Lab 1 - Orchestrating Multilingual Voice AI

Section Agenda

#	Task	Duration
1	Create the Autonomous AI Agent	20 min
2	Build & Import Actions in Webex Connect	25 min
3	Voice Flow Configuration & Channel Setup	20 min
4	Live Call Testing — English & Spanish	25 min

Introduction

Welcome to our first AI Agent activity at Cisco Live! In this lab, we will work with Cumulus Hospital — a multinational private hospital offering diagnostic services and medical specialties across the USA and Mexico.

Imagine this: a patient calls from Las Vegas or from Mexico City, needs a specialist, and wants to book an appointment — right now, in their own language. Your mission is to make that possible.

Key Learning

You will build an Autonomous AI Agent from scratch using the Webex AI Agent Studio, serving patients in both English and Spanish — handling information requests, validating schedules, and confirming appointments via SMS.

Section Independence

Section 1 is entirely independent. We have pre-configured the complete environment so you can focus exclusively on building your Autonomous AI Agent from scratch, connecting it to the Voice Flow, and validating the end-to-end multilingual experience.

The PodID Discipline

Always replace XXX with your assigned 3-digit ID (e.g., 001, 002). Your naming convention must start with the prefix Pod followed by your PodID (PodXXX). If you do not follow this convention exactly, you will overwrite your neighbor's work. Stay disciplined with your naming convention!

PodID as Patient ID

Always use the PodID as the Patient ID, (e.g., 001, 010, 060, 050).

Phase 1: Creating an Autonomous AI Agent

1.1 Access Control Hub & AI Agent Studio

Using Chrome, select the Admin_Lab Chrome Profile to open a new window.
Navigate to admin.webex.com and log in with your Admin POD credentials.

Pro-Tip

You can save the password on this Chrome Profile to make future login processes easier. Make sure you are using your POD credentials since all students are sharing the same tenant!

See how to access Control Hub
In the left panel on Control Hub, navigate to Services > Contact Center to open the Webex Contact Center administration.

Important

This navigation path is always required to access any Webex Contact Center configuration in this lab. We will NOT repeat this instruction again.
On the main page, navigate to Quick Links, find and click on Webex AI Agent to launch the AI Agent Studio in a new browser tab.

See how to navigate to AI Agent Studio

1.2 Create the Autonomous AI Agent

In the AI Agent Studio, click the black + Create agent button.

Pro-Tip

The AI Agent Studio offers templates to speed up initial bot creation. Since we are building from the ground up, we will start from scratch.
Click Start from scratch and then Next.

Select Autonomous as the agent type and fill in the fields as follows:

Field	Value
Agent Name	PodXXX_UseCase1_CumulusScheduler (Replace XXX with your PodID)
System ID	(Leave the one provided automatically by the system)
AI Engine	(Leave the default — Webex AI Pro 1.0)

In the Agent's Goal field, copy and paste the following:

You are the virtual agent for Cumulus Hospital. Your primary goal is to identify
the user's intent and provide accurate information based on the configured
Knowledge Base (KB). You will also assist the patient with appointment management
– scheduling and canceling medical appointments. Always be polite and professional
with the patients.

Click Create.

See how to create the AI Agent

1.3 Configure the AI Agent Profile

Once the agent is created, it will open automatically in the configuration page. Configure each section as follows:

Profile Tab

Field	Value
Time Zone	America/Los_Angeles
Welcome Message	Welcome to Cumulus Hospital. How can I help you today?

Instructions — copy and paste the following into the Instructions box:

A: Identity
You are the virtual assistant for the hospital. Your role is to identify the patient's
intent and perform actions accordingly. Examples of actions include: assisting patients
in scheduling, canceling, or rescheduling medical appointments; providing information on
services, hospital locations, facilities, medical laboratory and specialties offered by
the Hospital based on the Knowledge Base. Be empathetic, friendly, polite, and professional.

