The Global Community for Artificial Intelligence (AI) in Master Data Management (MDM) is designed to foster systematic knowledge transfer and exchange with other companies, researchers and experts.
Community members are provided with continuously updated content, comprising interesting lectures, latest research information, innovative use cases, lessons learned and how-to guidance.
Furthermore, design thinking workshops will be conducted in Europe and the United States of America to ensure continuous inflow of new ideas. The workshops also offer the chance to share thoughts and ideas as well as challenges that will be discussed within the community. Benefit from this great opportunity to find partners for co-innovation joining forces in the endeavor to bring first AI & MDM light house uses cases to life.
- Latest thinking, discussions and exchange for Artificial Intelligence in Master Data Management
- Exclusive insights for community members as well as guest lectures from AI experts and AI pioneers
- Experience AI through joined realization of selected use cases with latest AI technology
- 11.07.2019 Webinar: Rapid prototyping – How to create innovations customers love, Registration
- 01.08.2019 Webinar: Machine learning in business language, Registration
- 05.09.2019 Webinar: CRISP for value chain management, Registration
Community Members: MDM and AI Experts from the Following Companies
- B. Braun
- Böhringer Ingelheim Bosch
- Cardinal Health Colgate Palmolive
- Daichi Sankyō
- DB Schenker
- Eli Lilly
- F. Hoffmann-La Roche
- FMC Corporation
- Fresenius Medical Care
- General Mills
- Hartmann AG
- Johnson & Johnson
- Klingel Group
- Nestle Skin Care
- Norsk Medisinaldepot
- Paul Hartmann AG
- Philip Morris
- Procter & Gamble
- Rudolf Wild
- Roche Diagnostics
- Roche Pharma
- Takeda Pharmaceutical
- Tetra Pak
- Thyssen Krupp
- Vaillant Group
- Weidmüller Interface
Video Series – Ask the MDM expert
What is Artificial Intelligence and how will it affect MDM?
Artificial intelligence is a computer science that enables machines to mimic cognitive human behavior. As human behavior is quite complex, AI is typically divided into several disciplines e.g. robotics, planning, problem solving, language processing, machine learning, etc. Some of those capabilities will have no/small impact on MDM, some will have huge impact. However, if scientists were able to put all disciplines together into one general artificial intelligence machine, this machine would be able to perfectly mimic human behavior. Although this sounds futuristic and there are controversial discussions whether a generic artificial intelligent machine is possible at all, we will see the impact of AI in MDM with first use cases soon to be realized.
The new AI capabilities will change all areas of MDM practice.
Capability to solve complex problems
- Identification of typical user errors and reasoning why errors appear
- Considering transactional errors, frequent data changes, bug data etc. for autonomous development of standards
Capability to express emotions
- Applications that are able to detect how we are feeling and adjust accordingly
- Such applications could sense nonverbal behavior to become mood-aware to help uncertain/frustrated users
- Assessment of correlations and causality between mood and data quality
- Usability improvements based on measured customer satisfaction
Capability to create new/inventive ideas
- Ideation process for MDM related topics could be supported by computational creativity
- In combination with problem solving it could lead to innovative approaches for MDM issues
Capability to understand human language
- Deriving master data attributes from free text
- Replacing mouse and keyboard with voice command leading to new user experience and usability of MDM application
Capability to learn from experience
- Ensuring compliance of governance rules, principles and standards as steward
- Autonomous data maintenance based on training and experience
- Considering transactional errors for autonomous data corrections
Capability to plan towards a certain goal
- Use algorithms to plan data maintenance activities in advance
- Adjust and plan data cleansing activities based on upcoming transactions
Capability to control motion
- Not expected to have huge impact on data management itself
- However it could lead to new object definitions
Capability to create an ontological model
- Identification of the information existing in the company by describing objects, entities and relationships
- Building ontological model of the information by linking all elements to an information network
- Identification of missing information
Capability to perceive like humans
- Deriving master data attributes from pictures, videos and sound
- Replacing mouse and keyboard with gesture control leading to new user experience and usability of MDM application
Machine learning is typically implemented via neural networks. One good example for AI are artificial neural networks that are used to realize machine learning. Neural networks are mathematical models that mimic the behavior of neurons to perform complex tasks. It enables a machine to learn based on observational information. Artificial neural networks can be utilized for a variety of activities such as clustering, classification or pattern recognition. Learning happens via adjustment of bias and weights in training iterations. Thus the logic does not need to be defined upfront (like in classical programming) and the network does learn on its own how to solve tasks.
Example AI use cases for machine learning in MDM:
- Data plausibility checks based on unstructured and transactional data
- Predict data extensions and proactively maintain data
- Clustering: For example to define optimal set of payment terms based on customer/supplier and transactional data
- Chatbot to guide users and to explain data dependencies for improved usability of MDM tools. Also directly acting as data steward.
- Automatic data record linking/mapping and duplicate detection with no need for harmonization
Selected AI use cases for MDM
Based on the CAMELOT Artificial Intelligence Innovation funnel many use cases have been identified which solve information management challenges posed by AI. In three selected examples, CAMELOT outlines the relevance of AI for master data management and what solutions could look like.
Personal assistant that guides and assists users in the master data system
Research forecasts that chatbots will be responsible for cost savings of more than $8 billion per year by 2022, up from $20 million this year. Less (or even no) end-user training and support will be required as chatbots will guide users and answer MDM and tool related questions. AI speech and AI gesture control will allow to process commands and to capture nonverbal feedback. AI machine learning will process the information and will have access to the AI knowledge representation of the company.
Data quality watch guard that performs data plausibility, correctness and duplicate checks
AI will check for duplicates and data quality of records based on attributes as well as information outside of the MDM application such as: transactional data, CAD / PLM data, SOPs, etc. AI machine learning (e.g. a deep neural network) will perform the correction of typical/ individual / regional errors having access to the history of past changes. In combination wit AI reasoning and AI knowledge representation the solution allows to investigate the root cause of DQ issues autonomously.
AI enabled maintenance of contact person information based on business cards
AI machine perception will allow to read the business card. AI natural language processing will help to link the information provided with attributes like street, postal code, etc. AI machine learning enables to identify the company and to improve the performance overall.