Artiquare is pleased to announce its participation in M-AIR – Multimodal AI for Adaptive Intralogistics and Robotics, a research and development network focused on bringing advanced AI technologies into industrial logistics and robotic systems.
M-AIR brings together 16 companies and research institutions to explore new AI-based applications for production, logistics, and industry. The network is led by Cluster Mobility & Logistics [link] and AIR Artificial Intelligence Regensburg [link], based at TechBase Regensburg [link], and is designed to support the digitalization, automation, and connectivity of industrial intralogistics, with a particular focus on solutions that can benefit small and medium-sized enterprises.
For Artiquare, joining M-AIR is a natural step in our work at the intersection of industrial AI, foundation models, agentic systems, digital twins, and automation engineering.
Why M-AIR matters
Intralogistics is becoming increasingly complex. Warehouses, production environments, and industrial sites need to handle changing product variants, fluctuating demand, workforce constraints, and growing expectations for automation and resilience.
Traditional automation systems are powerful, but often rigid. They typically require predefined workflows, structured interfaces, and carefully engineered rules. The next generation of intralogistics systems will need to be more adaptive: able to understand context, combine information from multiple sources, interact naturally with operators, and support decision-making across dynamic environments.
This is exactly where M-AIR is focused.
The network aims to develop multimodal AI models for adaptive robotic systems that can be integrated into both existing and new intralogistics environments. Its focus includes the combination of technologies such as speech control, computer vision, data fusion, digital twins, and large language models. A particularly innovative aspect is the use of LLMs for interactive task description and process control.
From AI concepts to industrial use cases
M-AIR is not only about exploring AI as a technology. It is about translating AI capabilities into concrete industrial use cases.
The network has already identified several promising project directions, including speech-controlled mission definition for mobile robots, intelligent path planning for autonomous vehicles, adaptive bin-picking systems, AI-supported grasp-point detection, generative planning of industrial layouts, digital remodeling of existing facilities, automated safety assessment, and AI-supported product compliance management.
These topics reflect a broader shift in industrial automation: AI systems are moving from isolated models toward integrated assistants, copilots, and agents that can support real operational workflows.
For example, speech-controlled mission definition could make it easier for operators to describe logistics tasks in natural language, while AI-based planning systems translate those instructions into executable robot actions. Computer vision and multimodal models can help robots better understand their environment. Digital twins can provide the simulation and validation layer needed to test actions before they are deployed. LLM-based agents can act as orchestration layers, connecting user intent, enterprise data, physical systems, and automation tools.
Artiquare’s perspective
Sovereignty isn’t about nationalism. It’s about control.
Data sovereignty: Your data stays where you decide. No third-party access without your consent.
Operational sovereignty: Your system runs when you need it. No external dependencies that can fail or be revoked.
Strategic sovereignty: Your technology decisions aren’t constrained by foreign policy.
You can have all three. But not with the default architecture.
Connecting research and application
Our involvement in M-AIR also connects directly with our recent research contribution to the paper “Foundation-Model-Based Agents in Industrial Automation: Purposes, Capabilities, and Open Challenges.”
That work showed that foundation-model-based industrial agents are gaining momentum, especially for user assistance, monitoring, decision support, and engineering automation. It also highlighted that many systems are still at prototype or early validation stages, and that real-world deployment requires further progress in robustness, integration, latency, safety, and evaluation.
M-AIR is an opportunity to move this discussion closer to practice.
By working on concrete intralogistics and robotics scenarios, the network can help clarify where multimodal AI and agentic systems create real value, where conventional automation methods remain essential, and how hybrid architectures can combine the strengths of both.
Looking ahead
The runtime of M-AIR is scheduled from October 2025 to September 2026. During this period, Artiquare looks forward to contributing to the network’s exchange formats, project ideas, and applied research activities.
We are excited to collaborate with the M-AIR partners and to help advance the next generation of adaptive, intelligent, and human-centered intralogistics systems.
About M-AIR
M-AIR stands for Multimodal AI for Adaptive Intralogistics and Robotics. The research and development network brings together 16 companies and research institutions to develop AI-based applications for production, logistics, and industry. Artiquare GmbH is among the project partners.
We’re artiquare. We build reliable multi-agent AI for German industry.
Open source: github.com/artiquare/caa