B: Task / Action
MANDATORY INITIAL ACTION: ALWAYS perform the [customer_details] action at the start of
the interaction to obtain the patient's name "PersonaName". Use these data for the
actions below:

Scheduling an Appointment:

1. Greet the patient by "PersonaName" and ask for the desired medical specialty "Type".
   You also need to ask for the desired date "Time_Slot" and time "Time" for this
   appointment.

2. Consult the KB and offer to the patient the locations where the desired medical
   specialty "Type" or medical service is provided. Collect the desired "Location" from
   the patient. Verify if the "Type" is provided into the chosen "Location", otherwise
   propose the correct options to the patient.

3. Ask for the desired date and time for the appointment. You need to guarantee that the
   day chosen by the patient is a weekday for June 2026 and the desired hour follows the
   KB rules (DO NOT allow the patient to choose weekend days). Guarantee those rules are
   being followed before creating the appointment. Otherwise, propose different options
   to the patient based on the KB rules about time and days.

4. Create appointment: Convert the "Time_slot" into format "YYYY-MM-DD" (using current
   2026 year) and use the [create_appointment] action with the collected details to create
   the appointment. DO NOT mention SMS confirmation if you don't have customer ANI
   {{ PersonaANI }}. If you have {{ PersonaANI }} information, you can mention that an SMS
   will be sent shortly with the schedule confirmation.

C: Response Guidelines

- Be clear and simple: use short, clear sentences.
- Avoid complex, long, or ambiguous structures.
- Maintain a fluid, natural conversation similar to human speech, using intonation
  and pauses.
- When the client is interacting via voice, avoid using numbering, bullet points, or
  list formats; use continuous text, especially for numbers and facts.
- Avoid speaking web addresses like http://.
- Use the correct number format, such as speaking phone numbers digit by digit.
- For currency, use the correct format
  (e.g., 50.99 = fifty dollars and ninety-nine cents).
- When speaking dates, state the full month name and the full year.
- Make sure that the response does not contain any irrelevant text or formatting like
  newline characters, asterisks, dashes, or other special characters.
- DO NOT respond in markdown. Do not provide medical advice.
- You don't need to repeat the entire address every time you mention a "Location".
  If you have already mentioned the full address once, just mention the city.
- When the customer is using a digital channel, use numbering or bullet points to
  list options and put each option on a different line.

D: Error Handling and Alternative Responses

- Clarifications: Ask again when unsure of the patient's intent.
- Fallback Responses: If an action fails: "I am sorry, we were unable to process
  your request. Shall we try again?"
- Inform the patient about alternatives from the KB to help with scheduling availability.

Click Save changes.

See how to configure the AI Agent Profile

1.4 Create the Knowledge Base

Before configuring the Knowledge tab, we need to upload the KB document that will be used by our AI Agent.

Prerequisite

Ensure you have downloaded the Knowledge Base document to your local machine. Click Download and save the file to your Desktop as: UseCase1_AIAgentCumulus.pdf

On the left panel of the AI Agent Studio, click Knowledge to open the KB page.
Click + Create knowledge base and fill in the following:

Field Value

Knowledge Base Name PodXXX_UseCase1_AIAgentCumulus (Replace XXX with your PodID)

Click Create.
Click Upload Files, then Add file.
From your computer, select the file: UseCase1_AIAgentCumulus.pdf
Click Open to upload the file, then click Process Files.
Click Close and keep processing to return to the AI Agent configuration.

See how to create the Knowledge Base

1.5 Associate the Knowledge Base to the AI Agent

Go back to AI Agent on the left panel and click on your AI Agent to open its configuration.
Navigate to the Knowledge tab and select your Knowledge Base from the dropdown:

Field Value

Knowledge Base PodXXX_UseCase1_AIAgentCumulus

Remember to replace XXX with your PodID

Each student has their own Knowledge Base. Selecting the wrong one will break another student's lab.
Click Save, then click Publish. Add a note "V1" in the Publish window and click Publish again.

