, by Jürg Meierhofer, Expert Group Smart Services

The IoT Conference 2025, held on March 11 at the Trafo Hallen Baden, was a resounding success, offering an inspiring glimpse into the future of the Internet of Things (IoT). Under the theme “Shaping the Future of IoT – Technologies, Business Models, Security,” around 20 speakers from industry, research and development, as well as providers and users, presented innovative solutions and application examples to the audience of several hundred people.

While the entire Expert Group was invited to attend the conference under special conditions (including some free tickets), the group also actively contributed to the conference in two significant ways. At the start of the conference, Jürg Meierhofer participated in a panel discussion and shared thoughts about value creation models by IoT-based smart services.

Additionally, Gerrit Schatte, Head of Digital Products at Kistler, delivered a noteworthy presentation. He showcased impressive examples of how Kistler, a leading measurement systems manufacturer, is successfully creating entirely new business areas with digital products through the integration of hardware and software. These digital products enable Kistler not only to offer precise measurement systems but also comprehensive digital solutions that open up new markets and applications. Gerrit Schatte presented three case studies:

  • Onboard Rail Monitoring for online noise monitoring for rail head conditioning.
  • Structural Health Monitoring for monitoring the structural health of bridges.
  • Acceleration of flight tests through advanced data analysis and management platforms.

The presentation demonstrated how Kistler is developing innovative solutions through the integration of IoT and digital products, effectively addressing customer challenges and opening new business fields.

A big thank you goes to the organizers of asut for this inspiring exchange, which made the future of IoT tangible and provided important impulses for the further development and application of these technologies.

GeoAI in Switzerland: innovations, developments and perspectives

(German Version below)

March 11, 2025, by Nicolas Lenz, Stefan Keller and Reik Leiterer, Focus Topic Group Spatial Data Analytics

Shortly after the data innovation alliance was founded, geodata analyses were regularly discussed in a dedicated expert group. From 2020-2024, the Innovation Booster Databooster, an initiative managed by the data innovation alliance and powered by Innosuisse, has promoted innovation in the data-driven sector and contributed to the development of new technologies. Within the Databooster, spatial data was one of four focus topics. The transition to the follow-up initiative Innovation Booster Artificial Intelligence has been successfully implemented. The completion of the Databooster is the ideal moment to take stock: where does Switzerland stand in 2024 in terms of GeoAI use?

This document provides an overview of developments over the last four years, highlights changes in GeoAI technologies and application fields, and introduces the actors who have significantly shaped this transformation. It is aimed at researchers, industrial actors and decision-makers, and offers orientation and inspiration for the opportunities and challenges of the next phase of innovation.

GeoAI 2020-2024

The use of artificial intelligence in the GIS context has evolved in recent years from specialized applications to a wide range of innovative technologies. Four years ago, GeoAI was mainly used in the field of computer vision for object and pattern recognition in aerial images and for classifying point clouds. Since then, the range of applications has expanded significantly, including the field of evaluating time series to identify trends or anomalies.

Not only has the number of AI tools increased, but also their possible applications. For example, AI-supported geodata analyses have been used in the areas of mobility (timetable optimization, management of passenger flows), tourism (recommendation systems for excursion destinations and route difficulties), or in the insurance industry (risk assessments, forecasts).

The main reason for this development is that new AI technologies have become suitable for mass use. This has expanded their use not only in a research context, but also in business and public administration. The latter in particular has helped to make innovative applications visible to the general public.

Example: Use of GeoAI in federal agencies – pollen monitoring by MeteoSwiss (from a presentation by Bertrand Loison, FSO – SITC 2024)

A decisive factor for the dissemination of AI applications is the rapid development of language models as part of generative AI. These technologies extend existing applications and create entirely new possibilities. For example, speech-enabled geospatial systems enable improved search in large data collections or more intuitive use of GIS, such as for data collection form creation or to convert text queries into SQL or Python code. These new possibilities and the increased focus on user-friendliness have permanently changed the GeoAI provider landscape. Many new players are developing or using new technologies that manage geodata more efficiently and tap into unused potential. For example, it is difficult to grasp the influence of open source: Suffice to say that open source is playing an increasingly important role, whether as a software component or as an open language or foundation model. It promotes innovation, dissemination, transparency, trust and data protection, and ultimately leads to cost reduction and democratization (through low barriers to entry).

