From research to the floor
The transformative potential of AI for industrial core processes is well established. Only a fraction is realized in operations. We close this gap: with research directly at the process and implementations that arrive on the floor.
Carried by
01 / 07 · Starting point
The transformative potential of artificial intelligence for industrial core processes is considered empirically proven. Yet only a fraction of it is realized in operations. This discrepancy between theoretical potential and actual implementation is the center of our project.
Together with Stanford, TU München and UnternehmerTUM we therefore developed a project format that aims at impact, not analysis: through embedded research we capture the operational reality of end to end workflows inside the company and carry identified potential into real implementation through design partnerships.
of companies already use AI in at least one business function. The potential is recognized, adoption is everywhere.
Source: McKinsey, Superagency in the Workplace, 2025
of leaders call their company’s AI deployment mature: fully integrated into workflows, with measurable business outcomes.
Source: McKinsey, Superagency in the Workplace, 2025
02 / 07 · The project
Phase 1 makes the diagnosis in your processes. Phase 2 turns it into a deployed solution. You decide after each step whether to continue.
Phase 1 · Research
A research project investigates which challenges genuinely block AI implementation in industrial core processes and which applications demonstrably create value. The output is a structured analysis presenting realistically implementable AI use cases in your processes.
Phase 2 · Design partnership
In close cooperation with UnternehmerTUM we offer selected partners from phase 1 a design partnership with the research team. The highest impact use case is selected together and developed and implemented to a deployed solution.
03 / 07 · Method
We capture the operational reality of your workflows directly at the process, through immersion and shadowing, without pulling your teams out of their normal working mode.
Immersion and shadowing directly at the process, during live operations. No workshops, no questionnaires: we are where the work happens.
We need access to relevant processes and contacts during the stay. Nothing more. Your team keeps working as always.
Full NDA and safety induction. All company related information is evaluated in anonymized form only and never shared with third parties.
A closing meeting with your decision makers: real challenges, prioritized AI potential, impact assessment per process.
04 / 07 · Your outcome
The ratio is deliberately asymmetric: you open the door, we bring the research team, the method and the analysis.
What you contribute
What you get
Realistically implementable AI potential in your specific context. Focused on implementation, not theory.
Concrete implementation recommendations with an impact assessment per process.
Joint development and implementation of the highest impact solution in phase 2.
05 / 07 · Where we stand
This project is not an announcement. We are already researching inside partner plants across automotive, medtech and precision manufacturing, and the first design partnerships are forming.
01
Fieldwork
RunningDays on site, during live operations: maintenance, fault resolution, non standard interventions. Further phase 1 slots are open.
02
Finding
ValidatedAcross all conversations and plant visits, one application carried the highest confirmed value: capturing expert knowledge at the process and making it accessible.
03
Phase 2
StartingWith several partners we are moving from research into implementation: from the highest impact use case to a deployed solution on their lines.
Partner companies
Automotive · Medtech · Precision Manufacturing
Names disclosed under NDA
Research partners
TU München · Stanford University · UnternehmerTUM
06 / 07 · Team
A team out of TU München, with research at Stanford, Cambridge, MIT, IBM and Fraunhofer. Advised by people who have built and scaled industrial software before.
TU München · Stanford
Management and CS at TUM. Research on industrial energy flexibility at Stanford. AI implementation work at Langdock (Trumpf, Würth).
TU München · IFM Cambridge
Management and CS at TUM. Research at IFM Cambridge and MIT on AI in manufacturing. Built an enterprise grade AI system inside the TDK incubator.
TU München · Informatics & Physics
Informatics and Physics at TUM. ICML published machine learning work. Currently VP at TUM.ai. Quantum ML research at Fraunhofer and IBM.
Advisors
Thomas Kirchner
Co founder and former CEO, ProGlove
Prof. Thomas Bohné
Founder, Cyber Human Lab, University of Cambridge
Dr. Slavomir Tadeja
Postdoc, MIT LEAP Group, workforce development
Research and training environments
07 / 07 · Contact
For phase 1 we are looking for a small group of manufacturing companies where the research team can spend two to four days on site. The effort for your team stays deliberately minimal. The results are yours.