See how to associate the Knowledge Base

1.6 Test the AI Agent — Knowledge Base Validation

Before proceeding with the actions configuration, let's validate the AI Agent using the Knowledge Base only.

Before You Begin

Make sure your headset is connected to your computer before starting the preview.

Click Preview and Start a call.

Inside PodXXX_UseCase1_CumulusScheduler verify the following:

Location Information: The agent provides accurate information about hospital locations.
Services by Location: The agent describes services available at each location.
Facilities: The agent answers questions about hospital facilities.
Preparation Information: The agent provides preparation instructions for procedures.
Specialty Availability: The agent identifies which locations offer specific specialties.

Need help? See the full test script

Role	Message
👤 User	Which are the locations in the USA?
👤 User	Which services do we have in the New York location?
👤 User	Do you provide parking at the hospital?
👤 User	What is the preparation for a blood test?
👤 User	Which locations offer a cardiologist?

Session Inspection: Click on Sessions and select the transaction you just made to open the transcription window. When finished, close the session window and return to the Configuration tab of your AI Agent.

See how it works

1.7 Explore Language & Voice Settings

Go to the Conversation tab on your AI Agent configuration.

Key Learning — Language & Voice

By default, the AI Agent is created in English using the Jennifer voice. You can explore other languages and voices by changing the values in this tab.

💡 When making a live voice call using the Flow Builder, the language and voice settings configured in the flow will overwrite the values set here. This is how we can reuse the same bot in different languages for AI Agent voice calls.

See how to change language settings

Do you want to try in Spanish? See the full test script

Once you've switched the language to Spanish en-US, go ahead and update the welcome message to:

Gracias por llamar al hospital Cumulus, ¿cómo te ayudo hoy?

Role	Use	Message
👤 User	Location	¿Cuáles son las ubicaciones en los EE. UU.?
👤 User	Services	¿Qué servicios tenemos en la ubicación de Nueva York?
👤 User	Facilities	¿Ofrecen estacionamiento en el hospital?
👤 User	Preparation	¿Cuál es la preparación para un análisis de sangre?
👤 User	Specialty	¿Qué ubicaciones ofrecen un cardiólogo?

Phase 2: Building the Actions

2.1 Create the Webex Connect Flow - customer_details

Now we will create the Actions for this AI Agent. First, we need to launch Webex Connect, where the Action flows will be created and used inside the Autonomous AI Agent.

Navigate to Control Hub > Contact Center > Quick Links > Webex Connect to launch a new Chrome tab.
In Webex Connect, locate your Service named PodXXX_LTRCCT2001_CUMULUS and click to enter it.
Click on Flows and then Create Flow.
Fill in the following and click Create:

Field Value

Flow Name UseCase1_AIAgent_customer_details

Method New Flow

Template Start from Scratch

See how to create the flow in Webex Connect

2.2 Configure the AI Agent Event Node

From the Integrations window, click on AI Agent.
Copy and paste the following JSON in the Configure AI Agent Event field:
```
{ "PersonaID": "001" }
```
Click Parse and then Save.

2.3 Configure the HTTP Request Node

Click and drag the HTTP Request node to the canvas, positioning it beside the AI Agent node.
Click the green circle from the AI Agent Event to connect it to the HTTP Request node.

See how to connect the nodes

Double-click the HTTP Request node to open its configuration and fill in the following:

Field	Value
Method	GET
Endpoint URL	`https://api.airtable.com/v0/appvCQpTfOVVLcRmR/tblLGr9Ibkh79FPaD?filterByFormula={PersonaID}='$(n2.aiAgent.PersonaID)'`
Header	Authorization
Value	`It will be shared by your proctor`
Connection Timeout	5000
Request Timeout	5000

Click + Add Another Header and fill in the second pair:

Header	Value
Content-Type	application/json

Test the HTTP node: Click Test and enter your 3-digit POD number as the input. Verify the return is a valid JSON response. Click Save.

