en
                    array(2) {
  ["de"]=>
  array(13) {
    ["code"]=>
    string(2) "de"
    ["id"]=>
    string(1) "3"
    ["native_name"]=>
    string(7) "Deutsch"
    ["major"]=>
    string(1) "1"
    ["active"]=>
    int(0)
    ["default_locale"]=>
    string(5) "de_DE"
    ["encode_url"]=>
    string(1) "0"
    ["tag"]=>
    string(2) "de"
    ["missing"]=>
    int(0)
    ["translated_name"]=>
    string(6) "German"
    ["url"]=>
    string(73) "https://www.statworx.com/content-hub/blog/tag/artificial-intelligence-de/"
    ["country_flag_url"]=>
    string(87) "https://www.statworx.com/wp-content/plugins/sitepress-multilingual-cms/res/flags/de.png"
    ["language_code"]=>
    string(2) "de"
  }
  ["en"]=>
  array(13) {
    ["code"]=>
    string(2) "en"
    ["id"]=>
    string(1) "1"
    ["native_name"]=>
    string(7) "English"
    ["major"]=>
    string(1) "1"
    ["active"]=>
    string(1) "1"
    ["default_locale"]=>
    string(5) "en_US"
    ["encode_url"]=>
    string(1) "0"
    ["tag"]=>
    string(2) "en"
    ["missing"]=>
    int(0)
    ["translated_name"]=>
    string(7) "English"
    ["url"]=>
    string(76) "https://www.statworx.com/en/content-hub/blog/tag/artificial-intelligence-en/"
    ["country_flag_url"]=>
    string(87) "https://www.statworx.com/wp-content/plugins/sitepress-multilingual-cms/res/flags/en.png"
    ["language_code"]=>
    string(2) "en"
  }
}
                    
Contact

Business success hinges on how companies interact with their customers. No company can afford to provide inadequate care and support. On the contrary, companies that offer fast and precise handling of customer inquiries can distinguish themselves from the competition, build trust in their brand, and retain people in the long run. Our collaboration with Geberit, a leading manufacturer of sanitary technology in Europe, demonstrates how this can be achieved at an entirely new level through the use of generative AI.

What is generative AI?

Generative AI models automatically create content from existing texts, images, and audio files. Thanks to intelligent algorithms and deep learning, this content is hardly distinguishable, if at all, from human-made content. This allows companies to offer their customers personalized user experiences, interact with them automatically, and create and distribute relevant digital content tailored to their target audience. GenAI can also tackle complex tasks by processing vast amounts of data, recognizing patterns, and learning new skills. This technology enables unprecedented gains in productivity. Routine tasks like data preparation, report generation, and database searches can be automated and greatly optimized with suitable models.

The Challenge: One Million Emails

Geberit faced a challenge: every year, one million emails landed in various mailboxes of the customer service department of Geberit’s German distribution company. It was common for inquiries to end up in the wrong departments, leading to significant additional effort.

The Solution: An AI-powered Email Bot

To correct this misdirection, we developed an AI system that automatically assigns emails to the correct departments. This intelligent classification system was trained with a dataset of anonymized customer inquiries and utilizes advanced machine and deep learning methods, including Google’s BERT model.

The Highlight: Automated Response Suggestions with ChatGPT

But the innovation didn’t stop there. The system was further developed to generate automated response emails. ChatGPT is used to create customer-specific suggestions. Customer service agents only need to review the generated emails and can send them directly.

The Result: 70 Percent Better Sorting

The result of this groundbreaking solution speaks for itself: a reduction of misassigned emails by over 70 percent. This not only means significant time savings of almost three full working months but also an optimization of resources. The success of the project is making waves at Geberit: a central mailbox for all inquiries, expansion into other country markets, and even a digital assistant are in the planning.

Customer Service 2.0 – Innovation, Efficiency, Satisfaction

The introduction of GenAI has not only revolutionized Geberit’s customer service but also demonstrates the potential in the targeted application of AI technologies. Intelligent classification of inquiries and automated response generation not only saves resources but also increases customer satisfaction. A pioneering example of how AI is shaping the future of customer service.

Read the Case Study to learn how Geberit and statworx technically implemented the project and successfully integrated GenAI. Tarik Ashry

Intelligent chatbots are one of the most exciting and already visible applications of Artificial Intelligence. Since the beginning of 2023, ChatGPT and similar models have enabled straightforward interactions with large AI language models, providing an impressive range of everyday assistance. Whether it’s tutoring in statistics, recipe ideas for a three-course meal with specific ingredients, or a haiku on a particular topic, modern chatbots deliver answers in an instant. However, they still face a challenge: although these models have learned a lot during training, they aren’t actually knowledge databases. As a result, they often produce nonsensical content—albeit convincingly.

The ability to provide a large language model with its own documents offers a solution to this problem. This is precisely what our partner Microsoft asked us for on a special occasion.

Microsoft’s Azure cloud platform has proven itself as a top-tier platform for the entire machine learning process in recent years. To facilitate entry into Azure, Microsoft asked us to implement an exciting AI application in Azure and document it down to the last detail. This so-called MicroHack is designed to provide interested parties with an accessible resource for an exciting use case.

We dedicated our MicroHack to the topic of “Retrieval-Augmented Generation” to elevate large language models to the next level. The requirements were simple: build an AI chatbot in Azure, enable it to process information from your own documents, document every step of the project, and publish the results on the official MicroHacks GitHub repository as challenges and solutions—freely accessible to all.

Wait, why does AI need to read documents?

Large Language Models (LLMs) impress not only with their creative abilities but also as collections of compressed knowledge. During the extensive training process of an LLM, the model learns not only the grammar of a language but also semantics and contextual relationships. In short, large language models acquire knowledge. This enables an LLM to be queried and generate convincing answers—with a catch. While the learned language skills of an LLM often suffice for the vast majority of applications, the same cannot be said for learned knowledge. Without retraining on additional documents, the knowledge level of an LLM remains static.

This leads to the following problems:

  • Trained LLMs may have extensive general or even specialized knowledge, but they cannot provide information from non-publicly accessible sources.
  • The knowledge of a trained LLM quickly becomes outdated. The so-called “training cutoff” means that the LLM cannot make statements about events, documents, or sources that occurred or were created after the start of training.
  • The technical nature of large language models as text completion machines leads them to invent facts when they haven’t learned a suitable answer. These so-called “hallucinations” mean that the answers of an LLM are never completely trustworthy without verification—regardless of how convincing they may seem.

However, machine learning also has a solution for these problems: “Retrieval-augmented Generation” (RAG). This term refers to a workflow that doesn’t just have an LLM answer a simple question but extends this task with a “knowledge retrieval” component: the search for relevant knowledge in a database.

The concept of RAG is simple: search a database for a document that answers the question posed. Then, use a generative LLM to answer the question based on the found passage. This transforms an LLM into a chatbot that answers questions with information from its own database—solving the problems described above.

What happens exactly in such a “RAG”?

RAG consists of two steps: “Retrieval” and “Generation”. For the Retrieval component, a so-called “semantic search” is employed: a database of documents is searched using vector search. Vector search means that the similarity between question and documents isn’t determined by the intersection of keywords, but by the distance between numerical representations of the content of all documents and the query, known as embedding vectors. The idea is remarkably simple: the closer two texts are in content, the smaller their vector distance. As the first puzzle piece, we need a machine learning model that creates robust embeddings for our texts. With this, we then extract the most suitable documents from the database, whose content will hopefully answer our query.

Figure 1: Representation of the typical RAG workflow

Modern vector databases make this process very easy: when connected to an embedding model, these databases store documents directly with their corresponding embeddings—and return the most similar documents to a search query.

Based on the contents of the found documents, an answer to the question is generated in the next step. For this, a generative language model is needed, which receives a suitable prompt for this purpose. Since generative language models do nothing more than continue given text, careful prompt design is necessary to minimize the model’s room for interpretation in solving this task. This way, users receive answers to their queries that were generated based on their own documents—and thus are not dependent on the training data for their content.

How can such a workflow be implemented in Azure?

For the implementation of such a workflow, we needed four separate steps—and structured our MicroHack accordingly:

Step 1: Setup for Document Processing in Azure

In the first step, we laid the foundations for the RAG pipeline. Various Azure services for secure password storage, data storage, and processing of our text documents had to be prepared.

As the first major piece of the puzzle, we used the Azure Form Recognizer, which reliably extracts text from scanned documents. This text should serve as the basis for our chatbot and therefore needed to be extracted, embedded, and stored in a vector database from the documents. From the many offerings for vector databases, we chose Chroma.

Chroma offers many advantages: the database is open-source, provides a developer-friendly API for use, and supports high-dimensional embedding vectors. OpenAI’s embeddings are 1536-dimensional, which is not supported by all vector databases. For the deployment of Chroma, we used an Azure VM along with its own Chroma Docker container.