What role do the well-known GIS companies currently play?

Established Swiss GIS providers have long since discovered the use of GeoAI. Well-known providers such as EBP, ESRI Switzerland, Geocloud, GeoDataSolutions, Geowerkstatt, Hexagon or Meteotest (selection not exhaustive) have made significant progress in integrating AI into their GIS applications in recent years. They have successfully introduced AI technologies into their respective target industries and applied them together with clients from the private and public sectors.

Example: ESRI’s Data Science/Deep Learning product family – integrated into Microsoft Azure.

In addition to these established providers, several start-ups have managed to become recognized and established market players in recent years. Companies such as Gamaya, Meteomatics, Picterra and Pix4D have continuously developed their innovative technologies and firmly established them in various industries. Today, these companies stand for the successful transfer of GeoAI innovations from the start-up phase to marketable products and services that are used both nationally and internationally.

Momentum at start-ups and small companies

Young companies are bringing specialized solutions to market, using the latest technologies from AI. It’s about companies like Ageospatial, askEarth, ExoLabs, LaGrand, Litix or UrbanDataLab, which are using new technologies to solve specific challenges in handling geospatial data. In doing so, they are breathing fresh air into the market by pushing the boundaries of traditional applications and opening up geodata to new target groups.

Example: The Geoforge application from Ageospatial – Geospatial analysis with large language models.

These young companies show that the future of GeoAI will not be shaped by established players alone. They are introducing new ideas and approaches that could invigorate the market and set new standards. Whether these young companies will secure a long-term place among the established providers or how the market will develop as a whole remains to be seen. However, it is clear that innovation and adaptability are crucial in a dynamic environment such as GeoAI – for both emerging and established companies.

Innovative start-ups and small companies in geoinformatics

  • User-friendly use of GIS and geodata:
    • Ageospatial SARL (ageospatial.com)
      Voice-controlled geoinformation systems.
    • askEarth AG (ask.earth)
      Voice-controlled search engine for geodata.
    • Litix GmbH (litix.ch)
      Data extraction from document archives for map visualization and process support.
  • Focus on environment, agriculture and remote sensing:
    • ExoLabs AG (exolabs.ch)
      Analysis of Earth Observation data for environmental monitoring.
    • Gamaya SA (gamaya.com)
      Increasing efficiency and reducing CO₂ in agriculture.
    • Meteomatics AG (meteomatics.com)
      AI-supported weather data for environmental analysis.
    • Picterra (picterra.ch)
      Cloud-based GeoAI platform for analyzing Earth Observation images, with a focus on customer-specific AI solutions.
    • Pix4D SA (pix4d.com)
      Photogrammetry for point cloud analysis, strong in agricultural practice.
  • Focus on urban planning and infrastructure:
    • LaGrand GmbH (lagrand.ch)
      Analysis of temporal changes in images for informed urban decisions.
    • UrbanDataLab (urbandatalab.ch)
      Location data and analysis tools for risk models and location decisions.

What does the future hold?

The rapid development of GeoAI impressively demonstrates the potential of combining artificial intelligence and geoinformatics. Applications such as voice-controlled assistants in QGIS, ArcGIS or the AI-based search for public geoservices (via Geoharvester) are examples of upcoming innovations. These technologies will not only revolutionize the way we find and use geodata but also make it accessible to a wider audience.

Despite the many opportunities offered by GeoAI, the risks should not be underestimated. Particularly around generative AI, critical examination is necessary. The hype surrounding this technology can easily lead to an overestimation of its capabilities. Generative models deliver impressive, but not always reliable, results. Start-ups and other actors should be aware that these technologies require in-depth expertise and careful validation. The unreflective application of such systems carries the risk of making wrong decisions or losing the trust of users.

