CLIENTS

—

Bechtle AG
ciSioLogo_Blau-Gruen65_LR.png
kappniles_logo.png
hug-logo.png
mehnert.png
arte_logo.png
metatrader-5-logo_2x
ernst-abbe
integratedworlds_logo-claim_rg

We are proud to offer a wide range of AI-based solutions that help our customers in a variety of industries optimize their business processes.

Here is an overview of customer projects already successfully implemented by us. Our listed references were professionally supported by AI UI to improve their products and enhance the customer experience.

Quality assurance

BETTER THAN THE HUMAN EYE

QUALIFY YOUR PRODUCTS

Germany is world-renowned for high-quality mechanical engineering. This sector is dominated by companies that produce components for the global market with the highest precision and unmatched quality. This advantage must be maintained and expanded through relentless innovation and inventiveness. Detailed knowledge of processes and results is incorporated into the improvement of manufacturing. A few highly qualified engineers are irreplaceable in this process. Unfortunately, they have to spend a lot of time in quality control, a concentrated but grueling job.

It therefore stands to reason to use AI to automate quality control and implement it on a much larger scale. The only problem is that mechanical engineers rarely have experience in software development.

The Kapp Niles company therefore decided to use AIUI to bring AI into its operational processes, but without learning deep understanding in programming languages and AI frameworks. The unique user interface quickly created a neural network that had learned everything the engineers needed to know and provided quality assurance in checking interface structures. To the surprise of the project team, the AI was also able to provide statistics about defects in production more transparently than ever before, leading directly to manufacturing improvements.

The company was able to take an innovative step to stay significantly ahead of the competition and offer better products to their customers faster and at lower prices. This did not require the creation of a separate IT department.

  • AI image recognition
    Industrial image processing via AI algorithms, robust even with uneven exposure
  • Fully automatic control
    Fully automatic scanning and comprehensive evaluation
  • Quality management
    Quantitative defect detection now enables a better evaluation process that was previously subjective.

Process Analysis (1 day):

  • Diamonds are electroplated with nickel onto a steel surface.
  • Defects of various types can form in the process
  • Quality control is performed by an experienced operator under the microscope by 100% inspection.

Process evaluation (1 day):

  • The goal is the error-free, fully automated detection of voids in the diamond bed

AI Development (14 days):

  • Acquisition of a webcam microscope for digital image data acquisition.
  • Acquisition of ~300 images of various exposure, grain size, shape
  • Data labeling by Kapp Niles staff, i.e., marking the defects in AI-UI.
  • Training of artificial intelligence with AI-UI

Testing (1 day):

  • Use of extra images, to test AI prediction goodness on unseen data


Documentation (1 day):

  • AI-UI is designed so that documentation happens automatically

Go-live (1 day):

  • On the part of the customer, an interface between the camera microscope and AI-UI was established
  • AI-UI provided automatic quality analysis

eCommerce

GO DEEPER AND DEEPER

AUTOMATED AMAZON ANALYTICS

Merging a wide variety of data sources into a uniform database. Based on this, an AI can predictively perform automated actions in an existing eCommerce store or an Amazon seller account in order to permanently increase margins and sales.

Amazon has achieved a dominant position in e-commerce through price dumping and a relentless race to the lowest margins. As a distributor, Home & Garden 24 has been able to establish itself in this market. Over the years, the company realized that quite a few products offered have a sales cycle that depends on very many factors. The interrelationships are very complex and can actually only be grasped through years of experience. Therefore, the management deals daily with finding the best price, which can be significantly higher depending on the cycle – customers are willing to pay more in certain situations, for example, if a delivery can be guaranteed just before Christmas or if goods are guaranteed to be available for certain weather.

Due to the enormous cost pressure in eCommerce, Home & Garden 24 employs only a small IT department that covers a wide range of tasks but does not maintain specialist AI. The highest positions in the company spend too much time on the very strategic pricing of thousands of products, which is difficult to deliver.

Therefore, Home & Garden 24 was looking for a solution that would take the burden off management, not require further investment in training software engineers, not hand off mission-critical knowledge to the outside world, and be able to represent the inner workings of AI to better understand the complexity.

Home & Garden 24 decided to apply AI-UI to the problem. Within just a few weeks, this resulted in a neural network that looked at all input parameters over the past 2 years and was able to make automated price adjustments. The IT department will carry out the future optimization itself through the simple user interface and will get new insights into the decisions through the transparent presentation of the data and the levels of the network. No data will end up at Amazon or Google, where it will be used completely uncontrolled.