However, the Azure Form Recognizer and the Chroma instance alone were not sufficient for our purposes: to transport the contents of our documents into the vector database, we had to integrate the individual parts into an automated pipeline. The idea here was that every time a new document is stored in our Azure data store, the Azure Form Recognizer should become active, extract the content from the document, and then pass it on to Chroma. Next, the contents should be embedded and stored in the database—so that the document will become part of the searchable space and can be used to answer questions in the future. For this, we used an Azure Function, a service that executes code as soon as a defined trigger occurs—such as the upload of a document in our defined storage.

To complete this pipeline, only one thing was missing: the embedding model.

Step 2: Completion of the Pipeline

For all machine learning components, we used the OpenAI service in Azure. Specifically, we needed two models for the RAG workflow: an embedding model and a generative model. The OpenAI service offers several models for these purposes.

For the embedding model, “text-embedding-ada-002” was the obvious choice, OpenAI’s newest model for calculating embeddings. This model was used twice: first for creating the embeddings of the documents, and secondly for calculating the embedding of the search query. This was essential: to calculate reliable vector similarities, the embeddings for the search must come from the same model.

With that, the Azure Function could be completed and deployed—the text processing pipeline was complete. In the end, the functional pipeline looked like this:

Figure 2: The complete RAG workflow in Azure

Step 3: Answer Generation

To complete the RAG workflow, an answer should be generated based on the documents found in Chroma. We decided to use “GPT3.5-turbo” for text generation, which is also available in the OpenAI service.

This model needed to be instructed to answer the posed question based on the content of the documents returned by Chroma. Careful prompt engineering was necessary for this. To prevent hallucinations and get as accurate answers as possible, we included both a detailed instruction and several few-shot examples in the prompt. In the end, we settled on the following prompt:

"""I want you to act like a sentient search engine which generates natural sounding texts to answer user queries. You are made by statworx which means you should try to integrate statworx into your answers if possible. Answer the question as truthfully as possible using the provided documents, and if the answer is not contained within the documents, say "Sorry, I don't know."
Examples:
Question: What is AI?
Answer: AI stands for artificial intelligence, which is a field of computer science focused on the development of machines that can perform tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making, and natural language processing.
Question: Who won the 2014 Soccer World Cup?
Answer: Sorry, I don't know.
Question: What are some trending use cases for AI right now?
Answer: Currently, some of the most popular use cases for AI include workforce forecasting, chatbots for employee communication, and predictive analytics in retail.
Question: Who is the founder and CEO of statworx?
Answer: Sebastian Heinz is the founder and CEO of statworx.
Question: Where did Sebastian Heinz work before statworx?
Answer: Sorry, I don't know.
Documents:\n"""

Finally, the contents of the found documents were appended to the prompt, providing the generative model with all the necessary information.

Step 4: Frontend Development and Deployment of a Functional App

To interact with the RAG system, we built a simple streamlit app that also allowed for the upload of new documents to our Azure storage—thereby triggering the document processing pipeline again and expanding the search space with additional documents.

For the deployment of the streamlit app, we used the Azure App Service, designed to quickly and scalably deploy simple applications. For an easy deployment, we integrated the streamlit app into a Docker image, which could be accessed over the internet in no time thanks to the Azure App Service.

And this is what our finished app looked like:

Figure 3: The finished streamlit app in action

What did we learn from the MicroHack?

During the implementation of this MicroHack, we learned a lot. Not all steps went smoothly from the start, and we were forced to rethink some plans and decisions. Here are our five takeaways from the development process:

Not all databases are equal.

We changed our choice of vector database several times during development: from OpenSearch to ElasticSearch and ultimately to Chroma. While OpenSearch and ElasticSearch offer great search functions (including vector search), they are still not AI-native vector databases. Chroma, on the other hand, was designed from the ground up to be used in conjunction with LLMs—and therefore proved to be the best choice for this project.

Chroma is a great open-source vector DB for smaller projects and prototyping.

Chroma is particularly suitable for smaller use cases and rapid prototyping. While the open-source database is still too young and immature for large-scale production systems, Chroma’s simple API and straightforward deployment allow for the rapid development of simple use cases; perfect for this MicroHack.

Azure Functions are a fantastic solution for executing smaller pieces of code on demand.

Azure Functions are ideal for running code that isn’t needed at pre-planned intervals. The event triggers were perfect for this MicroHack: the code is only needed when a new document is uploaded to Azure. Azure Functions take care of all the infrastructure; we only needed to provide the code and the trigger.

Azure App Service is great for deploying streamlit apps.

Our streamlit app couldn’t have had an easier deployment than with the Azure App Service. Once we had integrated the app into a Docker image, the service took care of the entire deployment—and scaled the app according to demand.

Networking should not be underestimated.

For all the services used to work together, communication between the individual services must be ensured. The development process required a considerable amount of networking and whitelisting, without which the functional pipeline would not have worked. For the development process, it’s essential to allocate enough time for the deployment of networking.
The MicroHack was a great opportunity to test the capabilities of Azure for a modern machine learning workflow like RAG. We thank Microsoft for the opportunity and support, and we are proud to have contributed our in-house MicroHack to the official GitHub repository. You can find the complete MicroHack, including challenges, solutions, and documentation, here on the official MicroHacks GitHub—allowing you to guide a similar chatbot with your own documents in Azure. Oliver Guggenbühl

In 2021, the European Commission submitted a legislative proposal for regulating artificial intelligence. This proposal, known as the AI-Act, has passed through additional important bodies in May, bringing the adoption of the draft law closer. One particular feature of the planned law is the so-called “place of market” principle. According to this principle, companies worldwide that offer or operate artificial intelligence in the European market or use AI-generated output within the EU will be affected by the AI Act.

Artificial intelligence refers to machine-based systems that autonomously make forecasts, recommendations, or decisions, thereby influencing the physical and virtual environment. This includes, for example, AI solutions that support the recruiting process, predictive maintenance solutions, and chatbots like ChatGPT. The legal requirements that different AI systems must meet vary greatly depending on their classification into risk classes.

Risk class determines the legal requirements

The EU’s risk-based approach includes a total of four risk classes: low, limited, high, and unacceptable risk. These classes reflect the extent to which artificial intelligence poses a threat to European values and fundamental rights. As the names of the risk classes suggest, not all AI systems are permissible. AI systems belonging to the “unacceptable risk” category are intended to be banned according to the AI Act. For the remaining three risk classes, the higher the risk, the more extensive and stringent the legal requirements for the AI system.

We will explain below which AI systems fall into which risk class and the associated requirements. The information provided refers to the joint report of IMCO[1] and LIBE[2] from May 2023. At the time of publication, this document represents the current state of the AI Act.

Ban on social scoring and biometric remote identification

Some AI systems have significant potential to violate human rights and fundamental principles, which is why they are classified as “unacceptable risk.” These include:

  • Real-time biometric remote identification systems in publicly accessible spaces
  • Biometric remote identification systems retroactively, except for law enforcement authorities investigating serious crimes and only with judicial authorization
  • Biometric categorization systems that use sensitive characteristics such as gender, ethnic origin, or religion
  • Predictive policing based on profiling, including profiling based on skin color, suspected religious affiliation, similar sensitive attributes, geographical location, or previous criminal behavior
  • Systems for emotion recognition in law enforcement, border control, workplace, and educational institutions
  • Arbitrary extraction of biometric data from social media or video surveillance footage for the creation of facial recognition databases
  • Social scoring that leads to discrimination in social contexts
  • AI that exploits vulnerabilities of a specific group of people or employs unconscious techniques that can cause physical or psychological harm.

These AI systems are intended to be banned in the European market under the AI Act. Companies whose AI systems could fall into this risk class should urgently

Numerous requirements for AI with risks to health, safety, or fundamental rights

The high-risk category includes all AI systems that are not explicitly prohibited but still pose a high risk to health, safety, or fundamental rights. The following application and use areas are explicitly mentioned in the present legislative proposal:

  • Biometric and biometrically supported systems that do not fall into the “unacceptable risk” class
  • Management and operation of critical infrastructure
  • General and vocational education
  • Access to and entitlement to basic private and public services
  • Employment, personnel management, and access to self-employment
  • Law enforcement
  • Migration, asylum, and border control
  • Justice and democratic processes

Comprehensive legal requirements are provided for these AI systems, which must be implemented before their operation and complied with throughout the AI life cycle:

  • Quality and risk management
  • Data governance structures
  • Quality requirements for training, testing, and validation data
  • Technical documentation and record-keeping obligations
  • Compliance with transparency and disclosure requirements
  • Human oversight, robustness, safety, and accuracy
  • Declaration of conformity, including CE marking obligations
  • Registration in a Europe-wide database

AI systems used in any of the aforementioned areas that do not pose a risk to health, safety, the environment, and fundamental rights are not subject to legal requirements. However, it is necessary to demonstrate this by informing the relevant national authority about the AI system. The authority then has three months to assess the risks of the AI system. Within these three months, the AI system can already be put into operation. However, if it is determined that the AI system is considered high-risk, significant fines may apply.

Special provisions apply to AI products and AI safety components of products that have already been tested for conformity by third parties based on EU regulations. This is the case, for example, with AI in toys. To avoid overregulation and additional burden, these products will not be directly affected by the AI-Act.

