The rise of so-called “smart factory” technologies is leading to a race to modernize manufacturing plants and warehouse floors. As manufacturers seek to keep up with the competition, old equipment is being replaced by newer, more advanced machinery – and they struggle with higher transfer speeds. According to a study by Plex Systems, 50% of manufacturers accelerated their adoption of automation and digital systems during the pandemic. A separate report from Harris Poll, commissioned by Google, found that by June 2021, two-thirds of manufacturers will be using AI in their day-to-day operations.
Take those numbers with a grain of salt — they’re not from independent sources, after all. (Plex sells manufacturing equipment, Google sells cloud services to industrial customers.) Still, the success of startups like Invisible AI uses AI systems to simulate factory processes.
Invisible AI announced today that it has raised $15 million. CEO Eric Danziger said the platform, which provides on-the-job feedback to operators, is already in use at eight facilities owned by Toyota’s North America division, with eight more deployments planned in the next six months.
“Some of the biggest problems that can be solved by modern AI today are in manufacturing, where safety, quality and productivity are very important,” Danziger told TechCrunch in an email interview. “Everything in manufacturing, from safety audits to continuous improvement cycles, still relies on manual data collection using stopwatches and clipboards. We’re building intelligent solutions that help customers improve their assembly operations.”
Image Credits: Invisible AI
Danziger co-founded Invisible AI with Prateek Sachdeva in late 2018, whom he found working at lidar sensor startup Luminar. The two are using thousands of AI-enabled cameras in manufacturing facilities to monitor workers and large areas, including those that handle materials, mobile conveyors, and autonomous vehicles from one location to another.
Invisible AI technology can track employee activity from the box without an internet connection. Using edge computers and stereo cameras, the combined hardware-software platform has deep insights, enabling it to identify potential safety issues (eg, high-stress injuries) and quality issues across various assembly lines.
Danziger says the invisible AI can track any custom body movement or exercise, or search a factory for things like product ID. A recent new application has been rolled out to the platform allowing customers to monitor the movement of forklifts on the facility floor.
“We want to build solutions that use short-end AI tech, but can be deployed in minutes without any code or data collection — it just needs to work for scale and adoption,” he said. “It all starts with visibility.”
Employees may be right to think about the privacy implications. At worst, the technology can be harnessed for some intrusive purposes, such as enabling managers to punish employees in the name of productivity. Some companies use algorithms to do this, though not necessarily cameras. Amazon’s famous “Time Off Task” system worries warehouse workers about spending too much time away from their assigned tasks, such as scanning barcodes or sorting products into bins.
Unsurprisingly, Danziger was adamant that invisible AI would not support these use cases. He emphasized that the platform is currently unable to perform facial recognition and has an alternative Optional – Real-time face blurring capabilities “for customers who are incredibly sensitive about this topic.”
Image Credits: Invisible AI
“The best part of our system is that it’s completely edge-based and all the video processing and storage is in each camera, and it only leaves the camera when an end user is viewing video using the web interface,” Danziger said. “One hundred percent of customer data is inside their firewalls, which greatly reduces all security risks to this sensitive video data.”
Danziger gave no sign to customers. He can’t However, use invisible AI for surveillance – and US laws do not prohibit this. Each state has its own laws regarding surveillance, but most give employers broad discretion as long as the device used to monitor employees is clearly visible (with the exception of California’s AB-701). There is no federal law that expressly prohibits companies from monitoring their employees during the workday.
However we may feel about that fact, the invisible AI business is expanding – comparing the progress of competitors like Everguard, Intenseeye and Protext AI. Danziger said Invisible AI has 10 customers in markets such as automotive, aerospace and agriculture, and dozens of users per deployment and has applied for military and government contracts.
“Business was initially down during the pandemic, but increased product demand and many labor shortages have led to increased demand for our products,” Danziger said. “We have tremendous customer demand and we are trying to meet this funding now.”
Van Tuyl Companies led Invisible AI’s Series A with participation from FM Capital, 8VC, Sierra Ventures, K9 Ventures and West Coast Capital. That brings the 20-employee startup’s total to $21 million. Invisible AI plans to hire 10 employees by the end of the year.