Home & Garden 24 expects about 30% improved margins.

  • Data fusion
    Data from Amazon, the customer itself, and a weather service were merged and used for AI training.
  • AI demand forecasting
    Using historical data, a model was developed that estimates future demand and recalculates it daily.
  • Price adjustment
    If demand is expected to be very high, prices can now be adjusted upwards fully automatically in the customer’s ERP system to increase sales and prevent stock-outs.

Maschinenbau

Every day, billions of pieces of information are generated at every machine in manufacturing plants around the world, which contain information about the process that created them. The machines that carry out these processes are usually regulated or controlled and therefore automatically require the acquisition of this data, since without it they could not be controlled at all.

Existing machines can be easily upgraded with software in the form of AI. Control and measurement variables generated in the manufacturing process can be used to subsequently implement fully automated quality assurance or monitoring without having to purchase new machines.

Die Technische Universität Ilmenau ist eine forschungsstarke Universität in der Interdisziplinäre Spitzenforschung großgeschrieben wird. In rund 100 Fachgebieten an fünf Fakultäten der TU Ilmenau betreiben Wissenschaftler: innen innovative Grundlagenforschung und angewandte Forschung in Ingenieurwissenschaften, Informationstechnologien, Wirtschafts- und Medienwissenschaften, Mathematik und Naturwissenschaften auf höchstem Niveau.

  • Machine data acquisition
    Data such as, force, infeed distance, speed, torque and others were standardized together for AI training.
  • AI Quality Prediction
    By training IO and NIO processes, the AI was able to learn which measurement trends were related to defective products and which were okay.
  • Quality traffic light
    After successful AI training, a traffic light could be attached to the welding machine that makes statements about whether the welded joint is of good quality or not (IO / NOK) after each welding process based on the process data.

other:

  • Welding process digitized
  • Data transmission established
  • Fully automated quality control IO/NIO
  • Predictive maintenance foundation laid

ROBOTIcs

In industrial robotics, camera systems are very often used to detect objects and avoid collisions. However, the software that enables object detection is often unable to do this under varying light conditions because it is based on classical algorithms. Therefore, robots are often completely isolated from external light influences in their cells, which is no longer possible with the increasing use of cobots. With the help of an image recognition AI, it was achieved in this project that objects could be successfully found even without additional light sources and that all light influences could be eliminated.

The Mehnert Lab brings collaboration to Industry 4.0: In times of digital transformation and permanent changes in markets and industries, cooperation, synergy and knowledge exchange become essential for growth and progress. As a hub of industrial collaboration, the Mehnert Lab brings together different players in the value creation process, enabling the joint development of new business models and service concepts as well as the qualification of skilled workers.

  • AI training
  • Use of an object recognition AI in robotics
  • Annotation of images

prognose

In a collaborative development project, we were able to develop an AI prediction model. This led us to an innovative AI solution for delivery time forecasts for furniture shipping. In collaboration with Integrated Worlds GmbH, a platform was created on which automation solutions are made available via artificial intelligence and forecasting tools.

Integrated Worlds GmbH has been designing holistic and sustainable solution concepts since 1996. In doing so, they understand the business needs and requirements necessary for the success of their customers – whether flexibility, cost efficiency, scalability of solutions or Big Data and AI!

DIGITALIZATION

The EAH Jena and AI-UI try together not only in teaching, but also in consulting SMEs in the field of digitization and AI with the help of AIMS and know-how from past projects to provide knowledge.

We are proud to count the Ernst Abbe University of Applied Sciences Jena as our cooperation partner. The teaching institute is already using AIUI’s software very successfully for consulting and content transfer within various courses.

Practice-oriented and strong in research. Interdisciplinary and modern. Familiar and cosmopolitan.

 

This and much more characterizes the Ernst Abbe University of Applied Sciences Jena (in short: EAH Jena). Currently, about 4,400 young people are studying here in the Bachelor’s and Master’s programs in the fields of technology, economics, social sciences and health. On our campus, we live interdepartmental cooperation in teaching, studying, research, transfer and administration. And all this in a beautiful city where student life is not neglected either. Our motto? Innovation for quality of life. Health, precision, sustainability & networking.