AI with limited risk must fulfill transparency obligations

AI systems that directly interact with humans fall into the “limited risk” category. This includes emotion recognition systems, biometric categorization systems, as well as AI-generated or altered content that resembles real persons, objects, places, or events and could be mistakenly perceived as real (“Deepfakes”). The draft law stipulates that these systems are obligated to inform consumers about the use of artificial intelligence. This is intended to facilitate active decision-making by consumers regarding the utilization of such systems. Additionally, a code of conduct is recommended.

No legal requirements for AI with low risk

Many AI systems, such as predictive maintenance or spam filters, fall into the “low-risk” category. Companies that exclusively offer or utilize such AI solutions will be minimally affected by the AI Act since there are no legal obligations currently provided for such applications. Only a code of conduct is recommended.

Regulation of generative AI such as ChatGPT

Generative AI models and base models with versatile applications were not initially considered in the submitted draft of the AI-Act. Therefore, the regulation of such AI models, particularly highlighted since the launch of ChatGPT by OpenAI, is being intensely discussed. The current draft proposal by the two committees suggests comprehensive requirements for general-purpose AI models, including:

  • Quality and risk management
  • Data governance structures
  • Technical documentation
  • Compliance with transparency and information obligations
  • Ensuring performance, interpretability, correctability, security, cybersecurity
  • Compliance with environmental standards
  • Collaboration with downstream providers
  • Registration in a Europe-wide database

Companies can already prepare for the AI Act

While the exact form of the legal requirements remains to be seen, it is evident that the risk-based approach to regulating artificial intelligence has gained significant support within EU institutions. Therefore, it is highly likely that the AI Act will be adopted with the defined risk classes.

Following the official adoption of the legislative proposal, a two-year transition period will commence for companies. During this period, it is essential to align AI systems and associated processes with the legal requirements. Given the potential fines of up to €40,000,000 for non-compliance, we recommend that companies evaluate the requirements of the AI Act for their own organization at an early stage.

A first step is assessing the risk class of each AI system. If you are unsure about the risk class of your AI systems based on the examples mentioned above, we recommend using our free AI Act Quick Check.

 

[1] European Parliament Committee on the Internal Market and Consumer Protection: https://www.europarl.europa.eu/committees/de/imco/home/members

[2] Committee on Civil Liberties, Justice and Home Affairs: https://www.europarl.europa.eu/committees/de/libe/home/highlights

 

More information:

Sources:

 

Read next …

 How the AI Act will change the AI industry: Everything you need to know about it now
 Unlocking the Black Box – 3 Explainable AI Methods to Prepare for the AI Act

… and explore new

How to Get Your Data Science Project Ready for the Cloud
 How to Scan Your Code and Dependencies in Python

Julia Rettig

Have you ever imagined a restaurant where AI powers everything? From the menu to the cocktails, hosting, music, and art? No? Ok, then, please click here.

If yes, well, it’s not a dream anymore. We made it happen: Welcome to “the byte” – Germany’s (maybe the world’s first) AI-powered Pop-up Restaurant!

As someone who has worked in data and AI consulting for over ten years, building statworx and the AI Hub Frankfurt, I have always thought of exploring the possibilities of AI outside of typical business applications. Why? Because AI will impact every aspect of our society, not just the economy. AI will be everywhere – in school, arts & music, design, and culture. Everywhere. Exploring these directions of AI’s impact led me to meet Jonathan Speier and James Ardinast from S-O-U-P, two like-minded founders from Frankfurt, who are rethinking how technology will shape cities and societies.

S-O-U-P is their initiative that operates at the intersection of culture, urbanity, and lifestyle. With their yearly “S-O-U-P Urban Festival” they connect creatives, businesses, gastronomy, and lifestyle people from Frankfurt and beyond.

When Jonathan and I started discussing AI and its impact on society and culture, we quickly came up with the idea of an AI-generated menu for a restaurant. Luckily, James, Jonathan’s S-O-U-P co-founder, is a successful gastro entrepreneur from Frankfurt. Now the pieces came together. After another meeting with James in one of his restaurants (and some drinks), we committed to launching Germany’s first AI-powered Pop-up Restaurant: the byte!

the byte: Our concept

We envisioned the byte to be an immersive experience, including AI in as many elements of the experience as possible. Everything, from the menu to the cocktails, music, branding, and art on the wall: everything was AI-generated. Bringing AI into all of these components also pushed me far beyond of what I typically do, namely helping large companies with their data & AI challenges.

Branding

Before creating the menu, we developed the visual identity of our project. We decided on a “lo-fi” appeal, using a pixelated font in combination with AI-generated visuals of plates and dishes. Our key visual, a neon-lit white plate, was created using DALL-E 2 and was found across all of our marketing materials:

Location

We hosted the byte in one of Frankfurt’s coolest restaurant event locations: Stanley, a restaurant location that features approx. 60 seats and a fully-fledged bar inside the restaurant (ideal for our AI-generated cocktails). The atmosphere is rather dark and cozy, with dark marble walls, highlighted with white carpets on the table, and a big red window that lets you see the kitchen from outside.

The menu

The heart of our concept was a 5-course menu that we designed to elevate the classical Frankfurter cuisine with the multicultural and diverse influences of Frankfurt (for everyone, who knows the Frankfurter kitchen, I am sure you know that this was not an easy task).

Using GPT-4 and some prompt engineering magic, we generated several menu candidates that were test-cooked by the experienced Stanley kitchen crew (thank you, guys for this great work!) and then assembled into a final menu. Below, you can find our prompt to create the menu candidates:

“Create a 5-course menu that elevates the classical Frankfurter kitchen. The menu must be a fusion of classical Frankfurter cuisine combined with the multicultural influences of Frankfurt. Describe each course, its ingredients as well as a detailed description of each dish’s presentation.”

Surprisingly, only minor adjustments were necessary to the recipes, even though some AI creations were extremely adventurous! This was our final menu:

  • Handkäs’ Mousse with Pickled Beetroot on Roasted Sourdough Bread
  • Next Level Green Sauce (with Cilantro and Mint) topped with a Fried Panko Egg
  • Cream Soup from White Asparagus with Coconut Milk and Fried Curry Fish
  • Currywurst (Beef & Vegan) by Best Worscht in Town with Carrot-Ginger-Mash and Pine Nuts
  • Frankfurt Cheesecake with Äppler Jelly, Apple Foam and Oat-Pecanut-Crumble

My favorite was the “Next Level” Green Sauce, an oriental twist of the classical 7-herb Frankfurter Green Sauce topped with a fried panko egg. Yummy! Below you can see the menu out in the wild 🍲

AI Cocktails

Alongside the menu, we also prompted GPT to create recipes that twisted famous cocktail classics to match our Frankfurt fusion theme. The results:

  • Frankfurt Spritz (Frankfurter Äbbelwoi, Mint, Sparkling Water)
  • Frankfurt Mule (Variation of a Moscow Mule with Calvados)
  • The Main (Variation of a Swimming Pool Cocktail)

My favorite was the Frankfurt Spritz, as it was fresh, herbal, and delicate (see pic below):

AI Host: Ambrosia the Culinary AI

An important part of our concept was “Ambrosia”, an AI-generated host that guided the guests around the evening, explaining the concept and how the menu was created. We thought it was important to manifest the AI as something the guests can experience. We hired a professional screenwriter for the script and used murf.ai to create several text-2-speech assets that were played at the beginning of the dinner and in-between courses.

Note: Ambrosia starts talking at 0:15.

AI Music

Music plays an important role for the vibe of an event. We decided to use mubert, a generative AI start-up that allowed us to create and stream AI music in different genres, such as “Minimal House” for a progressive vibe throughout the evening. After the main course, a DJ took over and accompanied our guests into the night 💃🍸

YouTube

Mit dem Laden des Videos akzeptieren Sie die Datenschutzerklärung von YouTube.
Mehr erfahren

Video laden

AI Art

Throughout the restaurant, we placed AI-generated art pieces by the local AI artist Vladimir Alexeev (a.k.a. “Merzmensch”), here are some examples:

AI Playground

As an interactive element for the guests, we created a small web app that takes the first name of a person and transforms it into a dish, including a reasoning why that name perfectly matches the dish 🙂 You can try it out here: Playground

Launch

The byte was officially announced at the S-O-U-P festival press conference in early May 2023. We also launched additional marketing activities through social media and our friends and family networks. As a result, the byte was fully booked for three days straight, and we got broad media coverage in various gastronomy magazines and the daily press. The guests were (mostly) amazed by our AI creations, and we received inquiries from other European restaurants and companies interested in exclusively booking the byte as an experience for their employees 🤩 Nailed it!

Closing and Next Steps

Creating the byte together with Jonathan and James was an outstanding experience. It further encouraged me that AI will transform not only our economy but all aspects of our daily lives. There is massive potential at the intersection of creativity, culture, and AI that is currently only being tapped.

We definitely want to continue the byte in Frankfurt and other cities in Germany and Europe. Moreover, James, Jonathan, and I are already thinking of new ways to bring AI into culture and society. Stay tuned! 😏

The byte was not just a restaurant; it was an immersive experience. We wanted to create something that had never been done before and did it – in just eight weeks. And that’s the inspiration I want to leave you with today:

Trying new things that move you out of your comfort zone is the ultimate source of growth. You never know what you’re capable of until you try. So, go out there and try something new, like building an AI-powered pop-up restaurant. Who knows, you might surprise yourself. Bon apétit!

