Subscribe to Our Newsletter

Success! Now Check Your Email

To complete Subscribe, click the confirmation link in your inbox. If it doesn’t arrive within 3 minutes, check your spam folder.

Ok, Thanks

Andrew Ng's latest bet targets a $400bn industry where AI might actually work, because the alternative is waiting for technicians who do not exist

ResolveGrid uses computer vision and agentic AI to guide field service repairs in real time, and its origin inside Xerox's own service operation gives it something most AI startups lack: proof that it works before launch day

Defused News Writer profile image
by Defused News Writer
Andrew Ng's latest bet targets a $400bn industry where AI might actually work, because the alternative is waiting for technicians who do not exist
Photo by Emmanuel Ikwuegbu / Unsplash

ResolveGrid launched on Tuesday with a proposition that sounds like standard AI startup fare, an agentic platform that guides workers through complex tasks, until you consider the specific industry it targets and the economics that make the timing compelling.

Field service, the work of sending technicians to install, maintain and repair equipment at customer sites, is a roughly $400 billion global market facing a structural labour crisis. The skilled technicians who know how to diagnose and fix complex medical devices, telecommunications infrastructure and industrial equipment are ageing out of the workforce faster than they can be replaced. Equipment is simultaneously becoming more complex, incorporating software, sensors and connectivity that older technicians were never trained on and newer recruits lack the experience to troubleshoot.

ResolveGrid's platform addresses the gap by combining three capabilities. First, it ingests a customer's existing documentation, repair manuals, training videos and historical work notes, and organises them into a knowledge base the AI can draw on. Second, it uses computer vision to identify the specific equipment a technician is looking at through their device camera, matching it to the relevant documentation and generating step-by-step repair guidance in real time. Third, when the AI reaches the limits of its judgment, it can escalate to a remote human expert over video, with the full workflow context already assembled so the expert can diagnose the problem without starting from scratch.

The company said early deployments have cut dispatches in half, meaning issues that previously required sending a technician to site are now resolved remotely, and reduced first-visit failure rates, the expensive and frustrating outcome where a technician arrives, cannot fix the problem, and must return with different parts or expertise.

What distinguishes ResolveGrid from the dozens of AI-for-field-service pitches circulating in enterprise software is provenance. The technology originated inside Xerox, which operates one of the largest field service organisations in the world, servicing millions of devices across thousands of customer sites. The platform was built and tested against real repair workflows at scale before being spun out as an independent company.

The spinout was co-founded with AI Fund, the venture studio led by Andrew Ng, the Stanford professor and former Google Brain and Baidu AI chief whose previous investments include Landing AI and DeepLearning.AI. Ng described the skilled technician shortage as a structural problem that AI is well-positioned to address, a notably restrained claim from a figure who tends towards precision in his public statements about AI capabilities.

ResolveGrid is available now to enterprise customers in electronics, telecommunications, medical devices, transportation and manufacturing.

The recap

  • ResolveGrid launches agentic AI platform for field service operations.
  • Early customers report dispatches are cut in half.
  • Platform is available now to enterprise customers via ResolveGrid.ai.
Defused News Writer profile image
by Defused News Writer

Explore stories