AI Solutions for SME

"The question is not whether AI is worthwhile,
but rather in which processes it should be applied."

Ask us!

Book a meeting with us:

Or use our  contact form 

Physical AI

KI4KMU

Automotive SW 

NISAR – Ai for SME expertise from Munich in the in the picturesque Upper Bavaria

NISAR automates complex industrial processes with multi-agent AI and a human-in-the-loop approach, reducing manual effort and lowering costs. Advanced AI methods such as GenAI, LLMs, and knowledge graphs enable data integration, and moderator agents allow interaction in natural language.

 

Together we analyze your prcesses and identify potential for optimization by digitalization and AI. We consult in elaborating the solution and offer tailor-made solutions for your processes: 

  • We combine AI expertise with many years of automotive industry experience. The founders have published on safe AI applications in collaboration with TH Ingolstadt at BAY.AI as well presented at TÜV Safe.Tech, focusing on quality assurance and requirements management.

  • We support through our custom-developed tools, through consulting, and training to simplify development and operational processes. This includes IoT, connectivity, and cloud solutions to connect control units and vehicles to the platform, enable remote debugging, and facilitate remote data collection in factories.

Bild: Pixabay, designerpoint, Michael Siebert

We are active in the following organizations:







Services and Products

AI4SME: PAMAS – Process Automation with Multi-Agent AI System

Industrial processes are complex and often span multiple domains and expert roles. 

In combination withe the analysis of the processes we offer a solution for a safe and accurate automation of industry processes using our Multi-Agent System (PAMAS) framework, which is designed to adapt to and integrate into your processes.

AI engines only unfold their full potential in industrial settings when built upon a reliable ontological knowledge structure. While pure language models without context are prone to hallucination and incorrect outputs, coupling them with a domain knowledge model guarantees maximum precision and auditability.

Furthermore, a structured knowledge model seamlessly unifies disparate knowledge sources—ranging from technical datasheets and maintenance logs to ERP systems.

The Multi-Agent AI system mirrors and automates industrial processes. Implicit knowledge is captured and modeled in a knowledge graph. The agents retrieve process data from complex data sources. Human experts are provided with these AI agents to assist them in their daily tasks, thus, the experts remains ‘in the loop,’ shifting from ‘doing’ to “reviewing”, gaining capacity to focus on corner cases.

 

IoT Edge Box

Even today, data from remote sources is often not accessible. Be it the physical distance or the devices not being network-enabled at all. With our Edge Connect Box we tackle this challenge: it connects data sources such as devices, ECUs, machinery or vehicles to the cloud platform, making their data available for further analysis as well as enabling remote maintenance and remote debugging. Even remote pre-integration of vehicle ECUs is possible.

Automotive R&D Tools

Automotive R&D and Test Tools

  • NIMsys – Networked Intelligent Mobile System for HAD Testing – Connecting simulation and proving ground
  • AIBOX – Automotive IoT Box for Connectivity, Debugging and Testing
  • INtest – Automotive ECU Network Injection Testing
  • Scenario Recorder – Event-trigger for Scenario Recording
  • ROS Live Browser – ROS architecture and data flow visualization over-the-air
  • ADS – Architecture Design Studio (Tool and Middleware)

Learn more…

 

 

 

Highlight Projects

Multi-Agenten-System (MAS) für die Automatisierung von Qualitätsprozessen in der Halbleiterfertigung NISAR wurde beim Open Bosch Award 2024 als eines der TOP 5 Start-ups unter mehr als 100 Unternehmen in der GenAI-Meisterschaft ausgezeichnet. ■ Abwicklung komplexer Halbleiterfertigungsprozesse und -daten ■ Vollständige On-Premise-LLM-Lösung (lokale Serverlösung) ■ 100 % datensicher ■ Kosteneffizient und schnell ■ Bereit für die Integration in Altsysteme (Legacy-Systeme) ■ Qualitätskontrolle und Vorhersehbarkeit ■ Allgemeine Anwendungsmöglichkeit für F&E (Forschung & Entwicklung), Fertigung und Geschäftsprozesse

Multi-Agent System (MAS) for Semiconductor Manufacturing Quality Process Automation


