How to Build an Automated Order Confirmation Bot with NextGenSwitch Programmable Voice API

admin

Jun 11, 2026

Order confirmation is one of the most practical and high-value use cases for voice automation in e-commerce. Many online businesses still depend on manual agents to call customers and confirm orders before delivery. This process is time-consuming, expensive, and difficult to scale during high order volume.

With the NextGenSwitch Programmable Voice API, businesses can automate order confirmation calls using simple XML call flows, webhooks, DTMF input, speech recognition, and backend integration.

In this guide, we will build two types of order confirmation bots:

Basic Order Confirmation Bot: Uses pre-recorded audio with <Play> and keypad input.
Advanced Order Confirmation Bot: Uses <Say> with speech recognition and DTMF input.

Why Automate Order Confirmation?

Manual order confirmation creates several challenges:

Agents must call customers one by one
Fake orders increase delivery costs
Orders are delayed while waiting for confirmation
Customer response tracking becomes difficult
Call center cost increases with order volume

An automated order confirmation bot helps businesses:

Reduce fake orders
Improve successful delivery rates
Save agent time
Confirm orders faster
Operate 24/7
Scale to thousands of calls per day
Automatically update order status in CRM or e-commerce systems

Common Use Cases

An order confirmation bot can be used for:

E-commerce order confirmation
Cash-on-delivery order verification
Delivery address confirmation
Appointment confirmation
Subscription confirmation
Payment reminder calls
Product delivery follow-up
Customer verification before dispatch

This is especially useful for businesses using platforms like:

Shopify
WooCommerce
OpenCart
Magento
Laravel-based e-commerce systems
Node.js applications
Custom ERP or CRM platforms

Solution Architecture

The flow of a typical automated order confirmation bot is structured as follows:

Customer Places Order
        │
        ▼
E-commerce Platform
        │
        ▼
Backend Application
        │
        ▼
NextGenSwitch Programmable Voice API
        │
        ▼
Customer Receives Call
        │
        ▼
Customer Responds
        │
        ▼
Webhook Receives Result
        │
        ▼
Order Status Updated

The backend application controls the business logic. NextGenSwitch handles the voice call, audio playback, input collection, and webhook callback.

Part 1: Basic Order Confirmation Bot

The basic version is the easiest way to start. It uses:

<Play> to play a pre-recorded audio file.
<Gather> to collect keypad input.
digits callback variable to identify the customer’s choice.

This version does not use Text-to-Speech (TTS) or speech recognition.

Basic Bot Call Flow

Call Customer
        │
        ▼
Play Pre-recorded Audio
        │
        ▼
Customer Presses 1 or 2
        │
        ▼
Webhook Receives digits
        │
        ▼
Update Order Status

Basic XML Example Using Play

<Response>
  <Gather action="https://yourdomain.com/order-confirm" method="POST" input="dtmf" numDigits="1" timeout="10">
    <Play>https://yourdomain.com/audio/order-confirmation.mp3</Play>
  </Gather>
  <Play>https://yourdomain.com/audio/no-response.mp3</Play>
</Response>

Basic Audio Script

You can record this message and save it to your server:

Main Prompt: https://yourdomain.com/audio/order-confirmation.mp3
- Script: “Hello. This is an order confirmation call from ABC Store. Your order total is 1500 taka. Press 1 to confirm your order. Press 2 to cancel your order.”
Fallback Prompt: https://yourdomain.com/audio/no-response.mp3
- Script: “We did not receive your response. Our support team will contact you shortly.”

Basic Callback Example

If the customer presses 1, NextGenSwitch sends the result to your action URL:

POST /order-confirm
call_id=CALL-12345
digits=1
event_from=8801700000000
event_to=8801800000000
event_call_id=CALL-12345

If the customer presses 2:

POST /order-confirm
call_id=CALL-12345
digits=2
event_from=8801700000000
event_to=8801800000000
event_call_id=CALL-12345

Basic PHP Example

<?php
$callId = $_POST['call_id'] ?? '';
$digits = $_POST['digits'] ?? '';
$from = $_POST['event_from'] ?? '';
$to = $_POST['event_to'] ?? '';
$status = 'FOLLOW_UP';

if ($digits === '1') {
    $status = 'CONFIRMED';
} elseif ($digits === '2') {
    $status = 'CANCELLED';
}

// Example: update your database
// updateOrderStatus($callId, $from, $to, $status);

header('Content-Type: text/xml');

if ($status === 'CONFIRMED') {
    echo '<Response>
  <Play>https://yourdomain.com/audio/order-confirmed.mp3</Play>
</Response>';
} elseif ($status === 'CANCELLED') {
    echo '<Response>
  <Play>https://yourdomain.com/audio/order-cancelled.mp3</Play>
</Response>';
} else {
    echo '<Response>
  <Play>https://yourdomain.com/audio/follow-up-required.mp3</Play>
</Response>';
}
?>