See how to test the HTTP node

2.4 Configure Flow Outcomes

Click on the Configuration icon and select Flow Outcomes.
Verify that the first line Notify is Enabled.
Click the arrow to open Last Execution Status and configure as follows:

Field Value

Notification Settings Notify AI Agent

Define JSON Enter JSON

Copy and paste the following JSON:

See how to configure Flow Outcomes
Click Save and then click Make Live in the top-right corner to publish the flow.

Expected Warnings

You will see warnings advising the HTTP node has no termination — you can safely ignore them and click Make Live again. The flow will be published and ready to be used by your AI Agent.

2.5 Create the customer_details Action in AI Agent Studio

Navigate back to the Webex AI Agent Studio tab in Chrome.
Make sure you are on your AI Agent configuration. Go to Actions and click + Add actions. Select Fulfillment from the Create new action option.

Critical — Use Exact Names

Use exactly the action name and entity names specified below. Otherwise, your bot will not identify which action needs to be performed.
Fill in the fields as follows:

Field Value

Action Name customer_details

Action Description Ask Patient ID and use customer provided as "PersonaID" to get customer data like "PersonaName", "PersonaEmail".

See how to create the action
Click + New input entity and fill in the following:

Field Value

Entity Name PersonaID

Entity Type String

Entity Description 3-digit customer ID / Patient ID to be used as "PersonaID"

Click Add.
In the Fulfillment section, make sure Use Webex Connect Flow Builder is selected and configure as follows:

Field Value

Select Service PodXXX_LTRCCT2001_CUMULUS (Replace XXX with your PodID)

Select Flow UseCase1_AIAgent_customer_details

Click Add.

See how to complete the action

2.6 Import the create_appointment Flow

Due to time constraints, we will import the remaining action flow from an existing template.

Navigate to the Webex Connect tab. Find the Service ADMIN_LTRCCT2001_CUMULUS and click to enter it.

Do NOT modify anything inside this service

This service is shared across all students. Only Export the flow — do not change any configuration.
Navigate to Flows inside ADMIN_LTRCCT2001_CUMULUS. Find the flow UseCase1_AIAgent_create_appointment, click on the Actions arrow and choose Export. Download the JSON file to your computer.

See how to export the flow
Go back to Services and find your Service named PodXXX_LTRCCT2001_CUMULUS. Enter it, navigate to Flows, and click Create Flow.
Fill in the name UseCase1_AIAgent_create_appointment and select the JSON file you just exported to import it. Click Create.
Wait until the flow opens. Click Parse and Save on the first node.

See how to import the flow

Flow Architecture — What's Inside?

Take a moment to explore the nodes but do NOT change anything:

Node Function

Node 1 Data passed by AI Agent to this flow

Node 2 Airtable PUT that creates the scheduling / appointment

Remaining nodes Send SMS confirmation based on customer language
Click Make Live to publish your flow.

2.7 Create the create_appointment Action in AI Agent Studio

Navigate back to the Webex AI Agent Studio tab. Go to Actions and click + Add actions. Select Fulfillment from the Create new action option.

Critical — Use Exact Names

Use exactly the action name and entity names specified below.

Fill in the fields as follows:

Field	Value
Action Name	create_appointment
Action Description	Create an appointment for the patient using "PersonaID", "PersonaName" at the preferred day and time chosen by the patient. You must convert the date provided by the patient in the following format prior to executing the action: YYYY-MM-DD

Click + New input entity for each entity below:
use the arrow > to navigate from Entity 1 to Entity 8:

Entity 1 — PersonaIDEntity 2 — PersonaNameEntity 3 — HoursEntity 4 — Time_slotEntity 5 — TypeEntity 6 — LocationEntity 7 — LanguageEntity 8 — PersonaANI

Field	Value
Entity Name	PersonaID
Entity Type	String
Entity Description	3-digit customer ID / Patient ID to be used as "PersonaID"
Setting	✅ Required