The running Innovation Booster – Artificial Intelligence offers an opportunity to proactively address these challenges. The initiative builds on the successes of the Databooster and creates a platform to bring together researchers, companies and decision-makers. It will be crucial in helping to shape the next phase of GeoAI in Switzerland – responsibly, sustainably and with an eye to the future.

(Note: Some of the authors of this White Paper are involved in or act in an advisory capacity for the start-ups mentioned. The content presented is based on their expertise and market experience, regardless of their role in these companies.)

GeoAI in der Schweiz: Innovationen, Entwicklungen und Perspektiven

Verfasst durch die Fokusgruppe Spatial Data Analytics (N. Lenz, S. Keller & R. Leiterer)

Seit 2020 hat der Innovation Booster Databooster, eine von Innosuisse unterstützte Initiative der data innovation alliance, Innovationen im datengetriebenen Bereich gefördert und zur Entwicklung neuer Technologien beigetragen. Bereits kurz nach der Gründung der data innovation alliance wurden Geodatenanalysen regelmässig in einer eigenen Expertengruppe diskutiert. Innerhalb des Databoosters bildeten die räumlichen Daten eines von vier Fokusthemen. Ende 2024 wurde der Databooster erfolgreich abgeschlossen und der Übergang zur Nachfolge-Initiative Innovation Booster Artificial Intelligence umgesetzt. Der Abschluss des Databoosters ist der ideale Moment, um Bilanz zu ziehen: Wo steht die Schweiz 2024 bei der Nutzung von GeoAI?

Das vorliegende Dokument gibt einen Überblick über die Entwicklungen der letzten vier Jahre, beleuchtet die Veränderungen in den Technologien und Anwendungsfeldern von GeoAI und stellt die Akteure vor, die diese Transformation massgeblich geprägt haben. Es richtet sich an Forschende, industrielle Akteure und Entscheidungsträger und bietet Orientierung sowie Inspiration für die Chancen und Herausforderungen der nächsten Innovationsphase.

GeoAI 2020-2024

Die Nutzung von künstlicher Intelligenz im GIS-Kontext hat sich in den letzten Jahren von spezialisierten Anwendungen zu einem breiten Spektrum innovativer Technologien entwickelt. Vor vier Jahren wurde GeoAI vor allem im Bereich Computer Vision für die Objekt- und Mustererkennung auf Luftbildern und zur Klassifizierung von Punktwolken eingesetzt. Inzwischen hat sich das Spektrum der Anwendungen deutlich erweitert, unter anderem im Bereich der Auswertung von Zeitreihen im Hinblick auf die Erkennung von Trends oder Anomalien.

Nicht nur die Anzahl der KI-Werkzeuge hat zugenommen, sondern auch ihre Anwendungsmöglichkeiten. So wurden KI-gestützte Geodatenanalysen u.a. in den Bereichen Mobilität (Fahrplanoptimierungen, Lenkung von Passagierströmen), Tourismus (Empfehlungssysteme zu Ausflugszielen und Routenschwierigkeiten), oder im Versicherungswesen (Risikoabschätzungen, Prognosen) eingesetzt.

Hauptgrund für diese Entwicklung ist, dass neue KI-Technologien massentauglich wurden. Dadurch liess sich die Nutzung nicht nur im Forschungskontext, sondern auch in der Wirtschaft und der öffentlichen Verwaltung erweitern. Gerade letztere hat dazu beigetragen, dass innovative Anwendungen auch für die breite Bevölkerung sichtbar wurden.

