AI Agent with AR Interface
The core of the system is an LLM-based AI agent that goes far beyond simple object recognition. It accepts multimodal input, analyzes context, queries external data sources, and generates responses as AR scenes of any complexity.
5-Step AI Cycle
From voice command to interactive AR scene in milliseconds.
Input
Voice or text question in any language, or camera pointed at equipment
Analysis
LLM analyzes the request with context: VPS position, interaction history, sensor data, ERP state
Agentic Mode
AI autonomously determines which external data is needed: parts catalog, ERP, technical docs, knowledge base, sensor telemetry
ECS Generation
AI generates the response as an ECS structure — entities with position, geometry, material, animation, annotation components
AR Visualization
ECS structure is transmitted via WebSocket and rendered as an AR scene: 3D models, highlighted nodes, annotations, animations
Talk to It Like an Expert
The AI agent understands context, language, and spatial position — and responds with interactive AR scenes, not text.
«Show me all parts that need replacement during maintenance-500»
AI queries the maintenance schedule, finds the parts list, generates an ECS scene highlighting each part on the machine with catalog annotations.
«Was ist mit diesem Teil los?»
AI recognizes German, analyzes camera direction via VPS, identifies the component, queries sensor telemetry, generates a diagnostic AR report in German.
«Compare this part with an alternative»
AI finds the original and alternative in the catalog, generates an ECS scene with two 3D models side by side, difference annotations, and price comparison.
AR Responses as Entity Components
Every AI response is a set of entities, each with components that define how it appears and behaves in augmented reality.
Timer
Text
List
Buttons
Navigation
Marker
Detect
Media
3D Object
Audio
Static Instructions vs. Dynamic AI Generation
- Manually authored static AR overlays
- Fixed content for each equipment node
- Single language, no context awareness
- Months to create and update content
- Cannot answer unexpected questions
- AI dynamically generates AR scenes on the fly
- Context-aware: knows your position, equipment state, ERP data
- Any language — voice, text, or camera input
- New content generated instantly, no authoring needed
- Answers any question with an interactive AR visualization
See the AI Agent in Action
Schedule a demo to experience how the AI agent generates interactive AR responses in real time.
Request a Demo