See Example - Entity 1: PersonaID

Field	Value
Entity Name	PersonaName
Entity Type	String
Entity Description	This is the Patient name — use "PersonaName" obtained from the `customer_details` action
Setting	✅ Required

See Example - Entity 2: PersonaName

Field	Value
Entity Name	Hours
Entity Type	String
Entity Description	Time for the scheduling meeting chosen by the customer
Setting	✅ Required

See Example - Entity 3: Hours

Field	Value
Entity Name	Time_slot
Entity Type	Date
Entity Description	Date for the appointment chosen by the customer. Use the format YYYY-MM-DD
Setting	✅ Required

See Example - Entity 4: Time_slot

Field	Value
Entity Name	Type
Entity Type	String
Entity Description	Type of scheduling service desired by the patient that is offered by Cumulus Hospital. Consult KB to double check type and offer the location accordingly.
Setting	✅ Required

See Example - Entity 5: Type

Field	Value
Entity Name	Location
Entity Type	String
Entity Description	Hospital location (city and address) where the customer chose to take the medical service. It must match the Cumulus locations and service offers in this location according to the KB.
Setting	✅ Required

See Example - Entity 6: Location

Field	Value
Entity Name	Language
Entity Type	String
Entity Description	Global_Language being used by this interaction. Use value `{{ Global_Language }}`. If you don't have this value, use the language configured in this BOT.
Setting	✅ Required

See Example - Entity 7: Language

Field	Value
Entity Name	PersonaANI
Entity Type	String
Entity Description	Patient ANI to be used for sending SMS confirmation. Use the variable `{{ PersonaANI }}` to fill this information.
Setting	✅ Required

See Example - Entity 8: PersonaANI

In the Fulfillment section, make sure Use Webex Connect Flow Builder is selected and configure as follows:

Field Value

Select Service PodXXX_LTRCCT2001_CUMULUS (Replace XXX with your PodID)

Select Flow UseCase1_AIAgent_create_appointment

Click Add.

Save changes

2.8 Test the AI Agent — Full Actions Validation

It's TEST time! Click Preview and Start a call.

Inside PodXXX_UseCase1_CumulusScheduler verify the following:

Identity Capture: The agent performs the [customer_details] action at the start of every interaction.
Location & Services: The agent consults the KB to offer correct locations for the requested specialty.
Date & Time Validation: The agent only accepts weekdays in June 2026, between 8am and 6pm.
Appointment Creation: The agent creates the appointment and confirms the details using [create_appointment].
SMS Confirmation: If calling from a valid cell phone, an SMS confirmation is received.
Language Consistency: The entire interaction follows the configured language.

Need help? See the full test script — English

Role	Message
🤖 Prompt	Welcome to Cumulus Hospital. How can I help you today?
👤 User	I need to schedule an appointment
🤖 Prompt	(Agent performs `customer_details` action — asks for Patient ID)
👤 User	My ID is 001
🤖 Prompt	Hello [PatientName], what type of medical specialty are you looking for?
👤 User	I need a cardiologist
🤖 Prompt	(Agent consults KB and offers available locations for cardiology)
👤 User	New York
🤖 Prompt	What date and time would you prefer?
👤 User	June 10th at 10am
🤖 Prompt	(Agent creates appointment and confirms SMS will be sent)

Important

After testing, Publish your AI Agent with comment V2, so we can proceed to Task 3 and test it with a real live call.

Phase 3: Voice Flow Configuration

3.1 Export & Import the Voice Flow

Navigate to Control Hub > Contact Center > Flows.
Find the flow ADMIN_TEMPLATE_UseCase1_CumulusScheduling, click on the dots on the right and select Export to save the JSON file to your computer.

See how to export the Voice Flow
In Control Hub > Flows, click Manage Flows, then Create Flows from the dropdown menu.
In the new window, choose Import Flow and select the JSON file from the previous step. Click Next.
Give the unique name PodXXX_UseCase1Cumulus to this flow (Replace XXX with your PodID) and click Create Flow.
The new flow window will open. Click Get Started to begin editing.

