Warn them before they answer.
Tag suspected spam calls so your subscribers see a warning, using the same AI risk score the voice firewall gives every call.
A warning, not a surprise.
Suspected spam is tagged, so the subscriber sees a warning on the screen instead of an anonymous number.
Illustrative
Every call gets a risk score.
AI scores each call setup using the origin, the route, the CLI and the calling pattern history. Policy then connects, rejects or flags it in real time.
Low risk, the call rings normally
Illustrative · pick a risk level to try it
Spotted by SMS, labelled on voice.
When the SMS engine sees a scam number, the voice engine acts on calls from it a second later.
Illustrative
Mirror, score, decide, learn.
Calls are read from passive taps, and each outcome makes the next score better.
Passive taps copy SIP and SS7 signalling without touching routing.
AI scores the call setup against route and number reputation.
Policy connects, rejects or flags the call in real time.
Outcomes feed back into the models during daily retraining.
Illustrative
What the firewall looks at.
The label rests on checks the voice firewall already runs on every call.
SIP inspection
Every SIP INVITE checked for header anomalies and suspicious origins.
SS7 call control
ISUP and MAP signalling evaluated for illegitimate call setups.
CLI validation
Calling line identity checked against the route it actually arrived on.
Trunk monitoring
Per-trunk baselines reveal sudden shifts in call volume or destinations.
per node
Carrier-grade throughput
Processes 2,500+ voice calls per second on each node.
Unified policy
Voice rules live in the same policy engine as SMS rules.
Illustrative
Every verdict is logged, and reversible.
Each decision is recorded with its risk score. If a label is wrong, your team can reverse it.
Illustrative
Questions, answered.
It tags suspected spam calls so that subscribers see a warning before they answer. The label is applied when the call is scored, using the same risk decision the voice firewall makes for every call.
AI assigns a risk score to each call setup using the origin, the route, the calling line identity (CLI) and the calling pattern history. A policy then connects, rejects or flags the call in real time.
Yes. The voice firewall reads SIP and SS7 signalling through the same policy engine, so mixed cores are covered.
No. Passive taps copy SIP and SS7 signalling without touching routing, so legitimate traffic is not affected.
Yes. Every decision is logged with its risk score, and every verdict can be reversed by your team.
Yes. When the SMS engine sees a scam number, the voice engine acts on calls from it a second later, because both share one telemetry bus.
Outcomes feed back into the models during daily retraining, so verdicts and subscriber reports sharpen later decisions.
Yes. Operators and partners can offer the firewall as their own service, with full control over policy, routing and data.
Let subscribers see the warning.
Talk to partnerships about labelling suspected spam calls on your network.