A technician walks up to a faulty measurement device, scans the QR code on its display, and gets a plain-language diagnosis with step-by-step repair instructions, all before picking up the phone. David Lincoln, Digital Lead of ABB’s Measurement and Analytics Division, spoke with Lucian Fogoros of IIoT World at Hannover Messe 2026 about how the company’s MyMeasurement Assistant, now in version eight, resolves 80% of technical support issues through an AI copilot that reads device diagnostics and connects them to ABB’s library of instruction manuals, deliberately tuned to never guess.
How Does a QR Code Turn Into a Repair Instruction?
Every ABB measurement device displays a QR code that contains device type, serial number, hardware version, and diagnostic information. When a technician scans that code with the MyMeasurement Assistant, a web application, an Azure-based system called Gen AI Data Analyzer reads the diagnostic data and generates a health report about the equipment’s status.
For well-trained technicians, those health reports are already useful. ABB’s own service teams use them daily to troubleshoot devices. But for technicians who are less experienced with a specific piece of equipment, the diagnostic data can be difficult to interpret. The AI copilot reads the full health report and summarizes it in natural language: what is happening inside the device, what fault codes mean, and what the technician should do next.
The copilot pulls resolution steps directly from ABB’s library of instruction manuals. If a device throws an alarm number, the copilot finds the matching entry in the manual and tells the technician exactly how to resolve it. That is the prescriptive layer: detecting the fault and prescribing the exact repair steps from ABB’s own documentation.
What Happens When the AI Cannot Solve the Problem?
Internal testing at ABB shows that 80% of technical support issues are resolved through the MyMeasurement Assistant without contacting support or waiting for a specialist. The remaining 20% escalate automatically through a structured path that preserves all diagnostic context.
When the copilot cannot resolve an issue, it creates a case in ABB’s case management system. The transcript of the technician’s query and the full QR code diagnostic report are appended automatically, so the next person in the chain already understands the problem. The system then asks the technician to schedule a visual remote support call with a factory-trained ABB service specialist, someone who can look over their shoulder via video and guide the repair.
For the roughly 5% of cases where remote support still cannot resolve the issue, ABB sends a service technician to the site. That technician arrives already understanding the problem and carrying the correct spare parts, because the full diagnostic history traveled with the case. The result is faster resolution at every tier: self-service, remote, and on-site.
Why Did ABB Build a Copilot That Refuses to Guess?
ABB spent three years tuning the AI copilot on Microsoft AI Foundry, starting from a general-purpose model and narrowing it to respond only within the boundaries of ABB’s device documentation. The goal is a system with almost no risk of hallucination. “We don’t think that can serve customers. We won’t allow wrong answers. So we prefer no guesses,” according to the interview.
Microsoft AI Foundry provided the foundation: enterprise security, high-performance compute, and a base copilot that ABB then customized for its specific measurement devices. The partnership accelerated time to market, but the differentiation came from three years of tuning the model to ABB’s hardware, diagnostics, and documentation library.
Version eight of MyMeasurement Assistant also adds multi-language support. Technicians can now ask questions in their native language, which removes a significant barrier in global plants where English is a second language.
What Does Fleet-Wide Condition Monitoring Add?
ABB is also showcasing a new version of Gen AI Data Analyzer at Hannover Messe, expected to release after the summer, that extends the same AI architecture to fleet-level condition monitoring. A plant manager with hundreds of measurement devices spread across geographically distributed sites can now see the health status of the entire fleet in one view.
The system includes an anomaly detection algorithm that looks for patterns in telemetry data that might turn into failures in the coming days. All of this data feeds into the same copilot connected to ABB’s documentation library, so a plant manager can ask it to analyze live and historical telemetry, summarize trends, or generate a list of all active alarms across the fleet.
How University Research Shapes the Product Roadmap
ABB runs a long-standing partnership with Imperial College London’s chemical engineering department, including a summer research program where 20 to 30 students work in real plants. At a hackathon centered on ABB’s carbon capture plant, which has 18 measurement devices, students tested the new nameplate scanner feature in MyMeasurement Assistant. The scanner lets a technician photograph a device nameplate, and the application automatically generates a page with serial number and device details.
The students went further than testing. They developed concepts for virtual P&ID diagrams using copilot technology, building digital representations of the plant’s process flow. Some of those ideas are progressing into ABB’s product roadmap.
The broader observation from the partnership: industrial plants are often decades old and were never designed for digitalization. Connectivity is frequently absent. Students, unburdened by those constraints, tend to explore possibilities that experienced engineers have learned to work around rather than solve.
| Support Tier | What Happens | Issues Resolved | Context Preserved |
| Self-service (AI copilot) | Scan QR code, get AI diagnosis + prescribed fix from ABB manuals | 80% | Full diagnostic report |
| Remote specialist | Video call with factory-trained ABB technician | ~15% | Query transcript + QR report auto-appended |
| On-site service | ABB technician dispatched with correct spare parts | ~5% | Full case history, technician arrives prepared |
Frequently Asked Questions
1. What is prescriptive maintenance and how is it different from predictive?
Prescriptive maintenance goes beyond detecting that a fault exists or predicting that one will occur. It tells the technician exactly what to do about it, matching alarm codes to specific resolution steps from the equipment manufacturer’s documentation. ABB’s MyMeasurement Assistant uses an AI copilot connected to its instruction manual library to deliver those prescriptions automatically.
2. How does ABB prevent AI hallucination in industrial maintenance?
ABB spent three years tuning its copilot on Microsoft AI Foundry, restricting responses to information within ABB’s own device documentation and diagnostic data. The system is designed to produce no guesses. If it cannot resolve an issue, it escalates to a human specialist rather than generating an uncertain answer.
3. Can ABB’s MyMeasurement Assistant work across a fleet of devices?
Yes. A new version of Gen AI Data Analyzer, expected after summer 2026, provides fleet-wide condition monitoring across geographically distributed sites. It includes anomaly detection that identifies patterns likely to cause failures in the coming days, with the same AI copilot available to analyze telemetry data and summarize fleet health.
4. How does the QR code diagnostic system work?
Each ABB measurement device displays a dynamic QR code containing device type, serial number, hardware version, and diagnostic information. A technician scans the code with the MyMeasurement Assistant web application, which sends the data to an Azure-based Gen AI Data Analyzer that generates a health report and, through the AI copilot, a natural-language summary with prescribed repair steps.
—
This article is based on a video interview with David Lincoln, Digital Lead of ABB’s Measurement and Analytics Division, and Lucian Fogoros of IIoT World, recorded at Hannover Messe 2026. AI tools were used to help summarize and organize the content. Reviewed and edited by the IIoT World editorial team.
—
About the author