Basic Node.js Example

import express from 'express';
const app = express();

app.use(express.urlencoded({ extended: true }));

app.post('/order-confirm', async (req, res) => {
  const callId = req.body.call_id || '';
  const digits = req.body.digits || '';
  const from = req.body.event_from || '';
  const to = req.body.event_to || '';
  let status = 'FOLLOW_UP';

  if (digits === '1') {
    status = 'CONFIRMED';
  } else if (digits === '2') {
    status = 'CANCELLED';
  }

  // Example: update your database
  // await updateOrderStatus({ callId, from, to, status });

  res.type('text/xml');

  if (status === 'CONFIRMED') {
    return res.send(`
<Response>
  <Play>https://yourdomain.com/audio/order-confirmed.mp3</Play>
</Response>`);
  }
  if (status === 'CANCELLED') {
    return res.send(`
<Response>
  <Play>https://yourdomain.com/audio/order-cancelled.mp3</Play>
</Response>`);
  }
  return res.send(`
<Response>
  <Play>https://yourdomain.com/audio/follow-up-required.mp3</Play>
</Response>`);
});

app.listen(3000, () => {
  console.log('Order confirmation server running on port 3000');
});

Part 2: Advanced Order Confirmation Bot

The advanced version is more flexible and natural for the user. It uses:

<Say> for dynamic Text-to-Speech (TTS).
<Gather> for speech and keypad input simultaneously.
speech_result callback variable to read the customer’s spoken response.
digits as a DTMF fallback.
Backend logic to detect confirmation or cancellation.

This is highly useful when the order details (like customer name, product name, or total amount) are dynamic and specific to each customer.

Advanced Bot Call Flow

Call Customer
        │
        ▼
Say Dynamic Order Details
        │
        ▼
Customer Speaks or Presses Key
        │
        ▼
Webhook Receives speech_result or digits
        │
        ▼
Backend Detects Intent
        │
        ▼
Update Order Status

Advanced XML Example Using Say and Speech Gather

<Response>
  <Gather action="https://yourdomain.com/order-confirm" method="POST" input="speech dtmf" speechTimeout="auto" timeout="10" numDigits="1" transcript="true">
    <Say voice="female">
      Hello. This is an order confirmation call from ABC Store. You ordered one wireless headset. The total amount is 1500 taka. Press 1 or say confirm to confirm your order. Press 2 or say cancel to cancel your order.
    </Say>
  </Gather>
  <Say voice="female">
    We did not receive your response. Our support team will contact you later.
  </Say>
</Response>

Advanced Dynamic XML Example

In a real e-commerce integration, details like product name, price, and customer name are generated dynamically from your database:

<Response>
  <Gather action="https://yourdomain.com/order-confirm?order_id=1001" method="POST" input="speech dtmf" speechTimeout="auto" timeout="10" numDigits="1" transcript="true">
    <Say voice="female">
      Hello John. This is an order confirmation call from ABC Store. You ordered Samsung Wireless Earbuds. Your total amount is 2500 taka. Press 1 or say confirm to confirm your order. Press 2 or say cancel to cancel your order.
    </Say>
  </Gather>
  <Say voice="female">
    We did not receive your response. Our support team will contact you shortly.
  </Say>
</Response>

Gather Callback Variables

When the customer responds, NextGenSwitch sends the following callback parameters to your action URL:

Variable	Description
`call_id`	Current call ID
`digits`	DTMF keypad input
`speech_result`	Recognized speech text
`confidence`	Speech recognition confidence score (0.0 to 1.0)
`voice`	Captured voice/audio file path, if available
`rec_path`	Internal recording path
`event_from`	Caller phone number
`event_to`	Destination phone number
`event_call_id`	Original call ID

Callback Payload Examples

Spoken Response Callback

If the customer says: “Yes, please confirm my order.”

