Build a Hotel Guest Services Line with Telnyx Voice and SMS

Hotels lose guest satisfaction when the front desk is busy, off-shift, or stuck on the phone. A concierge voice and SMS line gives guests one number to call or text for room service, housekeeping, concierge help, or maintenance, and gives staff a clean queue of structured requests to work through.

The canonical code example is in the Telnyx code examples repo:

https://github.com/team-telnyx/telnyx-code-examples/tree/main/hotel-guest-services-python

What This Example Builds

This Python Flask example combines Telnyx Voice, AI Inference, Messaging, and Slack into a single concierge workflow:

• Inbound voice calls are answered, greeted by name when the caller ID matches a known room, and asked for the room number when it does not
• Inbound SMS from any guest phone is accepted the same way
• Each request is classified into room_service, housekeeping, concierge, or maintenance with Telnyx AI Inference
• Urgent phrases (fire, flood, leak, locked out, gas, medical) override the LLM classification and route the request with priority
• A Slack alert is posted for every new request, with department-specific emoji
• The guest gets an SMS confirmation on receipt and another SMS when staff mark the request complete

Why This Is A Useful Voice AI Example

This example is small enough to read in one sitting, but it covers the moving parts a developer needs for a real voice AI workflow:

• Call Control webhooks for call.initiated, call.answered, call.speak.ended, call.gather.ended, call.hangup
• Caller-ID-based room lookup and a regex-based room-number capture when the caller is unknown
• AI Inference with a constrained JSON system prompt for classification
• Slack integration for staff visibility
• SMS confirmation with idempotent webhook handling

That makes it a good starting point for any single-vertical phone service where one phone number handles a small set of structured requests.

Products Used

The example metadata lists:

telnyx_products: [Voice, AI Inference, Messaging]
integrations: [Slack]
language: python
framework: flask

Voice handles inbound calls and TTS. AI Inference classifies each request. Messaging sends the guest confirmations. Slack alerts staff.

Architecture

inbound phone call or sms
  -> telnyx voice/sms webhook
  -> flask app
  -> look up room by caller id
  -> (if unknown) ask for room number, extract it
  -> ai inference: classify department + urgency
  -> urgent phrases override classification
  -> append to service_requests
  -> sms confirmation to guest
  -> slack alert to staff
  -> staff: POST /requests/<id>/complete
  -> sms fulfilment to guest

State lives in memory (ROOMS, service_requests, calls) for the demo. For production, replace the room and request stores with your PMS database (Opera, Mews, Cloudbeds).

Run The Example

git clone https://github.com/team-telnyx/telnyx-code-examples.git
cd telnyx-code-examples/hotel-guest-services-python
cp .env.example .env
pip install -r requirements.txt
python app.py

Fill in the environment variables:

TELNYX_API_KEY=KEY...
TELNYX_PUBLIC_KEY="-----BEGIN PUBLIC KEY-----..."
MAIN_NUMBER=+1...
CONNECTION_ID=...
STAFF_SLACK_WEBHOOK=https://hooks.slack.com/services/...
AI_MODEL=moonshotai/Kimi-K2.6
PORT=5000

Expose the local webhook:

ngrok http 5000

Configure the Voice API application webhook URL:

https://<ngrok-id>.ngrok-free.app/webhooks/voice

Configure the Messaging Profile inbound webhook URL:

https://<ngrok-id>.ngrok-free.app/webhooks/sms

Demo Script

Call the Telnyx number from a phone whose caller ID is +15559001234 (room 101, guest Smith) or +15559005678 (room 205, guest Chen):

hi, this is room 205, can i order a club sandwich and a sparkling water please?

For the unknown-caller path, call from any other phone. When prompted for your room number, say "Room 205 please".

Expected result:

• Agent confirms by voice that the request is logged
• curl http://localhost:5000/requests shows the open request
• Slack shows the new request in the configured channel
• Guest receives an SMS confirmation

When staff are ready:

curl -X POST http://localhost:5000/requests/0/complete

Guest receives the fulfilment SMS.

Edge Cases The Example Handles

• No speech detected on a gather — reprompt the caller
• Unknown caller — ask for room number, extract it from the next utterance
• AI Inference returns malformed JSON — fall back to concierge / normal with raw text as the summary
• Slack webhook down — log and continue, SMS still goes out
• SMS send fails — log and continue, Slack still goes out
• Duplicate webhook delivery — event IDs are tracked for one hour and deduplicated
• Urgent phrases (fire, flood, locked out, gas leak) — overridden to urgency: urgent and posted to Slack with a :rotating_light: prefix even if the LLM classifies them otherwise

Production Considerations

Before using this pattern in production, add:

• Persistent storage: replace ROOMS and service_requests with your PMS database
• Redis-backed deduplication so event IDs survive restarts
• Queued Slack and SMS side effects so webhook responses stay under Telnyx's timeout
• Call recording for quality assurance (record_channels: "dual" on answer)
• Real paging (PagerDuty, Opsgenie) for urgent requests instead of Slack
• Multi-language system prompts and TTS voices per property