FAS turns call completion into a margin question

For outbound teams, a call marked “answered” is not always a conversation. False Answer Supervision (FAS) occurs when a carrier returns an answer signal even though the called party has not answered. The dialer may treat that signal as a connected call, while the call detail record (CDR) and the carrier bill record chargeable time.

That makes FAS a margin problem as well as a signaling problem. At volume, a short false connection repeated across many attempts can add trunk expense, occupy capacity and distort campaign reporting. But the bill is not calculated the same way on every route. Rates, minimums, rounding increments and contract terms matter. There is no single per-call cost that applies across providers.

Trace the billable interval

In a SIP call, the answer signal—often a 200 OK—can tell the dialer that a call connected. If that signal originates inside the network rather than at the called party, the connection may be false. Billing may begin at answer and continue until release, subject to the carrier’s charging rules. A call that never reaches a person can therefore create a charge without creating a useful conversation.

Start the investigation with a matched set of records: the dialer’s event timeline, SIP signaling, audio or media evidence where available, and the carrier CDR. Compare the answer and release times, duration, route and call outcome. Check whether the carrier’s billed duration matches the contract’s charging increments. A mismatch is a reason to investigate, not proof of fraud by itself. FAS can arise from signaling or network behavior; a billing dispute needs evidence tied to the affected calls.

Keep the calculation tied to actual trunk terms. For a rough exposure estimate, group suspect calls by route and billing interval, then apply the relevant rate and minimum charge. Include repeat attempts and the capacity they occupied, but do not count those as direct charges unless the contract or operating model supports that calculation. This produces a defensible range rather than a headline number built on assumed rates.

Do not confuse FAS with an AMD error

FAS and answering machine detection failures can both leave a team paying for calls that deliver no useful conversation, but they are different events. With FAS, the network may signal an answer where no called party answered. With an AMD false positive, a real person answers and the dialer classifies that response as a machine. The latter can waste a paid connection and lose a lead; the former can create a billable interval without a human answer at all.

Separate the two in reporting. Track suspected false answers by carrier route and compare those records with audio or other call evidence. Track AMD outcomes against the actual call result, including cases where a live person was dropped. A single “bad answered call” bucket hides whether the remedy belongs in carrier escalation, dialer logic or detection tuning.

That distinction also matters when deciding what to change. Do not quarantine a trunk based on one anomalous call. Look for a repeatable pattern, compare routes, and preserve the records needed to challenge billing. A previous guide to multi-carrier FAS detection and trunk isolation makes the same operational point: treat a suspect signal as evidence to investigate, not a verdict on its own.

What builders should measure

For each route, watch the gap between answer signaling and the first credible indication of a person or machine. Record the disposition, release timing, media presence and billed duration where those data are available. Review changes over time rather than relying on a single aggregate rate. If the CDR says connected but there is no corresponding answer in the media, that discrepancy is a useful investigation lead.

Billing timers deserve special attention. A false answer held open for longer can cost more than one released quickly, but the exact effect depends on the provider’s billing rules. Check whether the dialer, carrier and reporting system use comparable start and end points. Differences in timestamps or rounding can make a clean comparison difficult, so document the assumptions behind any estimate.

Product watch

The product information available for this digest describes AMDY.IO’s existing detection service, not a newly announced release or pricing change. AMDY.IO says it detects answering machines, live human responses and carrier false answers, and starts detection in one-eighth of a second. It lists native integrations with ViciDial, Asterisk and FreeSWITCH, plus a WebSocket API; compatibility information also names GoAutoDial, FreePBX, Issabel and 3CX. The company offers 50,000 free detections upon sign-up. These details do not establish a carrier billing rate or show that FAS charges will be avoided in every case.

The practical takeaway for dialer builders is to keep classification and billing evidence connected but distinct. Fast detection can help a system act on an answered call sooner; it cannot, by itself, settle whether a carrier’s answer signal was valid or whether a charge follows the trunk agreement. Measure the signal, inspect the call, and reconcile the record before turning an apparent FAS pattern into a cost claim.