POST /order-confirm
call_id=CALL-12345
speech_result=yes please confirm my order
confidence=0.96
event_from=8801700000000
event_to=8801800000000
event_call_id=CALL-12345

Keypad Input (DTMF) Callback

If the customer presses 1:

POST /order-confirm
call_id=CALL-12345
digits=1
event_from=8801700000000
event_to=8801800000000
event_call_id=CALL-12345

Advanced PHP Example Using `speech_result`

<?php
$callId = $_POST['call_id'] ?? '';
$digits = $_POST['digits'] ?? '';
$speechResult = strtolower($_POST['speech_result'] ?? '');
$confidence = $_POST['confidence'] ?? '';
$from = $_POST['event_from'] ?? '';
$to = $_POST['event_to'] ?? '';
$status = 'FOLLOW_UP';

if ($digits === '1') {
    $status = 'CONFIRMED';
} elseif ($digits === '2') {
    $status = 'CANCELLED';
} elseif (
    str_contains($speechResult, 'confirm') ||
    str_contains($speechResult, 'yes') ||
    str_contains($speechResult, 'okay')
) {
    $status = 'CONFIRMED';
} elseif (
    str_contains($speechResult, 'cancel') ||
    str_contains($speechResult, 'no')
) {
    $status = 'CANCELLED';
}

// Example: save callback data
// updateOrderStatus([
//     'call_id' => $callId,
//     'from' => $from,
//     'to' => $to,
//     'status' => $status,
//     'speech_result' => $speechResult,
//     'confidence' => $confidence
// ]);

header('Content-Type: text/xml');

if ($status === 'CONFIRMED') {
    echo '<Response>
  <Say voice="female">Thank you. Your order has been confirmed.</Say>
</Response>';
} elseif ($status === 'CANCELLED') {
    echo '<Response>
  <Say voice="female">Your cancellation request has been recorded.</Say>
</Response>';
} else {
    echo '<Response>
  <Say voice="female">We could not determine your response. Our support team will contact you shortly.</Say>
</Response>';
}
?>

Advanced Node.js Example Using `speech_result`

import express from 'express';
const app = express();

app.use(express.urlencoded({ extended: true }));

app.post('/order-confirm', async (req, res) => {
  const callId = req.body.call_id || '';
  const digits = req.body.digits || '';
  const speechResult = (req.body.speech_result || '').toLowerCase();
  const confidence = req.body.confidence || '';
  const from = req.body.event_from || '';
  const to = req.body.event_to || '';
  let status = 'FOLLOW_UP';

  if (digits === '1') {
    status = 'CONFIRMED';
  } else if (digits === '2') {
    status = 'CANCELLED';
  } else if (
    speechResult.includes('confirm') ||
    speechResult.includes('yes') ||
    speechResult.includes('okay')
  ) {
    status = 'CONFIRMED';
  } else if (
    speechResult.includes('cancel') ||
    speechResult.includes('no')
  ) {
    status = 'CANCELLED';
  }

  console.log({
    callId,
    from,
    to,
    digits,
    speechResult,
    confidence,
    status
  });

  // Example: update your database
  // await updateOrderStatus({
  //   callId,
  //   from,
  //   to,
  //   status,
  //   speechResult,
  //   confidence
  // });

  res.type('text/xml');

  if (status === 'CONFIRMED') {
    return res.send(`
<Response>
  <Say voice="female">Thank you. Your order has been confirmed.</Say>
</Response>`);
  }
  if (status === 'CANCELLED') {
    return res.send(`
<Response>
  <Say voice="female">Your cancellation request has been recorded.</Say>
</Response>`);
  }
  return res.send(`
<Response>
  <Say voice="female">We could not determine your response. Our support team will contact you shortly.</Say>
</Response>`);
});

app.listen(3000, () => {
  console.log('Advanced order confirmation server running on port 3000');
});

How to Use `speech_result`

The speech_result field contains the raw transcript of customer voice input. You should check for positive and negative intents inside your webhook logic.

Example Intent Mappings

Customer Says	`speech_result`	Suggested Status
”Yes”	`yes`	Confirmed
”Confirm”	`confirm`	Confirmed
”Please confirm my order”	`please confirm my order`	Confirmed
”Okay”	`okay`	Confirmed
”Cancel”	`cancel`	Cancelled
”No”	`no`	Cancelled
”I want to change my address”	`i want to change my address`	Follow Up
”Call me later”	`call me later`	Follow Up

Using AI for Better Intent Detection

Simple keyword matching works well for basic scenarios, but real-world voice responses can be complex. Customers might say:

“Can you deliver it tomorrow instead?”
“I ordered this by mistake.”
“Please confirm it, but change my delivery address.”

In these cases, you can forward speech_result to a Large Language Model (LLM) or Natural Language Processing (NLP) classifier to categorize customer intent.

For example, an AI analysis of the input might return:

{
  "intent": "address_change",
  "status": "FOLLOW_UP",
  "summary": "Customer wants to confirm the order but change the delivery address."
}

By identifying categories like confirm, cancel, address_change, delivery_question, payment_question, call_later, or human_agent_required, you can make your voice automation flows much more robust and customer-friendly.