Beispiel: Einsatz von GeoAI in den Bundesstellen – Pollenmonitoring durch MeteoSwiss (aus einer Präsentation von Bertrand Loison, BFS – SITC 2024)

Ein entscheidender Faktor für die Verbreitung von KI-Anwendungen ist die rasante Entwicklung von Sprachmodellen als Teil der generativen KI. Diese Technologien erweitern bestehende Anwendungen und schaffen völlig neue Möglichkeiten. Sprachgesteuerte Geodatensysteme ermöglichen beispielsweise verbesserte Suche in grossen Datensammlungen oder eine intuitivere Nutzung von GIS, wie beispielsweise für die Datenerfassungs-Formular-Erstellung oder aber um Textanfragen in SQL- oder Python-Code umzuwandeln.

Diese neuen Möglichkeiten und der verstärkte Fokus auf Benutzerfreundlichkeit haben die GeoAI-Anbieterlandschaft nachhaltig verändert. Viele neue Akteure entwickeln oder nutzen neue Technologien, die Geodaten effizienter bewirtschaften und ungenutzte Potenziale erschliessen. Der Einfluss von Open Source ist schwer zu erfassen. Nur so viel: Open Source spielt eine immer wichtigere Rolle, sei es als Softwarekomponente oder als offenes Sprach- oder Foundation-Modell. Es fördert Innovation, Verbreitung, Transparenz, Vertrauen, Datenschutz und führt letztlich zu Kostenreduktion und Demokratisierung (durch niedrige Einstiegshürden).

Welche Rolle spielen aktuell die bekannten GIS-Unternehmen?

Etablierte Schweizer GIS-Anbieter haben die Nutzung von GeoAI längst entdeckt. Bekannte Anbieter wie EBP, ESRI Schweiz, Geocloud, GeoDataSolutions, Geowerkstatt, Hexagon oder Meteotest (Auswahl nicht abschliessend) haben in den letzten Jahren bedeutende Fortschritte bei der Integration von KI in ihre GIS-Anwendungen erzielt. Sie haben erfolgreich KI-Technologien in ihre jeweiligen Ziel-Branchen getragen und zusammen mit den Auftraggebern aus dem privaten und öffentlichen Sektor angewendet.

Beispiel: Die Produktfamilie Data Science/Deep Learning von ESRI – integriert in Microsoft Azure.

Neben diesen etablierten Anbietern haben einige Start-ups der letzten Jahre den Schritt zu anerkannten und etablierten Marktteilnehmern geschafft. Firmen wie Gamaya, Meteomatics, Picterra und Pix4D haben ihre innovativen Technologien kontinuierlich weiterentwickelt und in verschiedenen Branchen fest verankert. Diese Unternehmen stehen heute für den erfolgreichen Transfer von GeoAI-Innovationen aus der Gründungsphase hin zu marktfähigen Produkten und Dienstleistungen, die sowohl national als auch international Anwendung finden.

Dynamik bei Start-ups und Kleinfirmen

Durch den Einstieg neuer, innovativer Start-ups erfährt die Geoinformatik aktuell eine neue Dynamik. Junge Unternehmen bringen spezialisierte Lösungen auf den Markt und nutzen dabei modernste Technologien aus der KI. Die Sprache ist von Firmen wie Ageospatial, askEarth, Exolabs, LaGrand, Litix oder UrbanDataLab, welche neue Technologien nutzen, um spezifische Herausforderungen im Umgang mit Geodaten zu lösen. Dabei bringen sie frischen Wind in den Markt, indem sie die Grenzen traditioneller Anwendungen erweitern und Geodaten für neue Zielgruppen erschliessen.

Beispiel: Die Geoforge Anwendung von Ageospatial – Geodatenanalyse mit Large Language Models.

Diese Jungunternehmen zeigen, dass die Zukunft der GeoAI nicht allein von etablierten Akteuren gestaltet wird. Sie bringen neue Ideen und Ansätze ein, die den Markt beleben und neue Standards setzen könnten. Ob diese Jungunternehmen sich langfristig einen Platz unter den etablierten Anbietern sichern oder wie der Markt sich insgesamt entwickeln wird, bleibt offen. Klar ist jedoch, dass Innovation und Anpassungsfähigkeit in einem dynamischen Umfeld wie der GeoAI entscheidend sind – sowohl für aufstrebende als auch für bereits etablierte Unternehmen.

