AI voice agents: inbound, outbound and conversational IVR
Voice agents that answer and make calls on PSTN, SIP, WebRTC and WhatsApp Voice, with sub-second turn-taking, barge-in and 30+ languages. When the call needs a person, the agent transfers it with the transcript. Runs on our cloud, or the full stack (STT, LLM, TTS, telephony) inside your data centre.
Sub-second turn-taking, barge-in, and a payment link sent on WhatsApp before the call ends. A human takes over only if the customer objects.
Get a call from the agentAgent: Hi, this is Maya calling from your insurer about your car policy, which ends on the 4th. Is now a good time?
Customer: Yeah, go ahead. Is the premium going up?
Agent: Slightly, it's 12,400 this year, up from 11,900, because of the rate revision. Your no-claim bonus is intact. Would you like to renew now? I can send the payment link on WhatsApp.
Customer: Hmm, can I get a discount?
Agent: I can't change the premium, but a colleague can look at options. I'll have them call you back before 6 today. Should I send the link anyway so you have it?
Customer: Sure, send it.
Agent: Done, it's on your WhatsApp. Thanks for your time.
An EnableX voice agent is an AI agent that holds a spoken conversation with a customer over the phone, WhatsApp Voice or in-app audio. It listens, understands, replies with sub-second turn-taking, takes actions in your systems, and transfers to a person with the transcript when needed. It runs in 30+ languages on EnableX telephony in 50+ countries, on our cloud or yours.
The quickest way to judge it is to hear it, then put it on a call type you already make. Book a 20-minute demo and we set the agent up on one of your real call flows: reminders, renewals, confirmations or inbound queries. It runs on your numbers, in your customers' language, with your team listening in. You read the transcripts and decide.
Same agent, same Studio, same telephony. Pick the direction the call goes in.
Callers hang up on agents that pause or talk over them. Four things stop that.
The agent detects the caller's language from the first sentence and continues in it. It handles mixed-language speech such as Hindi-English and Arabic-English without a language menu.
Multilingual voice AI →Picked up from the first sentence. No language menu.
Hindi-English and Arabic-English handled as one conversation.
Natural voices per language, or bring your own TTS voice.
Warm transfer to a live agent on the same call, or escalation to chat or a video call on the same platform. The agent's screen shows the transcript, intent and sentiment before they say hello, so the customer never repeats themselves. You set the triggers.
How handover works →No-code in Studio for CX and operations teams. One API request for developers.
Pick a template, connect your knowledge base, choose a language and a number, then call it from your own phone and listen.
Dialogs Studio →One request on a connected call opens a real-time, bidirectional audio stream to any STT, LLM and TTS, hosted or your own.
curl -X PUT https://api.enablex.io/voice/v1/call/$voice_id/stream \
-u APP_ID:APP_KEY \
-H "Content-Type: application/json" \
-d '{ "wss_host": "wss://your-host:port" }'const axios = require('axios');
const voiceId = 'YOUR_VOICE_ID';
const res = await axios.put(
`https://api.enablex.io/voice/v1/call/${voiceId}/stream`,
{ wss_host: 'wss://your-host:port' },
{ auth: { username: 'APP_ID', password: 'APP_KEY' } }
);
console.log(res.data);import requests
from requests.auth import HTTPBasicAuth
voice_id = "YOUR_VOICE_ID"
res = requests.put(
f"https://api.enablex.io/voice/v1/call/{voice_id}/stream",
auth=HTTPBasicAuth("APP_ID", "APP_KEY"),
json={"wss_host": "wss://your-host:port"}
)
print(res.json())Start on EnableX cloud with per-minute pricing. Move to private cloud, hybrid or fully on-premise without rebuilding: same product, same APIs. Air-gapped option for government. Controls aligned with ISO 27001 and SOC 2 Type II.
On-premise voice AI →Fastest start. Data in your chosen region.
Single-tenant in your own cloud account.
Call media in your data centre, orchestration on EnableX.
STT, LLM, TTS and telephony inside your network. Air-gapped option.
Routine calls are most of the volume and none of the judgement. The agent takes the routine, your people take the exceptions.
Three anonymised deployments. Names are withheld at our customers' request.
Build it: developer docs
Book a 20-minute demo and we will run the agent on one of your own call types, or start a free trial. No credit card required.
Related guides: On-Premise CPaaS · Voice AI compliance in India