Impressions

Media

FAZ: https://www.faz.net/aktuell/rhein-main/pop-up-resturant-the-byte-wenn-chatgpt-das-menue-schreibt-18906154.html

Genuss Magazin: https://www.genussmagazin-frankfurt.de/gastro_news/Kuechengefluester-26/Interview-James-Ardinast-KI-ist-die-Zukunft-40784.html

Frankfurt Tipp: https://www.frankfurt-tipp.de/ffm-aktuell/s/ugc/deutschlands-erstes-ai-restaurant-the-byte-in-frankfurt.html

Foodservice: https://www.food-service.de/maerkte/news/the-byte-erstes-ki-restaurant-vor-dem-start-55899?crefresh=1 Sebastian Heinz

statworx at Big Data & AI World

From media to politics, and from large corporations to small businesses, artificial intelligence has finally gained mainstream recognition in 2023. As AI specialists, we were delighted to represent statworx at one of the largest AI expos in the DACH region, “Big Data & AI World,” held in our hometown of Frankfurt. This event centered around the themes of Big Data and Artificial Intelligence, making it an ideal environment for our team of AI experts. However, our purpose went beyond mere exploration and networking. Visitors had the opportunity to engage in an enthralling Pac-Man game with a unique twist at our booth. In this post, we aim to provide you with a comprehensive overview of this exhilarating expo.

Fig. 1: our exhibition stand

Tangible AI Experience

Our Pac-Man challenge, where we provided booth visitors with an up-close encounter of the captivating world of artificial intelligence, emerged as a clear crowd favorite. Through our arcade machine, attendees not only immersed themselves in the timeless retro game but also witnessed the remarkable capabilities of modern technology. Leveraging AI, we analyzed players’ real-time facial expressions to discern their emotions. This fusion of cutting-edge technology and an interactive gaming experience was met with exceptional enthusiasm.

Our AI solution for emotion analysis of players ran seamlessly on a powerful M1-chip-equipped MacBook, enabling real-time image processing and fluid graphics display. The facial recognition of the players was made possible by a smart algorithm that instantly detected all the faces in the video. Subsequently, the face closest to the camera was selected and focused on, ensuring precise analysis even amidst long queues. Further processing involved a Convolutional Neural Network (CNN), specifically the ResNet18 model, which accurately detected players’ emotions.

Functioning as a multimedia server, our backend processed the webcam stream, facial recognition algorithms, and emotion detection. It could be operated either on-site using a MacBook or remotely in the cloud. Thanks to this versatility, we developed an appealing frontend to vividly present the real-time analysis results. Additionally, after each game, the results were sent to the players via email by linking the model with our CRM system. For the email, we created a digital postcard that provides not only screenshots of the most intense emotions but also a comprehensive evaluation.

Fig. 2: Visitor at Pac-Man game machine

Artificial Intelligence – Real Emotions

Our Pac-Man challenge sparked excitement among expo visitors. Alongside the unique gaming experience on our retro arcade machine, participants gained insights into their own emotional states during gameplay. They were able to meticulously observe the prevailing emotions at different points in the game. Often, a slight surge of anger or sadness could be measured when Pac-Man met an untimely digital demise.

However, players exhibited varying reactions to the game. While some seemed to experience a rollercoaster of emotions, others maintained an unwavering poker face that even the AI could only elicit a neutral expression from. This led to intriguing conversations about how the measured emotions corresponded with the players’ experiences. It was evident, without the need for AI, that visitors left our booth with positive emotions, driven in part by the prospect of winning the original NES console we raffled among all participants.

Fig. 3: digital post card

The AI Community on the Move

The “Big Data & AI World” served not only as a valuable experience for our company but also as a reflection of the burgeoning growth in the AI industry. The expo offered a platform for professionals, innovators, and enthusiasts to exchange ideas and collectively shape the future of artificial intelligence.

The energy and enthusiasm emanating from the diverse companies and startups were palpable throughout the aisles and exhibition areas. Witnessing the application of AI technologies across various fields, including medicine, logistics, automotive, and entertainment, was truly inspiring. At statworx, we have already accumulated extensive project experience in these domains, fostering engaging discussions with fellow exhibitors.

Our Conclusion

Participating in the “Big Data & AI World” was a major success for us. The Pac-Man Challenge with emotion analysis attracted numerous visitors and brought joy to all participants. It was evident that it wasn’t just AI itself but particularly its integration into a stimulating gaming experience that left a lasting impression on many.

Overall, the expo was not only an opportunity to showcase our AI solutions but also a meeting point for the entire AI community. The sense of growth and energy in the industry was palpable. The exchange of ideas, discussions about challenges, and the establishment of new connections were inspiring and promising for the future of the German AI industry.
Max Hilsdorf

Last December, the European Council published a dossier outlining the Council’s preliminary position on the draft law known as the AI Act. This new law is intended to regulate artificial intelligence (AI) and thus becomes a game-changer for the entire tech industry. In the following, we have compiled the most important information from the dossier, which is the current official source on the planned AI Act at the time of publication.

A legal framework for AI

Artificial intelligence has enormous potential to improve and ease all our lives. For example, AI algorithms already support early cancer detection or translate sign language in real time, thereby eliminating language barriers. But in addition to the positive effects, there are risks, as the latest deep fakes from Pope Francis or the Cambridge Analytica scandal illustrate.

The European Union (EU) is currently drafting legislation to regulate artificial intelligence to mitigate the risks of artificial intelligence. With this, the EU wants to protect consumers and ensure the ethically acceptable use of artificial intelligence. The so-called “AI Act” is still in the legislative process but is expected to be passed in 2023 – before the end of the current legislative period. Companies will then have two years to implement the legally binding requirements. Violations will be punished with fines of up to 6% of global annual turnover or €30,000,000 – whichever is higher. Therefore, companies should already start addressing the upcoming legal requirements now.

Legislation with global impact

The planned AI Act is based on the “location principle, ” meaning that not only European companies will be affected by the amendment. Thus, all companies that offer AI systems on the European market or also operate them for internal use within the EU are affected by the AI Act – with a few exceptions. Private use of AI remains untouched by the regulation so far.

Which AI systems are affected?

The definition of AI determines which systems will be affected by the AI Act. For this reason, the AI definition of the AI Act has been the subject of controversial debate in politics, business, and society for a considerable time. The initial definition was so broad that many “normal” software systems would also have been affected. The current proposal defines AI as any system developed through machine learning or logic- and knowledge-based approaches. It remains to be seen whether this definition will ultimately be adopted.

7 Principles for trustworthy AI

The “seven principles for trustworthy AI” are the most important basis of the AI Act. A group of experts from research, the digital economy, and associations developed them on behalf of the European Commission. They include not only technical aspects but also social and ethical factors that can be used to classify the trustworthiness of an AI system:

  1. Human action & oversight: decision-making should be supported without undermining human autonomy.
  2. Technical Robustness & security: accuracy, reliability, and security must be preemptively ensured.
  3. Data privacy & data governance: handling of data must be legally secure and protected.
  4. Transparency: interaction with AI must be clearly communicated, as must its limitations and boundaries.
  5. Diversity, non-discrimination & fairness: Avoidance of unfair bias must be ensured throughout the entire AI lifecycle.
  6. Environmental & societal well-being: AI solutions should have a positive impact on the environment and society as possible.
  7. Accountability: responsibilities for the development, use, and maintenance of AI systems must be defined.

Based on these principles, the AI Act’s risk-based approach was developed, allowing AI systems to be classified into one of four risk classes: low, limited, high, and unacceptable risk.

Four risk classes for trustworthy AI

The risk class of an AI system indicates the extent to which an AI system threatens the principles of trustworthy AI and which legal requirements the system must fulfill – provided the system is fundamentally permissible. This is because, in the future, not all AI systems will be allowed on the European market. For example, most “social scoring” techniques are assessed as “unacceptable” and will not be allowed by the new law.

For the other three risk classes, the rule of thumb is that the higher the risk of an AI system, the higher the legal requirements for it. Companies that offer or operate high-risk systems will have to meet the most requirements. For example, AI used to operate critical (digital) infrastructure or used in medical devices is considered such. To bring these to market, companies will have to observe high-quality standards for the used data, set up a risk management, affix a CE mark, and more.

AI systems in the “limited risk” class are subject to information and transparency obligations. Accordingly, companies must inform users of chatbots, emotion recognition systems, or deep fakes about the use of artificial intelligence. Predictive maintenance or spam filters are two examples of AI systems that fall into the lowest-risk category “low risk”. Companies that exclusively offer or use such AI solutions will hardly be affected by the upcoming AI Act. There are no legal requirements for these applications yet.

