Google Glass lives—and it’s getting smarter.

On Tuesday, Israeli software company Plataine demonstrated a new app for the face-mounted gadget that understands spoken language and offers spoken responses. Plataine’s app is aimed at manufacturing workers. Think of an Amazon Alexa for the factory floor.

The app points to a future where Glass is enhanced with artificial intelligence, making it more functional and easy to use. Plataine, whose clients include GE, Boeing, and Airbus, is also working to add image-recognition capabilities to its app.

The Israeli company showed off its Glass app at a Google conference in San Francisco promoting the company’s cloud computing business. It was built using AI services provided by Google’s cloud division, and with support from the company. Google is betting that charging other companies to access AI technology developed for its own use can help its cloud business draw customers away from rivals Amazon and Microsoft.

Jennifer Bennett, technical director to Google Cloud’s CTO, said adding Google’s cloud services to Glass could help make it a revolutionary tool for workers in situations where a laptop or smartphone would be too awkward.

“Many of you probably remember Google Glass from the consumer days—it’s baaack,” Bennett said, earning warm laughter, before introducing Plataine’s project. “Glass has become a really interesting technology for the enterprise.”

The session came roughly one year after Google abandoned its attempt to sell consumers on Glass and its eye-level camera and display, which proved controversial due to privacy concerns. Instead, Google relaunched the gadget as a tool for businesses called Google Glass Enterprise Edition. Pilot projects have involved Boeing workers using Glass on helicopter production lines, and doctors wearing it in the examining room.

Anat Karni, product lead at Plataine, slid on a black version of Glass Tuesday to demonstrate the app. She showed how the app could tell a worker clocking in for the day about production issues that require urgent attention, and show useful information for resolving problems via Glass’s display.

A worker can also talk to Plataine’s app to get help. Karni showed how a worker walking into a store room could say “Help me select materials.” The app would respond, orally and on the devices display, with which materials are needed and where they could be found. A worker’s actions could be instantly visible to factory bosses, synced into the software Plataine already provides customers such as Airbus to track production operations.

Plataine built its app by plugging Google’s voice-interface service, Dialogflow, into a chatbot-like assistant it had already built. It got support from Google, and also software contractor and Google partner Nagarro. Karni credits Google’s technology with an impressive ability to understand how variations in phrasing, or use of terms such as “yesterday” that can trip up chatbots, translates into a worker’s tasks and needs. “It’s so natural,” she says.

Karni told WIRED that her team is now working with Google Cloud’s AutoML service to add image-recognition capabilities to the app, so it can read barcodes and recognize tools, for example. AutoML, which emerged from Google’s AI research lab, automates some of the work of training a machine learning model, and has become a flagship of Google’s cloud strategy. The company hopes corporate cloud services will become a major source of revenue, arguing that Google’s expertise in machine learning and computing infrastructure can help other businesses. Diane Greene, the division’s leader, said last summer that she hoped to catch up with Amazon, far and away the market leader, by 2022.

Gillian Hayes, a professor who works on human-computer interaction at University of California Irvine, says the Plataine project and plugging Google’s AI services into Glass play to the strengths of the controversial hardware. She previously tested the consumer version of the app as a way to help autistic people navigate social situations.

“Spaces like manufacturing floors, where there’s no social norm saying it’s not OK to use this, are the spaces I think it will do really well,” Hayes says. Improvements to voice interfaces and image recognition since Glass first appeared—and disappeared—could help give the device a second wind. “Image and voice recognition technology getting better will make wearable devices more functional,” Hayes says.

More Great WIRED Stories