Temperature follows the shift.
Inspect the comparison behind the hypothesis. A relationship suggests a check; it does not prove the cause.
AI INVESTIGATIONS
Find out why
your tests fail.
AI compares your test results, investigates possible causes, and shows the evidence so your team can decide what to check next.
FOLLOW THE EVIDENCE
A voltage shift. Six possible directions.
Choose the scope, inspect a hypothesis, open its evidence, and decide which test could confirm it.
Compare the affected runs with the baseline.
The production investigation inherits the alert’s scope, time span, and drifting measurement. Explore subsets of this example to see why the comparison matters.
BASELINE · 250 RUNS3.3072 V
INCIDENT · 100 RUNS3.3685 V
MEAN SHIFT+61.3 mV
Temperature also changed: 24.2°C → 34.4°C. That is a hypothesis to examine, not yet a cause.
EVIDENCE + UNCERTAINTY
A probable cause is a testable direction.
Start from a yield drop or a drifting measurement. Compare stations, lots, versions and other recorded context; keep missing evidence visible.
The component lot is missing.
Add lot data and an independent reference reading to separate competing explanations.
HAND OFF THE NEXT CHECK
Give the team a finding they can follow.
Share the result with its alert context. Separate the confirming hardware test from the extra data the analysis needs.
Analysis outcomes and team access
A result can identify probable or interacting causes, remain inconclusive, or report insufficient data. Missing metadata stays visible.
Each analysis respects the initiating person’s role and access group. Teammates can read the shared result and launch their own analysis to inspect detailed evidence. Re-running replaces the latest analysis.
