Automate·Intermediate·25 min·Updated Sep 30, 2026

Monitor device health across the fleet

Analyze device telemetry to flag at-risk hardware before it fails, instead of reacting after a user reports a problem.

Download PDF

Good fit for

Microsoft 365

Excel

Teams

Works With

Claude

Copilot

Prerequisites

A device telemetry export (battery health, disk health, error/crash frequency) from your device management system

Business Outcome

Fewer unplanned device failures and the downtime they cause, by catching warning signs early.

Workflow Overview

The apps run in this order. At the AI step, pick whichever of these you have — they're alternatives, not a sequence.

Excel

Teams

Claude

Copilot

Copilot works differently from the others: with a paid Microsoft 365 Copilot license it's built right into the app above, no copy-pasting needed. Without a license, use it the same way as ChatGPT or Claude — in its own chat window.

Step 1: Define what "at risk" means from your telemetryExcel

Look back at devices that actually failed in the past year and check what their telemetry showed beforehand, then turn that into explicit thresholds.

Battery health below 60%, more than two disk errors a month, three or more unexpected shutdowns in 30 days — rather than flagging on a vague sense of "old device," which doesn't actually correlate with failure as reliably as people assume.

Step 2: Analyze the fleet exportCopilotClaude

Load the telemetry export into Excel and ask Copilot to flag devices matching the defined at-risk signals, stating the specific metric and value for each flag.

Prompt idea:

Using this device telemetry export, flag any device with battery health below 60%, disk errors logged more than twice this month, or 3+ unexpected shutdowns in the past 30 days. State which signal triggered each flag and the actual value.

Step 3: Prioritize by business impact, not just severityExcel

Rank flagged devices considering both how severe the signal is and who's using the device — a failing device for someone in a customer-facing role may need faster action than the same signal on a rarely used device.

Step 4: Route to IT with enough lead time to actTeams

Share the prioritized list with IT while there's still time to proactively replace or repair the device, rather than after it's already failed and someone's blocked.

Check the work

Spot-check flagged devices' actual telemetry against the source export to confirm the numbers are accurate.
Confirm the priority ranking reflects real business impact, not just raw signal severity.
Track how many flagged devices actually failed vs. false-positived, and tune thresholds accordingly over time.

Source: Microsoft Copilot Scenario Library — IT (2026)

Expected Outcome

A ranked list of at-risk devices with the specific telemetry signal driving the flag, routed to IT for proactive replacement or repair.

Is AI actually needed here?

Step 1 spells out exact thresholds — battery below 60%, disk errors, shutdown counts. Once the rule is explicit, step 2's AI flagging is just evaluating it row by row, which Excel already does natively.

Non-AI alternative: A Power Query transformation or nested Excel formula applies the same thresholds across the whole fleet export without AI, and refreshes automatically as new telemetry lands.