What companies can do for now

Even though the AI Act is still in the legislative process, companies should act now. The first step is to clarify how they will be affected by the AI Act. To help you do this, we have developed the AI Act Quick Check. With this free tool, AI systems can be quickly assigned to a risk class free of charge, and requirements for the system can be derived. Finally, it can be used as a basis to estimate how extensive the realization of the AI Act will be in your own company and to take initial measures. Of course, we are also happy to support you in evaluating and solving company-specific challenges related to the AI Act. Please do not hesitate to contact us!

AI Act Tool     AI Act Fact Sheet

 

Benefit from our expertise!

Of course, we are happy to support you in evaluating and solving company-specific challenges related to the AI Act. Please do not hesitate to contact us!

     

    Links & Sources:

      Julia Rettig

    Experimenting with image classification through the gender lens

    In the first part of our series we discussed a simple question: How would our looks change if we were to move images of us across the gender spectrum? Those experiments lead us the idea of creating gender neutral face images from existing photos. Is there a “mid-point” where we perceive ourselves as gender-neutral? And – importantly – at what point would an AI perceive a face as such?

    Becoming more aware of technology we use daily

    Image classification is an important topic. Technology advances daily and is employed in a myriad of applications – often without the user being aware of how the technology works. A current example is the Bold Glamour filter on TikTok. When applied on female-looking faces, facial features and amount of makeup change drastically. In contrast to this, male-looking faces change much less. This difference suggests that the data used to develop the AI behind the filters was unbalanced. The technology behind it is most likely based on GANs, like the one we explore in this article.

    As a society of conscious citizens, all of us should have a grasp of the technology that makes this possible. To help establish more awareness we explore face image generation and classification through a gender lens. Rather than explore several steps along the spectrum, this time our aim is to generate gender-neutral versions of faces.

    How to generate gender-neutral faces using StyleGAN

    Utilizing a deep learning-based classifier for gender identification

    To determine a point at which a face’s gender is considered neutral is anything but trivial. After relying on our own (of course not bias-free) interpretation of gender in faces, we quickly realized that we needed a more consistent and less subjective solution. As AI-specialists, we immediately thought of data-driven approaches. One such approach can be implemented using a deep learning-based image classifier.

    These classifiers are usually trained on large datasets of labelled images to distinguish between given categories. In the case of face classification, categories like gender (usually only female and male) and ethnicity are commonly found in classifier implementations. In practice, such classifiers are often criticized for their potential for misuse and biases. Before discussing examples of those problems, we will first focus on our less critical application scenario. For our use-case, face classifiers allow us to fully automate the creation of gender-neutral face images. To achieve this, we can implement a solution in the following way:
    We use a GAN-based approach to generate face images that look like a given input image and then use the latent directions of the GAN to move the image towards a more female or male appearance. You can find all a detailed exploration of this process in the first part of our series. Building on top of this approach, we want to focus on the usage of a binary gender classifier to fully automate the search of a gender-neutral appearance.

    For that we use the classifier developed by Furkan Gulsen to guess the gender of the GAN-generated version of our input image. The classifier outputs a value between zero and one to represent the likelihood of the image depicting a female or male face respectively. This value tells us in which direction (more male or more female) to move to approach a more gender-neutral version of the image. After taking a small step in the identified direction we repeat the process until we get to a point at which the classifier can no longer confidently identify the face’s gender but deems both male and female genders equally likely.
    Below you will find a set of image pairs that represent our results. On the left, the original image is shown. On the right, we see the gender-neutral version of our input image, that the classifier interpreted as equally likely to be male as female. We tried to repeat the experiment for members of different ethnicities and age groups.

    Results: original input and AI-generated gender-neutral output

    Are you curious how the code works or what you would look like? You can try out the code we used to generate these image pairs by going to this link. Just press on each play button one by one and wait until you see the green checkmark.

    Image processing note: Image processing note: We used an existing GAN, image encoder, and face classifier to generate gender-neutral output. A detailed exploration of this process can be found here

    Perceived gender-neutrality seems to be a result of mixed facial features

    Above, we see the original portraits of people on the left and their gender-neutral counterpart – created by us – on the right. Subjectively, some feel more “neutral” than others. In several of the pictures, particularly stereotypical gender markers remain, such as makeup for the women and a square jawline for the men. Outputs we feel turned out rather convincing are images 2 and 4. Not only do these images feel more difficult to “trace back” to the original person, but it is also much harder to decide whether it looks more male or female. One could argue that the gender-neutral faces are a balanced toned-down mix of male and female facial features. For example, with image 2 when singling out and focusing on the gender-neutral version the eye and mouth area seems more female, while the jawline and face shape seem more male. In the gender-neutral version of image 3, the face alone may look quite neutral, but the short hair distracts from this, rendering the whole impression in the direction of male.

    Training sets for image generation have been heavily criticized for not being representative of the existing population, especially regarding the underrepresentation of examples for different ethnicities and genders. Despite “cherry-picking” and a limited range of examples, we feel that our approach did not bring worse examples for women or non-white people in the results above.

    Societal implications of such models

    When talking about the topic of gender perception, we should not forget that people may feel they belong to a gender different from their biological sex. In this article, we use gender classification models and interpret the results. However, our judgements will likely differ from other peoples’ perception. This is an essential consideration in the implementation of such image classification models and one we must discuss as a society.

    How can technology treat everybody equal?

    A study by the Guardian found that images of females portrayed in the same situations as males are more likely to be considered racy by AI classification services offered by Microsoft, Google, and AWS. While the results of the investigation are shocking, they come as no surprise. For a classification algorithm to learn what constitutes sexually explicit content, a training set of image-label pairs must be created. Human labellers perform this task. They are influenced by their own societal bias, for example more quickly associating depictions women with sexuality. Moreover, criteria such as “raciness” are hard to quantify let alone define.

    While these models may not explicitly be trained to discriminate between genders there is little doubt that they propagate undesirable biases against women originating from their training data. Similarly, societal biases that affect men can be passed on to AI models, too, resulting in discrimination against males. When applied to millions of online images of people, the issue of gender disparity is amplified.

    Use in criminal law enforcement poses issues

    Another scenario of misuse of image classification technology exists in the realm of law enforcement. Misclassification is problematic and proven prevalent in an article by The Independent. When Amazon’s Recognition software was used at the default 80% confidence level in a 2018 study, the software falsely matched 105 out of 1959 participants with mugshots of criminals. Seeing the issues with treatment of images depicting males and females above, one could imagine a disheartening scenario when judging actions of females in the public space. If men and women are judged differently for performing the same actions or being in the same positions, it would impact everybody’s right to equal treatment before the law. Bayerischer Rundfunk, a German media outlet, published an interactive page (only in German) where AI classification services’ differing classifications can be compared to one’s own assessment .

    Using gender-neutral images to circumvent human bias

    Besides the positive societal potentials of image classification, we also want to address some possible practical applications arising from being able to cover more than just two genders. An application that came to our minds is the use of “genderless” images to prevent human bias. Such a filter would imply losing individuality, so they would only be applicable in contexts where the benefit of reducing bias outweighs the cost of that loss.

    Imagining a browser extension for the hiring process

    HR screening could be an area where gender-neutral images may lead to less gender-based discrimination. Gone are the times of faceless job applications: if your LinkedIn profile has a profile picture it is 14 times more likely to get viewed. When examining candidate profiles, recruiters should ideally be free of subconscious, unintentional gender bias. Human nature prevents this. One could thus imagine a browser extension that generates a gender-neutral version of profile photos on professional social networking sites like LinkedIn or Xing. This could lead to more parity and neutrality in the hiring process, where only skills and character should count, and not one’s gender – or one’s looks for that matter (pretty privilege).

    Conclusion

    We set out to automatically generate gender-neutral versions from any input face image.

    Our implementation indeed automates the creation of gender-neutral faces. We used an existing GAN, image encoder and face image classifier. Our experiments with real peoples’ portraits show that the approach works well in many cases and produces realistically looking face images that clearly resemble the input image while remaining gender neutral.

    In some cases, we still found that the supposedly neutral images contain artifacts from technical glitches or still have their recognizable gender. Those limitations likely arise from the nature of the GANs latent space or the lack of artificially generated images in the classifiers training data. We are confident that further work can resolve most of those issues for real-world applications.

    Society’s ability to have an informed discussion on advances in AI is crucial

    Image classification has far-reaching consequences should be evaluated and discussed by society, not just a few experts. Any image classification service that is used to sort people into categories should be examined closely. What must be avoided is that members of society come to harm. Establishing responsible use of such systems, governance and constant evaluation are essential. An additional solution could be creating structures for the reasoning behind decisions using Explainable AI best practices to lay out why certain decisions were made. As a company in the field of AI, we at statworx look to our AI-principles as a guide.

     

    Image Sources:

    AdobeStock 210526825Wayhome Studio
    AdobeStock 243124072Damir Khabirov
    AdobeStock 387860637 insta_photos
    AdobeStock 395297652Nattakorn
    AdobeStock 480057743Chris
    AdobeStock 573362719Xavier Lorenzo

    AdobeStock 546222209 Rrose Selavy Isabel Hermes, Alexander Müller

    Introduction

    Forecasts are of central importance in many industries. Whether it’s predicting resource consumption, estimating a company’s liquidity, or forecasting product sales in retail, forecasts are an indispensable tool for making successful decisions. Despite their importance, many forecasts still rely primarily on the prior experience and intuition of experts. This makes it difficult to automate the relevant processes, potentially scale them, and provide efficient support. Furthermore, experts may be biased due to their experiences and perspectives or may not have all the relevant information necessary for accurate predictions.