Innovative Start-ups und Kleinfirmen in der Geoinformatik

  • Nutzerfreundliche Nutzung von GIS und Geodaten
    • Ageospatial SARL (ageospatial.com)
      Sprachgesteuerte Geoinformationssysteme.
    • askEarth AG (ask.earth)
      Sprachgesteuerte Suchmaschine für Geodaten.
    • Litix GmbH (litix.ch)
      Datenextraktion aus Dokument-Archiven zur Kartenvisualisierung und Prozessunterstützung.
  • Fokus Umwelt, Landwirtschaft und Fernerkundung
    • ExoLabs AG (exolabs.ch)
      Analyse von Erdbeobachtungsdaten für Umweltüberwachung.
    • Gamaya SA (gamaya.com)
      Effizienzsteigerung und CO₂-Reduktion in der Landwirtschaft.
    • Meteomatics AG (meteomatics.com)
      KI-gestützte Wetterdaten für Umweltanalysen.
    • Picterra (picterra.ch)
      Cloudbasierte GeoAI-Plattform zur Analyse von Erdbeobachtungsbildern, mit Fokus auf kundenspezifische KI-Lösungen.
    • Pix4D SA (pix4d.com)
      Photogrammetrie für Punktwolkenanalysen, stark in der landwirtschaftlichen Praxis.
  • Fokus Stadtplanung und Infrastruktur:
    • LaGrand GmbH (lagrand.ch)
      Analyse zeitlicher Veränderungen in Bildern für fundierte urbane Entscheidungen.
    • UrbanDataLab (urbandatalab.ch)
      Standortdaten und Analysetools für Risikomodelle und Standortentscheidungen.

Was bringt die Zukunft?

Die rasante Entwicklung von GeoAI zeigt eindrucksvoll, welches Potenzial in der Verbindung von Künstlicher Intelligenz und Geoinformatik steckt. Anwendungen wie sprachgesteuerte Assistenten in QGIS, ArcGIS oder die KI-basierte Suche nach öffentlichen Geodiensten (via Geoharvester) stehen beispielhaft für bevorstehende Innovationen. Diese Technologien werden nicht nur die Art und Weise revolutionieren, wie wir Geodaten finden und nutzen, sondern sie auch für ein breites Publikum zugänglich machen.

Trotz der vielen Chancen, die GeoAI bietet, gilt es, die Risiken nicht zu unterschätzen. Besonders im Bereich der generativen KI ist eine kritische Auseinandersetzung notwendig. Der Hype um diese Technologie kann leicht dazu führen, ihre Fähigkeiten zu überschätzen. Generative Modelle liefern beeindruckende, aber nicht immer verlässliche Ergebnisse. Start-ups und andere Akteure sollten sich bewusst machen, dass diese Technologien fundiertes Fachwissen und sorgfältige Validierung erfordern. Die unreflektierte Anwendung solcher Systeme birgt die Gefahr, falsche Entscheidungen zu treffen oder das Vertrauen der Nutzer zu verlieren.

Mit dem laufenden Innovation Booster – Artificial Intelligence bietet sich die Gelegenheit, diesen Herausforderungen proaktiv zu begegnen. Die Initiative knüpft an die Erfolge des Databoosters an und schafft eine Plattform, um Forschende, Unternehmen und Entscheidungsträger zusammenzubringen. Sie wird entscheidend dazu beitragen, die nächste Phase von GeoAI in der Schweiz mitzugestalten – verantwortungsbewusst, nachhaltig und zukunftsorientiert.

(Hinweis: Einige der Autoren dieses Whitepapers sind an genannten Start-ups beteiligt oder in beratender Funktion tätig. Die dargestellten Inhalte basieren auf ihrer Expertise und Markterfahrung, unabhängig von ihrer Rolle in diesen Unternehmen.)

