Industrial AI startup Octonomy raises $20M in seed funding

Octonomy AI GmbH, a startup with an artificial intelligence platform for companies that operate heavy machinery, today announced that it has raised $20 million in funding.

Macquarie Capital Venture Capital led the seed round. It was joined by German bank NRW.Bank, TechVision Fund and Capnamic. The latter firm previously led a $5 million round for Octonomy in January.

The technical data that engineers require to troubleshoot a malfunctioning machine is often scattered across multiple systems. A robotic arm’s schematics and recent error logs, for example, might be stored in two separate applications. After identifying the component that’s causing the malfunction, engineers may have to log into a third application to determine whether a replacement part is available. 

Octonomy’s AI platform promises to automate much of the data review process. According to the company, its software uses an ensemble of AI agents to answer technical questions about industrial equipment. It claims the agents generate accurate responses more than 95% of the time.

Another selling point of the platform is that it’s relatively simple to deploy. During the setup process, companies don’t have to load schematics and other technical records manually into the software from the systems where those files are usually stored. As a result, it can be installed in under three weeks.

Octonomy accesses technical documentation by connecting to a company’s enterprise resource planning platform and other internal applications. It aggregates the data in a centralized repository to ease analysis. From there, Octonomy normalizes records and performs chunking. That’s the process of breaking up large documents into smaller files that can be processed more efficiently. 

Companies can organize their data into categories. For example, a manufacturer could create a repository that only contains data about the equipment installed in one specific plant. According to Octonomy, organizing data in this manner prevents its AI models from incorporating irrelevant information into prompt responses. 

The platform uses three sets of agents to answer user questions. The first agent type, which is called Support, is designed to answer relatively simple queries. There are also Consultancy agents, which can perform more advanced tasks such as explaining how two machine components interact with one another.

Prompt processing is coordinated by a third program known as the Supervisor agent. It analyzes each user request and determines which agent is best suited to generate an answer. 

“Our agents deliver verified 95%+ percent response quality, empowering teams often accustomed to 50% accuracy rates with standard AI platforms,” said Octonomy co-founder and Chief Executive Officer Sushel Bijganath. “With Macquarie Capital Venture Capital at our side, we gain an international partner to make Octonomy broadly available across Europe and the USA.”

Photo: Octonomy