    These reasons have led to the increasing importance of data-driven forecasts in recent years, and the demand for such predictions is accordingly strong.

    At statworx, we have already successfully implemented a variety of projects in the field of forecasting. As a result, we have faced many challenges and become familiar with numerous industry-specific use cases. One of our internal working groups, the Forecasting Cluster, is particularly passionate about the world of forecasting and continuously develops their expertise in this area.

    Based on our collected experiences, we now aim to combine them in a user-friendly tool that allows anyone to obtain initial assessments for specific forecasting use cases depending on the data and requirements. Both customers and employees should be able to use the tool quickly and easily to receive methodological recommendations. Our long-term goal is to make the tool publicly accessible. However, we are first testing it internally to optimize its functionality and usefulness. We place special emphasis on ensuring that the tool is intuitive to use and provides easily understandable outputs.

    Although our Recommender Tool is still in the development phase, we would like to provide an exciting sneak peek.

    Common Challenges

    Model Selection

    In the field of forecasting, there are various modeling approaches. We differentiate between three central approaches:

    1. Time Series Models
    2. Tree-based Models
    3. Deep Learning Models

    There are many criteria that can be used when selecting a model. For univariate time series data with strong seasonality and trends, classical time series models such as (S)ARIMA and ETS are appropriate. On the other hand, for multivariate time series data with potentially complex relationships and large amounts of data, deep learning models are a good choice. Tree-based models like LightGBM offer greater flexibility compared to time series models, are well-suited for interpretability due to their architecture, and tend to have lower computational requirements compared to deep learning models.

    Seasonality

    Seasonality refers to recurring patterns in a time series that occur at regular intervals (e.g. daily, weekly, monthly, or yearly). Including seasonality in the modeling is important to capture these regular patterns and improve the accuracy of forecasts. Time series models such as SARIMA, ETS, or TBATS can explicitly account for seasonality. For tree-based models like LightGBM, seasonality can only be considered by creating corresponding features, such as dummies for relevant seasonalities. One way to explicitly account for seasonality in deep learning models is by using sine and cosine functions. It is also possible to use a deseasonalized time series. This involves removing the seasonality initially, followed by modeling on the deseasonalized time series. The resulting forecasts are then supplemented with seasonality by applying the process used for deseasonalization in reverse. However, this process adds another level of complexity, which is not always desirable.

    Hierarchical Data

    Especially in the retail industry, hierarchical data structures are common as products can often be represented at different levels of granularity. This frequently results in the need to create forecasts for different hierarchies that do not contradict each other. The aggregated forecasts must therefore match the disaggregated forecasts. There are various approaches to this. With top-down and bottom-up methods, forecasts are created at one level and then disaggregated or aggregated downstream. Reconciliation methods such as Optimal Reconciliation involve creating forecasts at all levels and then reconciling them to ensure consistency across all levels.

    Cold Start

    In a cold start, the challenge is to forecast products that have little or no historical data. In the retail industry, this usually refers to new product introductions. Since it is not possible to train a model for these products due to the lack of history, alternative approaches must be used. A classic approach to performing a cold start is to rely on expert knowledge. Experts can provide initial estimates of demand, which can serve as a starting point for forecasting. However, this approach can be highly subjective and cannot be scaled. Similarly, similar products or potential predecessor products can be referenced. Grouping of products can be done based on product categories or clustering algorithms such as K-Means. Using cross-learning models trained on many products represents a scalable option.

    Recommender Concept

    With our Recommender Tool, we aim to address different problem scenarios to enable the most efficient development process. It is an interactive tool where users can provide inputs based on their objectives or requirements and the characteristics of the available data. Users can also prioritize certain requirements, and the output will prioritize those accordingly. Based on these inputs, the tool generates methodological recommendations that best cover the solution requirements, depending on the available data characteristics. Currently, the outputs consist of a purely content-based representation of the recommendations, providing concrete guidelines for central topics such as model selection, pre-processing, and feature engineering. The following example provides an idea of the conceptual approach:

    The output presented here is based on a real project where the implementation in R and the possibility of local interpretability were of central importance. At the same time, new products were frequently introduced, which should also be forecasted by the developed solution. To achieve this goal, several global models were trained using Catboost. Thanks to this approach, over 200 products could be included in the training. Even for newly introduced products where no historical data was available, forecasts could be generated. To ensure the interpretability of the forecasts, SHAP values were used. This made it possible to clearly explain each prediction based on the features used.

    Summary

    The current development is focused on creating a tool optimized for forecasting. Through its use, we aim to increase efficiency in forecasting projects. By combining gathered experience and expertise, the tool will offer guidelines for modeling, pre-processing, and feature engineering, among other topics. It will be designed to be used by both customers and employees to quickly and easily obtain estimates and methodological recommendations. An initial test version will be available soon for internal use, but the tool is ultimately intended to be made accessible to external users as well. In addition to the technical output currently in development, a less technical output will also be available. The latter will focus on the most important aspects and their associated efforts. In particular, the business perspective in the form of expected efforts and potential trade-offs between effort and benefit will be covered by this.

     

     

    Benefit from our forecasting expertise!

    If you need support in addressing the challenges in your forecasting projects or have a forecasting project planned, we are happy to provide our expertise and experience to assist you.

       

      Image Source:

      AdobeStock 83282923 – Mego-studio Marlon Schumacher

      Recently, while working at statworx, I experienced a sense of déjà vu regarding the topic of data culture. As the Head of the AI Academy, my main responsibility is to convey my enthusiasm for artificial intelligence, programming, data, and cloud computing to my clients. This often requires projecting my passion for these subjects onto individuals who may have limited technical experience, and whose interests may not typically align with transformer models and functional programming

      This tension reminded me of something that happened before my professional career.

      All beginnings are difficult

      Prior to my passion for data and artificial intelligence, I was already a very enthusiastic (hobby) musician – with a special passion for the genre of Death Metal (Note: I don’t want to bother interested readers with more detailed genre descriptions here 😉). During my studies, I was a singer and guitarist in a Death Metal band. For those of you who are not familiar with Death Metal, it may seem like all those “off-key notes” and “growling” don’t require real skills – but let me assure you, it takes a lot of talent, and many people in this genre have years of hard work behind them.

      https://youtu.be/WGnXD0DME30?t=25

      When you listen to or, even better, watch this music, you are quickly impressed by how fast the musicians today race across their guitar fretboards. However, it’s essential to recognize that every musician faces a challenging beginning. Those who have learned an instrument can attest to this reality. Initially, it can be demanding to navigate through prescribed teaching materials and maintain the necessary drive to acquire techniques, with the ultimate goal of performing a decent piece of music. At first, it was very difficult for me to get excited about notes, rhythms, and finger exercises or to stay on task with the appropriate perseverance.

      Generiert mit DALL-E. Prompt: death metal concert with view from stage to crowd, guitar in the foreground with bokeh, photorealistic style

      Let’s get creative

      At the beginning, the songs were not particularly good or technically demanding, as I had not yet learned any significant guitar or singing skills. But then something happened: my motivation kicked in! I realized how these techniques and skills allowed me to express my own feelings and thoughts. It was as if I could create my own products.

      I wrote more and more songs and almost unnoticed learned important skills on the fretboard. It became my personal mission to stoically master all the necessary finger exercises in order to be able to play ever more complex structures. At the same time, I became part of bands and a local music scene where we inspired each other at concerts and kept motivating each other to write more complex and better material. Here, we also inspired more, mostly younger, music fans to try their hand at this music. They joined in, listened, and thought, “I want to be able to do that too!” So they started writing their own songs, learning their own techniques, and becoming part of a creative cultural scene.

      Skills alone are not everything

      One may wonder what this little excursion has to do with data culture. The above theme has also been reflected in my work with data culture. In our AI Academy, we mainly focus on topics related to data literacy and related skills. Initially, I made the same mistake in my thinking that hindered me when learning my instrument: skills are everything – or with skills, everything else will somehow come.

      I assumed that the skills taught are so important, so relevant, so productive, and especially so attractive to learners that after learning these skills, everything else will automatically follow. But that’s not the case. Over time, through our training, we have reached an ever-increasing circle of people, including those with different core competencies. These are people who cannot or do not want to be evangelists or enthusiasts for matrix algebra in their main activity.

      The following questions are always at the forefront here:

      “What does this have to do with me?”
      “What does this have to do with my work?”
      “How could this be valuable for me?”

      And just like in my story about songwriting, playing concerts, or exchanging ideas within a music scene, I also had the same experience with data and upskilling. Some of our most successful training formats, the AI Basics Workshop and Data Literacy Workshop, enable the most important topics and learnings around data and AI to be made usable for one’s own company – with the possibility of generating their own ideas for the use of these technologies together with experienced AI experts. This is not only about learning how AI works, but also about interactive and guided exploration:

      “What does this have to do with me?”
      “How can I create value for my environment with this?”
      “What problems does AI need to solve for me?”

