Process optimization and data networking -
The central key topics for small and medium-sized enterprises
For companies, the use of artificial intelligence often means a change in their processes and presents them with challenges. Artificial intelligence is no longer only relevant for "global players", but is becoming increasingly important for medium-sized companies. Artificial intelligence makes it possible to adapt products and services to customer needs and create new business models. There is great potential in design, sales and predictive maintenance.
However, mid-sized companies often face challenges in introducing AI due to lack of know-how, skilled labor, high investment, complexity and lack of data. Technical skills, infrastructure, and company alignment influence preparation for AI deployment. Concerns about high barriers to entry, data sovereignty, assurance and certification of AI systems hinder adoption. There is great potential in midmarket data, but adoption can be challenging.
"AI-as-a-service" is recommended for companies with skills shortages or a small data base. NVIDIA and VMware offer AI for Enterprise Platform, a cloud-native software suite for AI and data analytics optimized, certified and supported by NVIDIA. The suite includes key NVIDIA technologies for rapidly deploying and scaling AI workloads in the hybrid cloud.
AI in production
Focus: Value chain optimization using AI-based applications
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Potentials along the value chain
The opportunities for SMEs include efficiency gains, advantages over competitors, and increased profits and earnings.
AI applications according to different value creation areas
Text processing
Cluster analyses
Cluster analyses
Sales & Logistics
- Warehousing, sorting, delivery by autonomous vehicles/robots
- AI-based demand and routine planning
- Route planning & delivery
- Warehouse logistics and picking
Production
- Automation technology
- Anomaly detection
- Predictive maintenance
- Cognitive assistants
- Further development of smart products for new business models
- Robotics
- Maintenance
- Quality management and control
- Process optimization and control
Supply chain
- Supply chain optimization
- Intelligent sales forecasting
- Demand forecasting to predict sales
Procurement/purchasing and ordering
- Inventory & Inventory Management
- Automated warehousing through autonomous vehicles
- Supply chain operations/ AI-based processing: taking over from order process to delivery
- Procurement market research
- Resource planning
- Supplier management
Corporate infrastructure, administration and human resources
- Takeover of routine tasks
- Partially automated applicant management
- Document management
- Fraud detection
- Forecasting
- Automatic document recognition
- Intelligent dunning procedures
- Energy management
- Staff scheduling
- Personnel Recruiting
- Facility Management
Service and customer management
- Automated customer review analyses
- Support for customer interaction (e.g. chatbots)
Research and development
- AI-supported simulation of product behavior
- Analyses for product development
- Product and process development
Marketing and sales
- Automated data collection and analysis
- AI support for customer interaction
- Dynamic price optimization, product portfolio optimization
- Targeted advertising/promotion
- Chatbot
- Visual product search
- Consulting/sales robots
- Sentiment analysis in customer service
- Digital assistance systems
Quality control and assurance
- Visual inspection of components for defects
- Predictive quality; optical and acoustic quality assurance
How can companies benefit from the use of AI
in industrial value creation?
Focus: Production
- Predictive analytics (e.g. monitoring and maintenance of production facilities)
- Optimized resource management (e.g. optimization of production and manufacturing schedules)
- Quality control (e.g. checking the condition of components)
- Intelligent assistance systems (e.g. support in manufacturing processes), value creation potential of AI in production
- Knowledge management (e.g., data models for complex engineering processes)
- Robotics (e.g. learning, self-regulating gripper systems)
- Autonomous driving and flying (e.g. driverless transport systems)
- Intelligent automation (e.g., automation of routine processes in manufacturing and assembly) and
- Intelligent sensor technology (e.g. pre-processing of data when monitoring production facilities).
Predictive maintenance uses data analysis and algorithms to make predictions about impending shutdowns and calculate optimal maintenance times to prevent downtime. Combining past maintenance experience with current data analytics creates a computer- and AI-assisted prediction of maintenance. This allows technicians to be equipped with the right spare parts at the right time, resulting in reduced maintenance activities.
Asset Performance Management (APM) comprises three technology components: condition-based maintenance (Asset Health Insights) based on digital IoT data, predictive maintenance using data analysis, and Equipment Maintenance Assistant, an AI-based support for maintenance technicians.
Another application example for AI technologies in process optimization using machine learning is the analysis of extensive data sets from complex systems, such as interlinked production lines. In complex systems in particular, there is the potential to uncover relationships that are not directly apparent to human observers because various preconditions occur that are beyond the observer's control. This can be used, for example, to uncover the causes of wandering bottlenecks in production. However, the successful use of AI depends on the quality of the data with which it has been trained in advance. It is only as good as the data on which it is based.
A distinction is made between perception-decision and execution assistance systems;
- Perception assistance systems improve information intake, from classic bills of materials to digital work instructions.
- Decision assistants support complex decisions through prioritization and data analysis.
- Execution assistants facilitate work through machine support, from simple hand pallet trucks to human-robot cooperation.
Data and AI-based value networks: potentials for SMEs
Platform
Suppliers
SME or large company
Company
SME or large company
Customer
Cloud service provider
- Data streams
- Service flows
- Payment flows
Importance of value networks
The use of AI leverages data from digitized objects, systems and environments to gain valuable information in real time. Data types such as operational, environmental or usage data are provided and combined, processed and analyzed on digital platforms. This enables scalable, individualized offers for customers through direct and indirect network effects.
Implementation of a value network
- Transparent network structure with clearly formulated value proposition of all parties involved
- Definition of a clear data strategy that determines the quality, relevance and availability of the data required for value creation
- Strategic and long-term (research) cooperation to generate own competencies in the area of data science
- Agile roll-out of the product to identify hurdles at an early stage and to be able to adapt the product
- Continuous review and adaptation of the business model and suitable financing models, e.g. revenue sharing
- Poorly communicated and coordinated projects
- Collecting data without clearly defined criteria and methods
- Complete outsourcing of AI competencies to external companies without continuous exchange and mutual coordination
- Rigid schedule setting and sticking to the original concept
- Adherence to classic business models and convincing network participants to bear the investment risk themselves