CarSignalOpen a complaint issue and start a diagnosis. CarSignal works the problem the way a good diagnostic tech does: gather evidence, rank hypotheses, and confirm on the car before committing money to a repair.
What goes in
The vehicle's exact spec, the customer's reported symptoms, and any DTC codes. Codes are grounded against canonical SAE definitions so bank and sensor mappings come out right, and hybrid or EV code ranges force the correct powertrain assumptions. Recalls and complaint patterns for the exact year, make, and model are pulled in automatically, along with real-world failure patterns for that platform.
What comes back
- Ranked hypotheses, each with a working probability
- A recommended repair with a confidence percentage
- The next test to run, with what a pass or fail would prove
- An evidence chain: the "because" behind every claim
- Suggested parts with alternates and labor-hour ranges, ready to become an estimate
The session is conversational: log test results and observations, answer its follow-up questions, and watch confidence move. Repair Buddy sits alongside for OEM-grounded procedure and spec questions.
The trust gate
A diagnosis is not quote-ready until a confirming test has been run on this car, and quotes will not send below 70% confidence without an explicit override. The customer-facing report reuses the same evidence chain, so what the customer reads is what the shop verified.
Diagnostics are metered per shop as part of your plan's AI usage.