use case
verified July 2026
TTS for order-status calls.
An order-status call wants one fact: where the package is. The caller’s tolerance is measured in seconds, the answer is a sentence long, and the traffic arrives in seasons. Short calls at high volume are a latency problem first and a cost-shape problem second.
01
The latency problem
When the whole call is one lookup and one sentence, the synthesis leg is most of what the caller perceives. First audio here is 107 ms measured — 108 ms at p95, server time on the production API — which leaves the rest of the turn budget for the order lookup itself.
Audio — Order-status output, two markets — dated production audio
generated 2026-07-16
Spanish
“Su pedido llega mañana antes del mediodía. El número de seguimiento ya está en su correo, y si necesita cambiar la dirección, puedo hacerlo ahora mismo.”
Dane · 9.7 s · global support · live translation
English
“Driver twelve is eight minutes out with two pallets, and dock four is clear. I've dropped the manifest in your inbox so you can check it before they pull in.”
Dane · 7.0 s · logistics · dispatch
02
The seasonal cost shape
Order-status traffic follows the promo calendar, and the peak weeks that triple call volume are the same weeks margin is thinnest. A character meter makes the voice bill peak with the season; a flat line holds at $150 while overflow spills to burst lines at $10 a line-day, only on the days the spike is real.
Table 1 — One order-status line across a season
| Month shape | Minutes / mo | On a meter | Gandr flat |
|---|---|---|---|
| Baseline | 4,000 | $120 – $400 | $150 |
| Promo month | 8,000 | $240 – $800 | $150 |
| Peak season | 12,000 | $360 – $1,200 | $150 |
See also
Related sheets.
use case
20 ports
Replacing IVR menus with voice agents
Tearing out press-1 menus for a conversational agent changes the latency and concurrency math. What the synthesis layer must do, with measured figures.
use case
8,000 min
TTS for AI customer-support agents
The latency budget and cost arithmetic for a voice support desk: why first audio decides whether callers interrupt, and what 8,000 minutes a month costs.
capability
$10
Burst lines: capacity priced by the day
Traffic past your committed lines spills to burst lines at $10 a line-day, capped in your dashboard — capacity for spike days without paying for them all year.
lab report
5,000 min
Per-character pricing, run against flat lines
Four voice workloads run through a $30–$100 per-million-character meter and through flat lines, with the break-even stated and the cases where the meter wins.
Thirty minutes, your production script, the live latency readout — measured in front of you.
Bring this use case to the call