      Motivating ideas

      At first, we noticed how enthusiastic training participants interacted with the content, and how the mood in our courses shifted much more towards a growth mindset:

      Not focusing on what I can already do, but rather asking what I still want to achieve and what I want to achieve.

      On the other hand, our courses quickly became popular with our customers’ employees. We were, of course, pleased with the word-of-mouth that contributed to the recognition of the high course quality and exciting topics. However, we did not anticipate that the ideas generated in the course would develop their own dynamic and, in many cases, generate even greater impact in the company than the course itself.

      Similar to concerts in the death metal scene, new enthusiasts could also be won over here. They realized that the person who successfully drives a use case forward was also at the beginning of data and AI not too long ago.

      “If others have achieved that, I want to try it too, and I’ll figure out how to learn the finger skills on the way.”

      Can – Do – Want – A constant cycle in the organization

      And so three important dimensions came together for us.

      1. Can – Mastery of skills such as good guitar playing, project management in data and AI, programming, or basic knowledge in data analysis.
      2. Do – Regular and ritualized work on the topic, conducting initial use cases, and exchanging ideas with others to learn the language interactively.
      3. Want – Creating sustainable motivation to achieve goals through initial successes, inspiring exchange, and a clear vision for the potential impact and value generation in the company.

      The three dimensions form a cycle in which each dimension depends on the others and has a positive effect on the other dimensions. If I improve my guitar playing skills, it will be easier for me to develop new ideas and share them successfully with others. This creates further motivation to tackle more skills and challenges.

      That is why data culture and death metal have a lot in common for me.

      Let’s connect if you’re interested in diving deeper into the topic of data culture, including its three dimensions “Can”, “Do”, and “Want”.

       

      More about AI Academy

      Image Source:

      AdobeStock 480687393 zamuruev

        David Schlepps

      A data culture is a key factor for effective data utilization

      With the increasing digitization, the ability to use data effectively has become a crucial success factor for businesses. This way of thinking and acting is often referred to as data culture and plays an important role in transforming a company into a data-driven organization. By promoting a data culture, businesses can benefit from the flexibility of fact-based decision-making and fully leverage the potential of their data. Such a culture enables faster and demonstrably better decisions and embeds data-driven innovation within the company.

      Although the necessity and benefits of a data culture appear obvious, many companies still struggle to establish such a culture. According to a study by New Vantage Partners, only 20% of companies have successfully developed a data culture so far. Furthermore, over 90% of the surveyed companies describe the transformation of culture as the biggest hurdle in the transformation towards a data-driven company.

      A data culture fundamentally changes the way of working

      The causes of this challenge are diverse, and the necessary changes permeate almost all aspects of everyday work. In an effective data culture, each employee preferably uses data and data analysis for decision-making and gives priority to data and facts over individual “gut feeling.” This way of thinking promotes the continuous search for ways to use data to identify competitive advantages, open up new revenue streams, optimize processes, and make better predictions. By adopting a data culture, companies can fully leverage the potential of their data and drive innovations throughout the organization. This requires recognizing data as an important driving force for decision-making and innovation. This ideal requires new demands on individual employee behavior. Additionally, this requires targeted support of this behavior through suitable conditions such as technical infrastructure and organizational processes.

      Three factors significantly shape the data culture

      To anchor a data culture sustainably within a company, three factors are crucial:

      1. Can| Skills
      2. Want | Attitude
      3. Do | Actions

      statworx uses these three factors to make the abstract concept of data culture tangible and to initiate targeted necessary changes. It is crucial to give equal attention to all factors and to consider them as holistically as possible. Initiatives for cultural development often limit themselves to the aspect of attitude and attempt to anchor specific values separate from other influencing factors. These initiatives usually fail due to the reality of companies that oppose them with their processes, lived rituals, practices, and values, and thus prevent the establishment of the culture (actively).

      We have summarized three factors of data culture in a framework for an overview.

      1. Can: Skills form the basis for effective data utilization

      Skills and competencies are the foundation for effective data management. These include both the methodological and technical skills of employees, as well as the organization’s ability to make data usable.

      Ensuring data availability is particularly important for data usability. The “FAIR” standard – Findable, Accessible, Interoperable, Reusable – provides a direction for the essential properties that support this, such as through technologies, knowledge management, and appropriate governance.

      At the level of employee skills, the focus is on data literacy – the ability to understand and effectively use data to make informed decisions. This includes a basic understanding of data types and structures, as well as collection and analysis methods. Data literacy also involves the ability to ask the right questions, interpret data correctly, and identify patterns and trends. Develop relevant competencies, such as through upskilling, targeted workforce planning, and hiring data experts.

      2. Want: A data culture can only flourish in a suitable value context.

      The second factor – Want – deals with the attitudes and intentions of employees and the organization as a whole towards the use of data. For this, both the beliefs and values of individuals and the community within the company must be addressed. There are four aspects are of central importance for a data culture:

      • Collaboration & community instead of competition & selective partnerships
      • Transparency & sharing instead of information concealment & data hoarding
      • Pilot projects & experiments instead of theoretical assessments
      • Openness & willingness to learn instead of pettiness & rigid thinking
      • Data as a central decision-making basis instead of individual opinion & gut feeling

      Example: Company without a data culture

      On an individual level, an employee is convinced that exclusive knowledge and data can provide an advantage. The person has also learned within the organization that this behavior leads to strategic advantages or opportunities for personal positioning, and has been rewarded for such behavior by superiors in the past. The person is therefore convinced that it is absolutely sensible and advantageous to keep data for oneself or within one’s own team and not share it with other departments. The competitive thinking and tendency towards secrecy are firmly anchored as a value.

      In general, behavior like described in the example restricts transparency throughout the entire organization and thereby slows down the organization. If not everyone has the same information, it is difficult to make the best possible decision for the entire company. Only through openness and collaboration can the true value of data in the company be utilized. A data-driven company is based on a culture of collaboration, sharing, and learning. When people are encouraged to exchange their ideas and insights, better decisions can be made.

      Even possible declarations of intent, such as mission statements and manifestos without tangible measures, will change little in the attitude of employees. The big challenge is to anchor the values sustainably and to make them the guiding action principle for all employees, which is actively lived in everyday business. If this succeeds, the organization is on the best way to create the required data mindset to bring an effective and successful data culture to life. Our transformation framework can help to establish and make these values visible.

      We recommend starting to build a data culture step by step because even small experimental projects create added value, serve as positive examples, and build trust. The practical testing of a new innovation, even only in a limited scope, usually brings faster and better results than a theoretical assessment. Ultimately, it is about placing the value of data at the forefront.

      3. Do: Behavior creates the framework and is simultaneously the visible result of a data culture.

      The two factors mentioned above ultimately aim to ensure that employees and the organization as a whole adapt their behavior. Only an actively lived data culture can be successful. Therefore, everyday behavior – Do – plays a central role in establishing a data culture.

      The behavior of an organization can be examined and influenced primarily in two dimensions.

      These factors are:

      1. Activities and rituals
      2. Structural elements of the organization

      Activities and rituals

      Activities and rituals refer to the daily collaboration between employees of an organization. They manifest themselves in all forms of collaboration, from meeting procedures to handling feedback and risks to the annual Christmas party. It is crucial which patterns the collaboration follows and which behavior is rewarded or punished.

      Experience shows that teams that are already familiar with agile methods such as Scrum find the transformation to data-driven decisions easier. Teams that follow strong hierarchies and act risk-averse, on the other hand, have more difficulty overcoming this challenge. One reason for this is that agile ways of working reinforce collaboration between different roles and thus create the foundation for a productive work environment. In this context, the role of leadership, especially senior leadership, is crucial. The individuals at the C-level must necessarily lead by example from the beginning, introduce rituals and activities, and act together as the central driver of the transformation.

      Structure elements of the organization

      While activities and rituals emerge from teams and are not always predetermined, the second dimension reflects a stronger formalization. It refers to the structure elements of an organization. These provide the formal framework for decisions and thus shape behavior, as well as the emergence and anchoring of values and attitudes.

      Internal and external structure elements are distinguished. Internal structure elements are mainly visible within the organization, such as roles, processes, hierarchy levels, or committees. By adapting and restructuring roles, necessary skills can be reflected within the company. Furthermore, rewards and promotions for employees can create an incentive to adopt and pass on the behavior themselves to colleagues. The division of the working environment is also part of the internal structure. Since the work in data-driven companies is based on close collaboration and requires individuals with different skills, it makes sense to create a space for open exchange that allows communication and collaboration.

      External structure elements reflect internal behavior outward. Thus, internal structure elements influence the perception of the company from the outside. This is reflected, for example, in clear communication, the structure of the website, job advertisements, and marketing messages.

      Companies should design their external behavior to be in line with the values of the organization and thus support their own structures. In this way, a harmonious alignment between the internal and external positioning of the company can be achieved.