A Last Farewell: Databooster Finale

December 10, 2024, by Gundula Heinatz Bürki, data innovation alliance

The Innovation Booster Databooster, powered by Innosuisse, was a success story as a program. 4 years of promoting innovation – from community building to organizing ideation and co-creation sessions to co-financing innovative ideas. So it was time to come together one last time and offer a platform to a selected number of innovation teams with fascinating ideas to tell their personal success story. More than 50 participants gathered at the Technopark Zurich for the Databooster Finale on December 05, 2024, to share the experiences and to set-up the next steps with a collaborative and open innovation mindset. It was truly inspiring to learn so many innovative ways how data-driven methods can improve the lives of humans, animals and society as a whole.

Read on to gain an insight into the projects that the Databooster has supported over the course of its activities. If you want to see the full overview of all supported projects, have a look at our website.

The keynote speaker Kerstin Johansson Baker from the Federal Statistical Office informed the audience about the usage of Data Science for Public Good. Kerstin presented some projects of the Data Science Competence Center for the Swiss public administration and how researchers can collaborate with the center. It was interesting to hear how far the public sector has come in making data available – not only in terms of data quality, e.g. with regard to standardized metadata, but also in terms of access options that enable seamless integration into third-party applications. A treasure trove of data that is still underused by the private sector and offers great potential for developing innovative products and services.

The project presentations started with Polina Veltmann and Romain Meisterhans from Smilamind, who presented the professional support for efficient treatment of youth mental health and wellbeing with smart remote monitoring for. It was important to co-create with young people the solution, to collect feedback from stakeholders to optimize clinical effectiveness, and of course to ensure the correct handling of legal requirements and ethical-moral issues.

In the following presentation, Mel Zürcher from rewoso informed the audience about a medical database supporting medical outcomes with quick insights in the current state of science. Together with the Bellevue Medical Group, the team addressed the challenge of the limited availability of medical data in structural form. Finally, they shifted the idea to an individual dashboard for patients.

The next presenter Fabio Lambruschi from Sourceway Sarl presented their project about cervical spine assessment for the range of motion. Cervical spine pathology is becoming increasingly prevalent leading to significant social and healthcare costs worldwide. Together with the stakeholders, they developed an AI-driven mobile app as technology for everyone. It creates economic value by reducing the need for in-person clinical visits, thus lowering healthcare costs and enabling faster preliminary clarification for a larger part of the population.

Nic Lenz from Litix demonstrated how to access a large archive of documents in the field of geology from the past 60 years. The innovation team used interviews with potential customers from public administration, notary office, fire fighters and insurance companies to develop a vision and to test the idea with AI-powered maps. This idea once again highlighted the great added value that can be found in existing data archives when using current methods and technologies.

Joana Kühne from Artificialy presented the innovation journey together with Forbo to improve the tile manufacturing process using AI technology. The involved stakeholder got a deeper understanding of the process and specific defect types .It was nicely demonstrated here how collaboration between larger industrial organisations, agile SMEs, and research can create added value for all parties involved.

In the next project presentation from Martin Beth (OST), the challenges of the building industry were discussed. The team addressed the existing needs of digitalization in the building industry and in the frame of Building Information Modeling (BIM). The team investigated this challenge and built a first solution from a rule-based to an AI-based quality management of BIM models.

Maaz Sheikh from Ageospatial Sàrl presented their innovation journey together with the University of Geneva to address the challenges of geoinformation systems in Switzerland. The team identified the limits and potentials of GeoLLMs and created a proof of concept with Swiss Federal data. This approach used location intelligence for supporting informed decisions by different stakeholders from the insurance industry as well as from municipalities.

Other highlights included the presentation from Benoît Buchs from PrecisionFly. Their project aims at AI-based pipe inspection with the city of Winterthur. Additionally, Wolfram Willuhn from Plutinsus demonstrated how they could optimise the scaling of the heating system through stakeholder collaboration and building a digital twin. Johannes Löckinger from AI-Tails explained the challenges of identifying and tracking animal pain. First, the team realized that they have to ask the right questions to the stakeholder to understand the user and market interest. The team designed a solution filling an existing gap in the market for pets and livestock.