TRADING

Automated stock trading with AI uses large amounts of financial data to perform real-time analysis and make appropriate trading decisions. Our software for online trading can be used through MetaTrader. Advantages that accrue to the numerous users are fast execution of buying and selling activities and avoidance of human errors and impulse actions.

EROTIK

PRECISION AT ITS FINEST

Bereits bestehende Maschinen lassen sich einfach mit Software in Form von KI aufrüsten. Steuer und Messgrößen, die im Fertigungsprozess erzeugt werden, können verwendet werden um nachträglich eine vollautomatische Qualitätssicherung oder Überwachung zu implementieren, ohne neue Maschinen anschaffen zu müssen.

Weltweit fallen in Fertigungsbetrieben an jeder Maschine täglich Milliarden Informationen an, die Aussagen über den Prozess, durch den Sie entstanden sind in sich tragen. Die Maschinen, die diese Prozesse ausführen sind meist geregelt oder gesteuert und bedürfen daher ganz automatisch einer Erfassung dieser Daten, da sie sich ohne diese überhaupt nicht kontrollieren ließen.

Warum also nicht diese Daten nutzen, um weitere Einblicke in den Prozess zu erhalten. Doch wie soll das gehen?

Im Vorliegenden Beispiel wurde ein Rührreibschweißprozess genutzt, um zu zeigen wie man mit den entstehenden Prozessdaten zu Qualitätsaussagen kommen kann, ohne weitere Messtechnik einkaufen zu müssen. Zwei Bleche (Kupfer und Aluminium) werden miteinander über einen rotierenden Stempel, unter hohem Druck verschweißt und die daraus resultierenden Prozessdaten wie Kraft, Weg, Strom, Drehzahl, Drehmoment und Zeit wurden aufgezeichnet. Dabei wurden absichtlich hunderte von Proben als IO produziert und weitere Proben leicht verunreinigt um NIO Schweißverbindungen zu erzeugen. 

Die Idee ist, dass bei verunreinigten Blechen auch Unterschiede in den Prozessparametern auftreten. Diese möchte man aber nicht aufwändig, händisch heraussuchen. Daher nutzt man einfaches KI Training, indem man dem Algorithmus zeigt welche Daten „IO“ sind und welche „NIO“, sodass die KI selbstständig herausfindet worin die Unterschiede innerhalb der Prozessabläufe liegen.

Mit ausreichend Trainingsdaten können hier hohe Genauigkeiten erreicht werden und das ohne zusätzliche Messsysteme.

Im Ergebnis lassen sich die Daten von jedem Prozessschritt nutzen, um ein kontinuierliches Qualitätssicherungskonzept umzusetzen.

Detaillierte Ausführungen und Erläuterungen finden Sie unter der folgenden Veröffentlichung in Zusammenarbeit mit der TU Ilmenau.

DOI: 10.1177/1464420720912773

  • Maschinendatenerfassung
    Daten wie, Kraft, Zustellweg, Drehzahl, Drehmoment und andere wurden standardisiert für das KI Training zusammengeführt.
  • KI Qualitätsprognose
    Durch Antrainieren von IO- und NIO Prozessen, konnte die KI erlernen, welche Messwertverläufe mit fehlerhaften Produkten zusammenhängen und welche okay sind.
  • Qualitätsampel
    Nach erfolgreichem KI Training, konnte an der Schweißmaschine eine Ampel angebracht werden, die nach jedem Schweißprozess auf Basis der Prozessdaten Aussagen darüber trifft, ob die Schweißverbindung von guter Qualität ist oder nicht (IO / NIO).

Sonstiges:

  • Schweißprozess digitalisiert
  • Datenübertragung etabliert
  • Vollautomatische Qualitätskontrolle IO/NIO
  • Predictive Maintenance Grundstein gelegt

DOCUMENTS

Document processing in many companies and communities is associated with high bureaucratic effort. AI-UI and ciSio have joined forces to use software robots and intelligent document processing to automate processes that were previously impossible to automate.

ciSio translates and takes the complexity out of digitization. Exciting new technologies become available. With ciSio everything becomes easier: the introduction, application, operation and procurement. Ideally suited for medium-sized businesses. Through know-how transfer, ciSio makes medium-sized businesses independent.

By using the AI-UI internal AI document chatbot, it is possible to ask questions to a document and receive automated answers.

No matter if invoice number, Iban, address, or handwritten answer texts on forms of any kind, our AI automatically delivers the answer to your systems, so that processes that could not be automated so far, become automatable.

