The central user interface
We provide a standard user interface like ChatGPT. This is your development and deployment center for AI automations. You control the system with natural languages
The plugin system
Plugins provide access to your data and extend the capabilities with customized AI models
Usability is a king
Get started page
- Prompt templates for AI communication
- Best practices for different use cases
Voice-guided prompt tailoring
- Speak to the AI in natural speech
- The AI will ask questions step by step
- Finally it takes all information and build a perfect prompt
The agent system for task automation
Via an agent system, checks and evaluations can be automated, for example those of the vibration behaviour or the evaluation of a shift
Automated machine monitoring
Sound processing
Up to 24 Piezo sensors build a sound grid for the detection and localization of machine failures. The results are visualized directly on the digital twin of the machine
Image processing
Up to 24 cameras capture every detail of the current state of the machine and all mechanical settings
NeRF for Live 3D
The system can handle up to 24 cameras per machine. That is enough to capture all details and build a 3D live model in real time. So the user can travel in time and space to see what happened
Blender for AR/VR applications
Allows parts to be searched and viewed in the integrated 3D model, with a VR/AR approach to present all information
Data analysis
Generative data scientist
- Generate data loading functions
- Generate data processing functions
- Generate plots
Generative how-to
- Give instructions for tasks on the machine
- Explains the work step-by-step
Information retrieval
Generative AI revolutionizes on-the-job training, sculpting tailored learning experiences from vast data, ensuring every employee reaches their full potential
Semantic search for documents
The system swiftly pinpoints and delivers the exact knowledge of the machine professionals need to operate
Visual question answering
The system understands visual concepts, transforming provides images from the shop floor into insightful knowledge
Domain knowledge teaching
Trained on specialized domain knowledge the system is ensuring professionals mastering of the nuances of the field
Detection algorithms
Specialized algorithms process the machine data and give analytical insights for the time series
CONDITION MONITORING
Finds clusters and anomalies in multi-variable data and builds score to simplify the information
PREDICTIVE MAINTENANCE
Track the development of the machine and find damage indicators
DEPENDENCY FINDER
Correlates data, applies process stability and important analysis
Compliance features
-
Plug in local AI models, which are automatically used
-
Log files for documentation and control
-
Multi User Interface with shared and private tabs
-
Content filtering by local LLM to block sensitive and GPR-related data for OpenAI and Microsoft
-
Auto-translate with local model in 200+ languages in text and speech
Integration in 4 phases
The project progresses through four major phases, starting with the adaptation of an “octopus” software prototype to a machine and ensuring data and IT compliance. This foundation enables the development and refinement of the software through alpha and beta stages, incorporating stakeholder feedback, rigorous testing, and broader machine integration. The project culminates in a comprehensive rollout, with a timeline highlighting a prototype implementation in Brazil and wider beta deployment by Q1/23
Finding the right AI development strategy takes TIME and EFFORT
With Metric Space it takes less time