NISAR was recognized as one of THE TOP 5 startups under 100+ companies for GenAI Championship at the Open Bosch Award 2024
■ Handling of complex semiconductor manufacturing processes and data
■ Fully on-premise LLM solution
■ 100% data secure
■ Cost efficient and fast
■ Ready for integration with legacy systems
■ Quality control and predictability
■ General application possibility for R&D, Manufacturing, and Business Processes

Zwischen den Entwicklungsschritten ermöglicht die KI-gestützte Workflow-Automatisierung eine effiziente und präzise Wissensübertragung. ■ In jedem Entwicklungsschritt analysiert die KI-Engine die Informationen und schlägt nachfolgende Schritte vor – basierend auf Daten aus vorherigen Schritten und aktuellen Entscheidungen. ■ Verknüpfung verschiedener Tools (IDEs, JIRA usw.) und Ermöglichung der Workflow-Automatisierung durch Large Language Models (LLM). ■ Anpassung der KI-Engine durch Optimierung mit dem privaten Wissen der Nutzer und den spezifischen Entwicklungsaufgaben. ■ Die KI-gestützte Engine umfasst Web-Frontend, Tool-Plugins, Backend-Datenbank, KI-Prompts, private Wissensdatenbank und eine Schnittstelle zum (lokalen) LLM. ■ Der Prozess bleibt durch die Mitwirkung und Bestätigung von Experten sicher und funktional zuverlässig, um eine KI-„Blackbox“ zu vermeiden. ■ Das interaktive Dialogsystem lässt nicht nur den Experten teilhaben, sondern Studien zeigen, dass dadurch auch die Leistung des LLM verbessert wird.

AI-Assisted V-Model

Between the development steps, the AI-assisted workflow automation enables efficient and accurate knowledge transmission
■ In each development step the AI engine analyses the information and proposes succeeding steps based on data from preceding steps and from current decisions
■ Connecting various tools (IDEs, JIRA, etc.) and enabling workflow automation through Large Language Models (LLM)
■ Customization of AI engine by optimization with users’ private knowledge and development tasks
■ AI-assisted engine includes Web-Frontend, Tool-plugins, backend data base, AI prompts, private knowledge base and interface to (local) LLM
■ The process remains securely and functional safe through expert participation and confirmation to avoid any AI black box
■ The interactive dialogue system not only let’s the expert participate, but studies show that also to performance of the LLM is improved

 

 

About us

The NISAR team combines a strong academic beckground in Ai, Robotics and CV with many years of experience in IoT and Cloud technology as well as in the automotive industry, in  SW and system engineering as well as manufacturing and homologation. We also decided to build our team immediately with two pillars, simultaneously in Germany and in China, the most dynamic market at the moment.

Our mission is to increase efficiency in industrial processes such as automotive and R&D (interview at Bayern Innovativ), quality and manufacturing through a modern cloud-based tool landscape for automating knowledge-driven processes with a strong focus on usability and efficient workflows.

 

 

 

Success Stories: Our Partners

Research Background

Industrial Expertise

Intrigued?

Feel free to contact us: info@nisar.ai

The Locations

Currently we are operating from two locations to cover the two biggest markets, Munich for EMEA and Shanghai for the Asia Pacific region.

 

Wir unterstützen die IHK-Initiative Pack ma’s digital, um der Wirtschaft in Oberbayern bei der Digitalisierung weiter zu helfen und die Zukunft des Standorts zu sichern.

Wir sind aktives Mitglied bei Bayern Innovativ und fördern Digitalisierung, Transformation und den Wissenstransfer zu angewandter KI etwa durch Vorträge und Events.

Als aktives Mitglied des Bayerischen Unternehmensverbands Metall und Elektro e. V. unterstützen wir die bayerische Industrie beim Kompetenzaufbau im Bereich digitaler Transformation.

In Kooperation mit TÜV SÜD tragen wir mit Fachvorträgen und Weiterbildungen zur Stärkung von Qualitäts- und Sicherheitsstandards in der Technologiebranche bei.

The Multi-Agent AI system mirrors and automates industrial processes. Implicit knowledge is captured and modeled in a knowledge graph. The agents retrieve process data from complex data sources. Human experts are provided with these AI agents to assist them in their daily tasks, thus, the experts remains ‘in the loop,’ shifting from ‘doing’ to “reviewing”, gaining capacity to focus on corner cases.

Learn more…