With $7.5 million in seed funding, the German business Juna.ai GmbH, which uses artificial intelligence to increase factory efficiency, started today.
The investment was spearheaded by Kleiner Perkins. Prominent venture capitalist John Doerr, the Stockholm-based fund Norrsken VC, and other sponsors joined it.
Matthias Auf der Mauer is the CEO of Juna.ai. Sensirion Holding AG, a publicly traded sensor manufacturer, purchased his prior industrial software firm. Christian Hardenberg, the company’s chief technical officer and co-founder, was formerly employed at Delivery Hero SE in a similar capacity. A hosted AI platform called Juna.ai claims to be able to assist firms in figuring out how to run their manufacturing lines more effectively.
Applications from SAP SE, Snowflake Inc., and other enterprise software vendors can be integrated with the company’s platform. It gathers technical information from clients’ plants, including equipment temperature readings, via such integrations. The raw data is then transformed by Juna.ai’s AI algorithms into dashboards that show where a factory’s operations need to be improved.
A few aspects of the platform are designed to assist manufacturers in increasing the reliability of their output. A factory’s daily production volume, the rate at which manufacturing flaws appear, and other relevant metrics can all be measured by Juna.ai. Additionally, the software produces suggestions for enhancements.
Customers can also use Juna.ai’s platform to reduce the energy consumption of factories. The software breaks down electricity usage by product for manufacturers who make a variety of goods. Additionally, Juna.ai detects instances in which a production line uses more electricity than normal and offers technical information that can be utilized for troubleshooting.
The business claims that clients can use the data generated by its platform to comply with ISO 50001 standards for their plants. This compilation of best practices aims to assist manufacturers in reducing their electricity consumption. Furthermore, Juna.ai creates reports from the power usage data it gathers that can be used to prove adherence to production standards.
Custom AI models power the platform’s internal operations. According to the company, it used a technique called reinforcement learning to create its models. Trial and error is used to train neural networks based on this technology: They get positive feedback following each learning milestone and practice a task repeatedly to figure out the best approach to complete it.
To create AI agents, Juna.ai refines its models using data from clients. These are specialized neural networks that require little or no human intervention to complete tasks. Each of Juna.ai’s agents is made to track and improve a distinct set of industrial KPIs, the company claims.
At the moment, the company employs six people. The recently disclosed seed round will be used by Juna.ai to expand its clientele and hire more employees.