Basic vs Advanced Bot Comparison

Feature	Basic Bot	Advanced Bot
Audio Method	`<Play>`	`<Say>`
Input Method	DTMF only	Speech + DTMF
Dynamic Order Details	Limited	Yes
Uses digits	Yes	Yes
Uses speech_result	No	Yes
Best For	Simple confirmation	AI-ready natural confirmation
Deployment Complexity	Low	Medium
Customer Experience	Standard keypad interaction	Conversational & more natural

Best Practices

To build a reliable order confirmation bot:

Keep confirmation messages short and to the point.
Mention the business name clearly.
Instruct customers exactly what to press or say.
Always include a DTMF fallback for speech-to-text workflows.
Log every incoming webhook payload for troubleshooting.
Save the call_id with the corresponding order record.
Store speech_result and confidence scores for auditing.
Use fallback messages inside <Gather> to handle no-responses.
Send unclear or complex responses to human follow-up queues.
Avoid using unsupported <Gather> attributes.

Example Database Schema

You may want to store these fields in your database for tracking voice confirmation status:

Field	Example Value
`order_id`	`1001`
`customer_phone`	`8801700000000`
`call_id`	`CALL-12345`
`digits`	`1`
`speech_result`	`yes please confirm`
`confidence`	`0.96`
`status`	`CONFIRMED`
`created_at`	`2026-06-11 10:30:00`

Final Workflow Example

Here is how the complete business flow looks once integrated:

New Order Created
        │
        ▼
Generate Call XML
        │
        ▼
NextGenSwitch Calls Customer
        │
        ▼
Customer Presses Key or Speaks
        │
        ▼
Webhook Receives digits or speech_result
        │
        ▼
Backend Updates Order Status
        │
        ▼
Warehouse Receives Confirmed Order

Conclusion

NextGenSwitch Programmable Voice API makes it easy to build automated order confirmation systems using simple XML call flows and standard backend applications.

The Basic Order Confirmation Bot is perfect for businesses that want a simple and reliable solution using pre-recorded audio and keypad input.
The Advanced Order Confirmation Bot is better for businesses that need dynamic order details, speech recognition, and AI-powered response handling.

By combining <Play>, <Say>, <Gather>, digits, and speech_result, you can automate order verification, reduce fake orders, improve delivery success, and scale customer communication efficiently.

For further reference, check out the NextGenSwitch Programmable Voice API documentation to customize these call flows to suit your specific business logic.

Return to blog

Continue Exploring NextGenSwitch

Go deeper with implementation and platform pages:

Programmable Voice API Docs Virtual PBX Product Contact Center Solution

How to Build an Automated Order Confirmation Bot with NextGenSwitch Programmable Voice API

admin

Why Automate Order Confirmation?

Common Use Cases

Solution Architecture

Part 1: Basic Order Confirmation Bot

Basic Bot Call Flow

Basic XML Example Using Play

Basic Audio Script

Basic Callback Example

Basic PHP Example

Basic Node.js Example

Part 2: Advanced Order Confirmation Bot

Advanced Bot Call Flow

Advanced XML Example Using Say and Speech Gather

Advanced Dynamic XML Example

Gather Callback Variables

Callback Payload Examples

Spoken Response Callback

Keypad Input (DTMF) Callback

Advanced PHP Example Using `speech_result`

Advanced Node.js Example Using `speech_result`

How to Use `speech_result`

Example Intent Mappings

Using AI for Better Intent Detection

Basic vs Advanced Bot Comparison

Best Practices

Example Database Schema

Final Workflow Example

Conclusion

Continue Exploring NextGenSwitch

Talk to our AI voice agent

Continue the conversation on WhatsApp

How to Build an Automated Order Confirmation Bot with NextGenSwitch Programmable Voice API

admin

Why Automate Order Confirmation?

Common Use Cases

Solution Architecture

Part 1: Basic Order Confirmation Bot

Basic Bot Call Flow

Basic XML Example Using Play

Basic Audio Script

Basic Callback Example

Basic PHP Example

Basic Node.js Example

Part 2: Advanced Order Confirmation Bot

Advanced Bot Call Flow

Advanced XML Example Using Say and Speech Gather

Advanced Dynamic XML Example

Gather Callback Variables

Callback Payload Examples

Spoken Response Callback

Keypad Input (DTMF) Callback

Advanced PHP Example Using speech_result

Advanced Node.js Example Using speech_result

How to Use speech_result

Example Intent Mappings

Using AI for Better Intent Detection

Basic vs Advanced Bot Comparison

Best Practices

Example Database Schema

Final Workflow Example

Conclusion

Continue Exploring NextGenSwitch

Advanced PHP Example Using `speech_result`

Advanced Node.js Example Using `speech_result`

How to Use `speech_result`