Before the networking apéro, the Databooster outcomes were presented after four successful years funded by Innosuisse. We would like to thank everyone who has supported us over by supporting our ideation events, evaluating the innovation ideas, submitting innovative projects, fuelling ideas or working behind the scenes. Without your help this would not have been possible.

Christoph Heitz, as president of the Leading House, concluded that the innovation journey of the IB Databooster came to an end, but new opportunities are arising with other Innovation Boosters, such as the IB Artificial Intelligence, and other instruments. Innovation never sleeps – and the data innovation alliance will continue to be at the forefront of innovation in the field of data and AI.

The Power of Remote Service for Value Creation

December 15, 2024, by Jürg Meierhofer, Expert Group Smart Services

Remote services are

  • Economic: faster response to customer need, less costs
  • Environmental: less emissions by travelling
  • Social: better plannable work time

Thanks to René Vogel (Mr. Vision), we have gained an impressive insight into how the new possibilities of AR / VR can support this. Many thanks to STAUFEN.INOVA AG (Thomas Spiess) for hosting us in their wonderful premises in a super central location in Zürich.

First Steps Towards AILAS – the AI Law Supporter

A team of researchers is currently developing AILAS (AI Law Supporter) as a tool for AI product/service developers to find out what kind of legal issues they face, and where they get help to address them. The Data Ethics Expert Group served as Sounding Board for this project that is supported by the AI Innovation Booster.

10 members of the Expert Group provided a feedback to the first version of the questionnaire presented by Janmejay Singh, who has the operative lead in this project. The questionnaire, structured along the logic of the EU AI act, will be further improved based on this feedback for the first Shaping Workshop that is planned for early December 2024. At this Shaping Workshop, a mock version of the tool will be presented to a stakeholder group that include both potential users (such as AI startups) and supporters (such as consultants) of AILAS. A second Shaping Workshop is planned for January 2025.

General Assembly of the data innovation alliance at HWZ

November 5, 2024, by Milena Perraudin, data innovation alliance

On November 5, the members and board of the data innovation alliance gathered for the annual General Assembly at the Hochschule für Wirtschaft Zürich (HWZ). The event, attended by 30 participants, opened with a keynote from Bertrand Loison of the Federal Statistical Office (FSO). During the assembly, members unanimously approved the 2023 financial report, the 2025 budget, and the board’s discharge. Updates on personnel changes and upcoming events were also shared, offering a comprehensive look at this year’s achievements and future plans.

Keynote: The Three Tasks of the Federal Statistical Office (FSO)

We were honored to welcome Bertrand Loison, Vice-Director of the FSO and Head of Data Science, AI, and Statistical Methods, as our keynote speaker. He outlined the FSO’s core mission of of bridging critical questions with statistical insights through three main responsibilities:

  1. Creating, managing, and disseminating statistical information – this includes standardizing heterogeneous datasets.
  2. Managing federal statistical registers – ensuring the multiple use of data through platforms like the National Data Management (NaDB) and the I14Y Interoperability Platform.
  3. Supporting public organizations in data science and AI – fostering data science principles across sectors and aiding in the evolution of these disciplines.

Bertrand Loison also presented the FSO’s strategic focus, emphasizing privacy-preserving data science and evidence-based policymaking. He concluded by introducing the Competence Network for Artificial Intelligence (CNAI), whose mission is to promote the sustainable adoption and public trust in AI and other emerging technologies.

General Assembly (GA)

The GA started with a review of the 2023 Annual and Financial Reports, which were unanimously approved by the members present. The 2025 budget and the discharge of the board was also unanimously approved.

