Bring intelligence into the physical workflow.
Robotics turns software assumptions into motion, timing, wear, safety, and real-world variation. We connect the intelligent layer to the operational system around it.
Embodied AI combines perception, reasoning, and action in a physical system. Enterprise deployment requires edge inference, controlled orchestration, human supervision, safety engineering, fleet operations, and integration with business workflows.
The integration layer matters
- Perception and sensor pipelines
- Edge inference and latency budgets
- Task orchestration and system interfaces
- Human supervision and exception handling
- Telemetry, fleet data, and maintenance
- Security and bounded operating zones
Robotics experience measured in uptime
Before MH – Applied AI, Marcel Haas led engineering for the technical foundation supporting AutoStore warehouse robotics in the Swisslog environment. Over multiple years, he and his team supported highly available systems at international customer sites and helped keep robot operations available around the clock.
The public CNEXT project record documents the operating context: Warehouse Manager, conveyor and picking systems, structured systems engineering, real-time computing, monitoring, automation, and worldwide 24×7 support. Following KUKA’s 2014 acquisition of Swisslog, Marcel later transferred the systems, operating standards, and working knowledge to the successor operations team in Vietnam.
- High-availability foundations for 24×7 robot operations
- International on-site deployment and service
- Integration across warehouse management, conveyors, and picking
- Monitoring, incident response, runbooks, and operational handover
- Knowledge transfer across teams, countries, and time zones
Misty: embodied AI connected to live business data
Marcel’s current skills context also includes the CNEXT AI Labs Misty project developed by the team he co-leads. The public project record describes the open Misty robotics platform connected to Azure Speech for recognition and output, Azure OpenAI for reasoning and video analysis, and Power Automate plus Logic Apps for live ERP integration.
The prototype recognizes faces, conducts natural conversations, names objects, and responds to real-time business-system data. Misty’s official platform also exposes functionality through HTTP APIs and WebSocket messages and supports AI, natural-language processing, machine learning, computer vision, sensors, and face, object, and speech recognition. This is attributed as CNEXT AI Labs team work rather than a solo Marcel Haas implementation.
- Azure Speech for spoken input and output
- Azure OpenAI for reasoning and video analysis
- Power Automate and Logic Apps for live ERP data
- Face recognition, object naming, and natural conversation
- Open HTTP APIs and WebSocket robotics interfaces
Safety is a specialist discipline.
We do not substitute AI enthusiasm for machinery, workplace, product, or functional-safety expertise. Robotics missions include the qualified specialists and approvals their physical risk requires.
Start where software meets motion.
High-value first work often sits in perception, operator assistance, simulation, fleet intelligence, or business-system integration before autonomous movement expands.
Bring us the mission, not a shopping list.
In the first conversation we map the operational outcome, constraints, deployment environment, and the shortest credible path to evidence.
Book an AI deployment consultation