Answers to common questions about how ResolveGrid works, how it's deployed, and what customers see.
ResolveGrid helps field technicians fix complex equipment the first time they show up. The AI walks them through troubleshooting step by step, using your own approved manuals and workflows. Computer vision lets the AI see what the technician sees and check the work. And if the technician gets stuck, one click brings in a live remote expert over augmented reality video.
We're headquartered in Mountain View, California. ResolveGrid was co-founded with Andrew Ng's AI Fund and builds on the CareAR technology originally developed at Xerox. We focus on industries where repeat visits are expensive and senior expertise is hard to come by, including complex electronics, telecommunications, medical devices, transportation, and manufacturing.
With a remote video tool, getting help depends on getting the right person on a call. ResolveGrid works differently. The AI guides the technician first, using answers drawn from your own documentation, and a human expert joins only if the issue calls for one.
That changes a few things in practice. The AI is available whether or not an expert is free. Every technician in every market gets the same guidance. And when an expert does join, they see the full workflow, the photos and video, and everything the AI has already tried, so nobody starts over. Our AR remote assistance product, ResolveGrid Assist, is one layer of the platform.
CareAR started inside Xerox as augmented reality remote support technology, and it was proven across one of the world's largest service organizations. ResolveGrid was launched as an independent company, co-founded with AI Fund, to bring that technology to market and build on it with agentic AI, computer vision, and training.
Today CareAR is part of ResolveGrid. When we talk about Assist, we mean the CareAR Assist remote support capability, with live video, AR annotations, and expert guidance. Existing CareAR customers are still fully supported, and the CareAR help center at help.carear.com is still the place to go for Assist documentation.
ResolveGrid isn't tied to any one type of equipment. Administrators set up each equipment model as an asset and attach its manuals, service guides, videos, and workflows. From then on, the AI guides technicians using the material for that specific model.
Customers use it across a wide range of hardware. That includes industrial robots, laboratory analyzers, medical imaging systems, HVAC and building systems, telecom and network equipment, printers and multifunction devices, ATMs and cash recyclers, self-service kiosks, and drive-through and point-of-sale systems.
ResolveGrid doesn't rely on a general AI model's best guess about your equipment. Administrators upload your content (manuals, quick-start guides, service bulletins, SOPs, videos, and workflows), attach it to the right equipment model, and publish it. The AI Assistant and the Training agent are designed to answer only from that published content.
When a technician starts a session and names or scans the equipment, the AI locks onto that model and keeps every recommendation specific to it. Anything that's been uploaded but not published yet stays invisible to the AI, so administrators get a chance to review content before it ever reaches a technician.
Any AI built on a language model can make mistakes, so we designed ResolveGrid with several layers of protection. Answers are generated from your published documentation rather than the model's general training, which greatly reduces the chance of an invented procedure or part number. They cite the passages they draw from, so a technician can check the original. And when the AI can't find a grounded answer, it's designed to say so rather than guess.
On top of that, workflows can require visual verification of critical steps, and a live remote expert is always one click away. That gives technicians a human to turn to for safety-critical decisions.
The ResolveGrid AI Assistant works in English, French, German, Italian, Portuguese, Spanish, Indonesian, Japanese, Malay, and Simplified Chinese. It normally follows the user's portal language setting, so a technician in Brazil and a technician in Japan can work from the same knowledge base, each in their own language.
Yes. During an AI-guided session, the technician clicks one button to request a live expert. Available experts get notified, and when one accepts, the technician gets a prompt to join the live session.
The expert doesn't start cold. Everything carries over, including the workflow, the AI conversation, and any photos or video the technician already captured. From there, the expert can use ResolveGrid Assist's AR annotations to point at specific parts on the technician's live camera view.
ResolveGrid comes in three deployment models, all on the same platform.
Managed cloud is fully hosted by ResolveGrid, and you choose which AI provider powers the language models.
On-premises runs in your own data center or private cloud, using local LLMs on your own GPUs or approved AI providers through your own gateway, so your IT team controls the infrastructure and the data.
Air-gapped is completely isolated from outside networks and uses local LLMs. It's built for defense, critical infrastructure, and other highly regulated environments.
Workflows, computer vision, and expert escalation work the same in every model. What changes is where the platform runs and which AI models power it. Cloud deployments can use leading commercial models, while on-premises and air-gapped deployments run the strongest open-weight models your hardware supports.
Yes. In an air-gapped deployment, the whole platform runs inside an isolated network with no connection to the public internet, and that includes the language models. Prompts, documents, and session data never leave the environment.
Live expert escalation still works in an air-gapped deployment. The expert joins from inside the isolated network, so the session never leaves the environment. That opens up AI guidance to places that have usually been shut out of it, such as defense facilities, secure manufacturing sites, and critical infrastructure operators.
ResolveGrid is designed to sit on top of the systems you already run. It connects with field service management (FSM), CRM, ticketing, asset management, contact center, and knowledge platforms, so job details flow in and resolution data flows back out.
The AI Assistant can also be embedded in other applications. A technician working in an existing portal or support tool can use ResolveGrid without switching apps.
An administrator creates a training task for one equipment model, describes what it should cover and what counts as passing, and pins the manuals, videos, and workflows to study. The AI writes the certification quiz from those success criteria, so nobody has to write questions by hand.
The technician opens My Training and clicks Start. The AI training agent walks them through the materials, answers their questions, and then gives the quiz, which unlocks only after they've gone through the study materials. Passing marks the task complete with a score and date. Administrators can see who passed, who's overdue, and overall pass rates on a dashboard, with automated reminders and CSV exports for audits.
Yes. Administrators or training coordinators can set up user groups, such as all technicians in a region or all new hires, and assign a task or a sequence of tasks to the whole group with a due date. Anyone added to the group later is enrolled automatically.
Reminder emails go out on a schedule. They nudge people before the due date, repeat when a task is overdue, respect quiet hours, and can escalate to a manager if needed. An administrator can also send an immediate reminder from the dashboard to anyone who's falling behind.
Here's what early ResolveGrid customers have reported.
A 50% reduction in dispatches, as AI agents triage and resolve issues remotely before a truck ever rolls.
An 82% first-time fix rate, with guided troubleshooting and visual verification cutting down on failed first visits.
A 63% reduction in service time, from faster diagnosis onsite and fewer repeat visits.
An 85% increase in remote resolution rates, from combining AI self-solve with live expert help.
Results vary depending on the equipment and where a team is starting from.
Request a demo at resolvegrid.ai/demo, and our team will walk you through the platform using equipment that's relevant to your industry.
Most customers start small, with one piece of equipment or product line where repeat dispatches or expert bottlenecks are costing the most. Getting that equipment onboarded means uploading the documentation and videos you already have and publishing a knowledge base, which takes hours, not weeks. From there, teams add more equipment, set up training, and roll out to more technicians, partners, or customers.
Browse our field service AI glossary for plain-language definitions, or contact our team.