Glossary · Billing
What does concurrency mean for AI calls?
Concurrency (calls and streams)
Concurrency is how many calls, or speech streams, a system runs at the same moment. It's the number of lines open at once, and with call length it sets how many calls you can finish in an hour.
On a call
A lender uploads a list of borrowers for EMI reminders. With five lines open and calls of under a minute, the list moves steadily all afternoon. With one line, the same list would run late into the evening.
AI agents from ₹1 a minute.¹
Explained
Here is how it works, in plain words.
Concurrency is the voice world’s version of “how many counters are open”. Each live call holds one line for as long as it lasts. When every line is busy, the next call waits.
It decides how fast a campaign finishes. A rough rule: calls per hour equals lines open times the seconds in an hour, divided by the average seconds per attempt. Short calls and quick no-answers turn over faster. Long conversations hold a line for minutes.
Telephone engineers have measured this for a century. Traffic on a line is counted in erlangs, after the Danish engineer A. K. Erlang. One erlang is one line kept busy for the whole hour. The ITU’s traffic-engineering standard still uses the term.
Concurrency limits show up in several places at once. Your voice AI plan has one for calls. Your telephone carrier has one for channels. Each phone number has a sensible pace before networks start flagging it as spam. The slowest of these sets your real speed.
Text-to-speech has its own version. Each sentence being synthesised holds a slot while it’s made. A voice agent’s call needs a slot most of the time it’s speaking, so size the speech limit a little above your peak number of calls.
What happens at the limit matters as much as the number. Some systems reject the extra call with an error, and your code has to retry. Others queue it and dial when a line frees up. Queuing is friendlier, but a queued call rings later than you asked, so check the status of anything time-sensitive.
More concurrency isn’t always better. Calling hundreds of people a minute from a few numbers is a quick way to get them flagged. Match the lines to your numbers, and the numbers to a human pace.
In Voice Infra
This is what Voice Infra does about it.
Calls over your concurrency limit wait in a queue and dial as soon as a line frees up, so going over the limit doesn't drop a call. The limit is a setting on our side, and raising it needs no change to your code. Text-to-speech plans each include a number of requests at once, listed on our pricing page.