DIGITALISIERUNG

Bereits bestehende Maschinen lassen sich einfach mit Software in Form von KI aufrüsten. Steuer und Messgrößen, die im Fertigungsprozess erzeugt werden, können verwendet werden um nachträglich eine vollautomatische Qualitätssicherung oder Überwachung zu implementieren, ohne neue Maschinen anschaffen zu müssen.

Weltweit fallen in Fertigungsbetrieben an jeder Maschine täglich Milliarden Informationen an, die Aussagen über den Prozess, durch den Sie entstanden sind in sich tragen. Die Maschinen, die diese Prozesse ausführen sind meist geregelt oder gesteuert und bedürfen daher ganz automatisch einer Erfassung dieser Daten, da sie sich ohne diese überhaupt nicht kontrollieren ließen.

Warum also nicht diese Daten nutzen, um weitere Einblicke in den Prozess zu erhalten. Doch wie soll das gehen?

Im Vorliegenden Beispiel wurde ein Rührreibschweißprozess genutzt, um zu zeigen wie man mit den entstehenden Prozessdaten zu Qualitätsaussagen kommen kann, ohne weitere Messtechnik einkaufen zu müssen. Zwei Bleche (Kupfer und Aluminium) werden miteinander über einen rotierenden Stempel, unter hohem Druck verschweißt und die daraus resultierenden Prozessdaten wie Kraft, Weg, Strom, Drehzahl, Drehmoment und Zeit wurden aufgezeichnet. Dabei wurden absichtlich hunderte von Proben als IO produziert und weitere Proben leicht verunreinigt um NIO Schweißverbindungen zu erzeugen. 

Die Idee ist, dass bei verunreinigten Blechen auch Unterschiede in den Prozessparametern auftreten. Diese möchte man aber nicht aufwändig, händisch heraussuchen. Daher nutzt man einfaches KI Training, indem man dem Algorithmus zeigt welche Daten „IO“ sind und welche „NIO“, sodass die KI selbstständig herausfindet worin die Unterschiede innerhalb der Prozessabläufe liegen.

Mit ausreichend Trainingsdaten können hier hohe Genauigkeiten erreicht werden und das ohne zusätzliche Messsysteme.

Im Ergebnis lassen sich die Daten von jedem Prozessschritt nutzen, um ein kontinuierliches Qualitätssicherungskonzept umzusetzen.

Detaillierte Ausführungen und Erläuterungen finden Sie unter der folgenden Veröffentlichung in Zusammenarbeit mit der TU Ilmenau.

DOI: 10.1177/1464420720912773

  • Maschinendatenerfassung
    Daten wie, Kraft, Zustellweg, Drehzahl, Drehmoment und andere wurden standardisiert für das KI Training zusammengeführt.
  • KI Qualitätsprognose
    Durch Antrainieren von IO- und NIO Prozessen, konnte die KI erlernen, welche Messwertverläufe mit fehlerhaften Produkten zusammenhängen und welche okay sind.
  • Qualitätsampel
    Nach erfolgreichem KI Training, konnte an der Schweißmaschine eine Ampel angebracht werden, die nach jedem Schweißprozess auf Basis der Prozessdaten Aussagen darüber trifft, ob die Schweißverbindung von guter Qualität ist oder nicht (IO / NIO).

Sonstiges:

  • Schweißprozess digitalisiert
  • Datenübertragung etabliert
  • Vollautomatische Qualitätskontrolle IO/NIO
  • Predictive Maintenance Grundstein gelegt

Discover our AI LAB!

BEGINNER TUTORIALS FOR AI-UI

On the page of our AI LAB you will find video tutorials, descriptions and ready-to-download project files for direct import into AI-UI.

Have a look around and find your problem solution in the shown tutorials.

get started now

Trading

Effective trading through AI

The use of big data allows AI to develop models that identify and leverage patterns and trends. AIMS makes it possible to realize fast and flexible adjustments to market changes and make better decisions. All this based on trained patterns, emotionless and according to system.

Image recognition

Reliable labeling and accurate quality assurance through AI.

Artificial intelligence is effectively used to analyze images. By recognizing patterns and objects, it can be used for classification, object tracking and image segmentation. Automated detection of visually detectable defects that are difficult, less reproducible, and time-consuming for human inspectors to identify enables faster and more accurate quality control of your products.

Consent Management Platform by Real Cookie Banner