Christoph Heitz, president of the data innovation alliance, presented key updates on the alliance’s two four-year Innovation Booster programs, which are powered by Innosuisse.  The Innovation Booster Databooster, now in its final year, successfully organized multiple outreach events and funded 20 shaping workshops along with 13 deep dives. The Databooster Finale will be held on December 5, 2024, at Technopark Zürich. More information and registration here. The Innovation Booster Artificial Intelligence currently in its first year, hosted over 20 outreach events and funded a total of 10 ideas, including six shaping workshops and four deep dives.

Following these program updates, two personnel changes were announced. Gundula Heinatz-Bürki, Managing Director of the data innovation alliance, will step down in February 2025 after seven years of dedicated service. Her work has been instrumental in shaping the alliance, and we extend our deepest gratitude for her commitment. We also thank Matthias Werner, who will be leaving the board at the end of 2024 after departing from Trumpf. His contributions, particularly during the 2022 strategy review, have left a lasting impact.

We would also like to acknowledge the ongoing support of board members Matthias Brändle (La Mobilière), Andrea Dunbar (CSEM & EPFL), Hans Peter Gränicher (D ONE), Christoph Heitz (ZHAW IDP), and Anne Herrmann (FHNW Institute for Market Supply & Consumer Decision-Making).

The assembly concluded with an exciting announcement: the 12th IEEE Swiss Conference on Data Science (SDS2025) will be held in Zurich on June 26-27, 2025. Calls for Participation are now open, and we invite all members to contribute their expertise to make this event a success, following our motto, “Together we move faster.” Save the date and find more information here. 

After the General Assembly, attendees enjoyed a networking aperitif, fostering connections and further collaboration in a convivial atmosphere.

The Role of Documents in RAG Applications. From Documents to AI-Ready Data

October 25, 2024, by Benjamin von Deschwanden and Martin Keller, Acodis

Acodis together with data innovation alliance organized a webinar, diving into the essential process of turning complex documents into AI-ready data, setting the foundation for implementing Retrieval-Augmented Generation (RAG) in real-world scenarios.

You  can watch it on demand here https://www.acodis.io/webinar-the-role-of-documents-in-rag-applications

The webinar is designed for data and analytics professionals aiming to bridge the gap between their current data and the AI capabilities of tomorrow. You’ll discover the critical role documents play in RAG applications and gain practical insights to help you streamline data extraction, uphold data quality, and maintain traceability. With step-by-step guidance, you’ll leave equipped to elevate your organization’s data processes, ensuring every document can be harnessed as a valuable AI asset.

What You’ll Learn:

  1. Mastering Document Complexity
    Not all documents are created equal; they vary greatly in format, structure, and language. This section of the webinar breaks down the different types of documents and offers techniques for managing complexity, enabling you to create a foundation for accurate data extraction.
  2. Achieving High Data Accuracy and Traceability
    Clean, reliable data is critical to successful AI applications. You’ll learn advanced techniques for extracting high-quality data and ensuring traceability so your AI outputs are consistent, dependable, and ready for complex applications like analytics, intelligent automation, and advanced querying.
  3. Implementing RAG in Your Organization
    RAG (Retrieval-Augmented Generation) has the potential to change the way your organization approaches data retrieval and interaction. In this segment, we cover actionable steps to incorporate RAG into your workflows, helping you maximize the value of your data and enhance AI-driven solutions for customer support, internal knowledge management, and more.

Why You Should Watch This Webinar

In an increasingly competitive market, being able to effectively harness data can be a defining advantage. The ability to transform documents into AI-ready data enables you to:

  • Improve analytics and insights across departments
  • Enhance the efficiency of internal processes
  • Reduce time spent on manual data handling through automation

Ideal for data architects, analytics leaders, and digital transformation teams, this webinar provides a clear roadmap to navigate the complexities of document-based data and unlock its full potential with cutting-edge RAG applications.

To dive deeper into AI-ready data strategies, download our detailed whitepaper, AI-Ready Data Explained, covering more techniques and real-world examples for successful data transformations.