      First small steps can already create significant changes

      Our experience has shown that the coordinated design of skills, willingness, and action results in a sustainable data culture. It is now clear that a data culture cannot be created overnight, but it is also no longer possible to do without it. It has proven useful to divide this challenge into small steps. With first pilot projects, such as establishing a data culture in just one team and initiatives for particularly committed employees who want to drive change, trust is created in the cultural shift. Positive individual experiences serve as a helpful catalyst for the transformation of the entire organization.

      The philosopher and visionary R. Buckminster Fuller once said, “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.” Because with the advancement of technology, companies must be able to adapt to fully tap their potential. This allows decisions to be made faster and more accurately than ever before, drives innovation, and increasingly optimizes processes. The sustainable establishment of a data culture will give companies a competitive advantage in the market. In the future, data culture will be an essential part of any successful business strategy. Companies that do not embrace this will be left behind.

      However, the use of data is a major problem for many companies. Often, data quality and data compilation stand in the way. Even though many companies already have data solutions, they are not optimally utilized. This means that much information remains unused and cannot be incorporated into decision-making.

       

      Sources:

      [1] https://hbr.org/2020/03/how-ceos-can-lead-a-data-driven-culture

      Image: AdobeStock 569760113 Annsophie Huber

      A data culture is a key factor for effective data utilization

      With the increasing digitization, the ability to use data effectively has become a crucial success factor for businesses. This way of thinking and acting is often referred to as data culture and plays an important role in transforming a company into a data-driven organization. By promoting a data culture, businesses can benefit from the flexibility of fact-based decision-making and fully leverage the potential of their data. Such a culture enables faster and demonstrably better decisions and embeds data-driven innovation within the company.

      Although the necessity and benefits of a data culture appear obvious, many companies still struggle to establish such a culture. According to a study by New Vantage Partners, only 20% of companies have successfully developed a data culture so far. Furthermore, over 90% of the surveyed companies describe the transformation of culture as the biggest hurdle in the transformation towards a data-driven company.

      A data culture fundamentally changes the way of working

      The causes of this challenge are diverse, and the necessary changes permeate almost all aspects of everyday work. In an effective data culture, each employee preferably uses data and data analysis for decision-making and gives priority to data and facts over individual “gut feeling.” This way of thinking promotes the continuous search for ways to use data to identify competitive advantages, open up new revenue streams, optimize processes, and make better predictions. By adopting a data culture, companies can fully leverage the potential of their data and drive innovations throughout the organization. This requires recognizing data as an important driving force for decision-making and innovation. This ideal requires new demands on individual employee behavior. Additionally, this requires targeted support of this behavior through suitable conditions such as technical infrastructure and organizational processes.

      Three factors significantly shape the data culture

      To anchor a data culture sustainably within a company, three factors are crucial:

      1. Can| Skills
      2. Want | Attitude
      3. Do | Actions

      statworx uses these three factors to make the abstract concept of data culture tangible and to initiate targeted necessary changes. It is crucial to give equal attention to all factors and to consider them as holistically as possible. Initiatives for cultural development often limit themselves to the aspect of attitude and attempt to anchor specific values separate from other influencing factors. These initiatives usually fail due to the reality of companies that oppose them with their processes, lived rituals, practices, and values, and thus prevent the establishment of the culture (actively).

      We have summarized three factors of data culture in a framework for an overview.

      1. Can: Skills form the basis for effective data utilization

      Skills and competencies are the foundation for effective data management. These include both the methodological and technical skills of employees, as well as the organization’s ability to make data usable.

      Ensuring data availability is particularly important for data usability. The “FAIR” standard – Findable, Accessible, Interoperable, Reusable – provides a direction for the essential properties that support this, such as through technologies, knowledge management, and appropriate governance.

      At the level of employee skills, the focus is on data literacy – the ability to understand and effectively use data to make informed decisions. This includes a basic understanding of data types and structures, as well as collection and analysis methods. Data literacy also involves the ability to ask the right questions, interpret data correctly, and identify patterns and trends. Develop relevant competencies, such as through upskilling, targeted workforce planning, and hiring data experts.

      2. Want: A data culture can only flourish in a suitable value context.

      The second factor – Want – deals with the attitudes and intentions of employees and the organization as a whole towards the use of data. For this, both the beliefs and values of individuals and the community within the company must be addressed. There are four aspects are of central importance for a data culture:

      Example: Company without a data culture

      On an individual level, an employee is convinced that exclusive knowledge and data can provide an advantage. The person has also learned within the organization that this behavior leads to strategic advantages or opportunities for personal positioning, and has been rewarded for such behavior by superiors in the past. The person is therefore convinced that it is absolutely sensible and advantageous to keep data for oneself or within one’s own team and not share it with other departments. The competitive thinking and tendency towards secrecy are firmly anchored as a value.

      In general, behavior like described in the example restricts transparency throughout the entire organization and thereby slows down the organization. If not everyone has the same information, it is difficult to make the best possible decision for the entire company. Only through openness and collaboration can the true value of data in the company be utilized. A data-driven company is based on a culture of collaboration, sharing, and learning. When people are encouraged to exchange their ideas and insights, better decisions can be made.

      Even possible declarations of intent, such as mission statements and manifestos without tangible measures, will change little in the attitude of employees. The big challenge is to anchor the values sustainably and to make them the guiding action principle for all employees, which is actively lived in everyday business. If this succeeds, the organization is on the best way to create the required data mindset to bring an effective and successful data culture to life. Our transformation framework can help to establish and make these values visible.

      We recommend starting to build a data culture step by step because even small experimental projects create added value, serve as positive examples, and build trust. The practical testing of a new innovation, even only in a limited scope, usually brings faster and better results than a theoretical assessment. Ultimately, it is about placing the value of data at the forefront.

      3. Do: Behavior creates the framework and is simultaneously the visible result of a data culture.

      The two factors mentioned above ultimately aim to ensure that employees and the organization as a whole adapt their behavior. Only an actively lived data culture can be successful. Therefore, everyday behavior – Do – plays a central role in establishing a data culture.

      The behavior of an organization can be examined and influenced primarily in two dimensions.

      These factors are:

      1. Activities and rituals
      2. Structural elements of the organization

      Activities and rituals

      Activities and rituals refer to the daily collaboration between employees of an organization. They manifest themselves in all forms of collaboration, from meeting procedures to handling feedback and risks to the annual Christmas party. It is crucial which patterns the collaboration follows and which behavior is rewarded or punished.

      Experience shows that teams that are already familiar with agile methods such as Scrum find the transformation to data-driven decisions easier. Teams that follow strong hierarchies and act risk-averse, on the other hand, have more difficulty overcoming this challenge. One reason for this is that agile ways of working reinforce collaboration between different roles and thus create the foundation for a productive work environment. In this context, the role of leadership, especially senior leadership, is crucial. The individuals at the C-level must necessarily lead by example from the beginning, introduce rituals and activities, and act together as the central driver of the transformation.

      Structure elements of the organization

      While activities and rituals emerge from teams and are not always predetermined, the second dimension reflects a stronger formalization. It refers to the structure elements of an organization. These provide the formal framework for decisions and thus shape behavior, as well as the emergence and anchoring of values and attitudes.

      Internal and external structure elements are distinguished. Internal structure elements are mainly visible within the organization, such as roles, processes, hierarchy levels, or committees. By adapting and restructuring roles, necessary skills can be reflected within the company. Furthermore, rewards and promotions for employees can create an incentive to adopt and pass on the behavior themselves to colleagues. The division of the working environment is also part of the internal structure. Since the work in data-driven companies is based on close collaboration and requires individuals with different skills, it makes sense to create a space for open exchange that allows communication and collaboration.

      External structure elements reflect internal behavior outward. Thus, internal structure elements influence the perception of the company from the outside. This is reflected, for example, in clear communication, the structure of the website, job advertisements, and marketing messages.

      Companies should design their external behavior to be in line with the values of the organization and thus support their own structures. In this way, a harmonious alignment between the internal and external positioning of the company can be achieved.

      First small steps can already create significant changes

      Our experience has shown that the coordinated design of skills, willingness, and action results in a sustainable data culture. It is now clear that a data culture cannot be created overnight, but it is also no longer possible to do without it. It has proven useful to divide this challenge into small steps. With first pilot projects, such as establishing a data culture in just one team and initiatives for particularly committed employees who want to drive change, trust is created in the cultural shift. Positive individual experiences serve as a helpful catalyst for the transformation of the entire organization.

      The philosopher and visionary R. Buckminster Fuller once said, “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.” Because with the advancement of technology, companies must be able to adapt to fully tap their potential. This allows decisions to be made faster and more accurately than ever before, drives innovation, and increasingly optimizes processes. The sustainable establishment of a data culture will give companies a competitive advantage in the market. In the future, data culture will be an essential part of any successful business strategy. Companies that do not embrace this will be left behind.

      However, the use of data is a major problem for many companies. Often, data quality and data compilation stand in the way. Even though many companies already have data solutions, they are not optimally utilized. This means that much information remains unused and cannot be incorporated into decision-making.

       

      Sources:

      [1] https://hbr.org/2020/03/how-ceos-can-lead-a-data-driven-culture

      Image: AdobeStock 569760113 Annsophie Huber