Google Cloud's first product tailored for the manufacturing industry is far from groundbreaking. But for the third-place provider, the effort to win against AWS and Microsoft in the sector goes beyond just releasing new tools.
Google has a reputation for suddenly killing off products, which may be acceptable in the consumer market where options are usually more plentiful and prices much lower; in enterprise technology, however, it's a major problem. Companies don't want to invest the time and money it takes to deploy software only to find out soon after that it's no longer supported by the vendor.
Under CEO Thomas Kurian, Google Cloud has been trying to change that stigma, including putting a more rigid process in place for cancelling products, according to managing director Dominik Wee. It's also rolling out dedicated sector offerings, a testament to the focus on a core group of industries.
On Tuesday, the company released Visual Inspection AI, a system to detect product defects. While other vendors offer similar solutions and the tech giant has been using it internally for years, Google Cloud is banking on its ease of use as the key differentiator.
"It isn't new. What we are offering in concept isn't new at all," Wee, who leads Google Cloud's manufacturing efforts, told Protocol. But, "by making the learning easier … you can scale it out much more. With the previous approaches you needed highly-trained specialists. Now, this technology is so easy to use that the people you have on the shop floor can use it."
Unlike past systems which could require hundreds of photos of defective products to train the models, Google says its tool can go into production with very few examples. That incremental benefit shouldn't be ignored, Wee argued, particularly for a sector that is already well underway in adopting advanced technology — a push that insiders refer to as "industry 4.0." As companies scale out use of AI-backed inspection systems like Google's, for example, they can use the data generated to link damaged products to broader process improvements.
"Most manufacturing companies have come to a point where they've run out of ideas," he said, referring to the ceiling that companies are hitting in achieving double-digit investment returns on projects. "If you're a plant floor manager, you're not looking for 50% improvement. You're looking for the next 2%, and the next 2%."
Alongside building a product roadmap, Google Cloud has made other important steps to support a push into the manufacturing industry. It is hiring top talent like long-time SAP executive Hans Thalbauer and Accenture's Suchitra Bose. The company is making a purposeful push to nab SAP customers; the German software firm has huge penetration in the industrial segment. And Google Cloud struck a key partnership with Siemens, one of the industry's most important providers of factory tooling that is now trying to position itself as an automation player.
While many manufacturers have already invested in digitization, tech like digital twins or iterative manufacturing that shows promise but is far from wide adoption is poised to significantly expand the amount of data, often from the factory floor, that industrial companies will need to store and analyze quickly and constantly. That requires a combination of powerful AI algorithms with edge computing tech. And that's why Google Cloud sees such a big opportunity.
"There's enough value in it for us by the data sitting in Google Cloud," Wee said. "We don't see our play as moving into the application. Other people do that much better than we do. For us, this is entirely a play around complementarity and a bet on future data priorities."
But Microsoft and AWS see the same opportunity. Back in February, AWS released a similar AI-powered defect detector for manufacturers. Still, Google Cloud's goal is broader than providing manufacturing-related cloud and AI tools. It's hoping to establish itself as a "connective tissue" for the industry, one that can bring together the worlds of IT and operational technology, referring to products like internet-enabled sensors that monitor machines that Siemens and others sell.
"They're very foreign to each other," said Wee. "The big opportunity is them coming together. It's been discussed for a long time, but visual inspection demonstrates how we are taking the next step on the journey. It exemplifies where we want to go in the long term, we want to close this gap."
