AMD timeout, call routed before detection completed
Answering machine detection didn't finish in the allowed window, call routed as human by default.
Symptoms
- •Voicemails reaching agents despite AMD being enabled
- •Campaign logs show AMD timeout messages
- •False-positive rate on human calls unusually low, but live-human-to-voicemail-drop ratio is off
Common causes
- •AMD max-time too short for typical greeting length
- •Audio path slow, RTP jitter delaying early-packet classification
- •AMD algorithm struggling with specific greeting patterns (common in certain demographic groups)
How to fix it
- 1Increase AMD total_analysis_time in campaign: from 5000 to 6000-7000ms
- 2Review RTP jitter stats in `rtp set debug on` output
- 3A/B test AMD tuning parameters on a small campaign first
VICIPanel's AMD tuning UI shows timeout rate alongside false-positive and false-negative rates so you can optimize all three together.
Related issues
More Asterisk errors
Asterisk not running on the dialer box
The Asterisk process is dead, no calls can be placed or received until it's restarted.
Campaign dial level stuck at zero
Campaign auto-dial level drops to 0 and never recovers, dialing stops.
Dial prefix invalid for campaign
Campaign is configured with a dial prefix that doesn't match any Asterisk dialplan extension, every dial fails instantly.
Hopper empty, dialing paused
The pre-dial lead cache is empty and the fill cron hasn't run, dialing stops until new leads enter the hopper.
Stop troubleshooting. Start operating.
VICIPanel's health dashboard catches most of these issues before they impact your agents — with plain-English diagnostics and one-click remediation.