See how to import the Voice Flow

3.2 Edit the Voice Flow

Verify Edit Mode is ON on the top toolbar.
Find the node VAV2 (VirtualAgentV2) and click on it to edit.
In the Virtual Agent dropdown, select your Autonomous AI Agent:

Field Value

Virtual Agent PodXXX_UseCase1_CumulusScheduler

In the State Event section, open the dropdown and fill in the Event Data field with the following JSON:

{'WelcomeMsg': '{{WelcomeMsg}}', 'PersonaANI': '{{NewPhoneContact.ANI}}', 'Global_Language': '{{Global_Language}}'}

Why Are We Passing These Variables?

Variable	Purpose
`WelcomeMsg`	Passes the welcome message from the flow to the AI Agent
`PersonaANI`	Passes the caller's phone number for SMS confirmation
`Global_Language`	Passes the language selected by the patient in the IVR

See how to edit the Voice Flow

Toggle Validation to ON to ensure no configuration errors exist.
Click Publish. Ensure the version label is set to latest, then click Publish flow.

3.3 Configure the Channel

In Control Hub > Contact Center, find Channels.
Find your Channel PodXXX_Cumulus_Voice and click to edit. Configure as follows:

Field Value

Routing Flow PodXXX_UseCase1Cumulus

Music on Hold defaultmusic_on_hold.wav

Version Label Latest

Click Save.

See how to configure the Channel

3.4 Update the Welcome Message

Navigate back to the AI Agent Studio tab and click on your AI Agent to open its configuration.
In the Profile tab, update the Welcome Message field:
copy and paste the complete value:

Field Value

Welcome Message {{WelcomeMsg}}
Click Save Changes, then click Publish. Add a comment V3 and click Publish again.

See how to update the Welcome Message

Phase 4: Live Call Testing

4.1 English Path Validation

Step 1 — Initiate Call

Dial your assigned Channel number from your mobile device or Webex App. (you can find it under the lab access instructions).

🔗 You can use your cell phone or log in to the Webex App using the additional user assigned to your POD to make the call.

Note: For detailed guidance, refer to the Getting Started section.

Step 2 — Language Selection

When the IVR prompts for language, press 1 for English.

Step 3 — Triage Interaction

Test Script — English

Provide your Patient ID when asked (e.g., 001) - (you can find it under the lab access instructions).
Request an appointment: "I need to schedule an appointment with a cardiologist"
Choose your preferred location from the options provided by the AI Agent
Select a date: Any weekday in June 2026, between 8am and 6pm**
Confirm the appointment details
(If calling from a valid cell phone) → Verify the SMS confirmation is received ✅

Need help? See the full test script — English

Role	Message
🤖 Prompt	Welcome to Cumulus Hospital. How can I help you today?
👤 User	I need to schedule an appointment
🤖 Prompt	(Agent performs `customer_details` action — asks for Patient ID)
👤 User	My ID is 001
🤖 Prompt	Hello [PatientName], what type of medical specialty are you looking for?
👤 User	I need a cardiologist
🤖 Prompt	(Agent consults KB and offers available locations for cardiology)
👤 User	New York
🤖 Prompt	What date and time would you prefer?
👤 User	June 10th at 10am
🤖 Prompt	(Agent creates appointment and confirms SMS will be sent)

4.2 Spanish Path Validation

⚠️ Before You Begin — Configure Your Browser for Spanish

It is recommended that you configure your browser environment to support Spanish before starting this path. This ensures the Agent Desktop correctly reflects the Spanish locale settings.

In Chrome, click the three-dot menu → Settings.
In the search bar, type Language.
Locate Spanish in the language list. If it is not set as the primary language, click the three dots next to it and select Move to the top.
Reload your browser page to apply the changes.

Step 1 — Initiate Call

Dial your assigned Channel number from your mobile device or Webex App. (you can find it under the lab access instructions).

Step 2 — Language Selection

When the IVR prompts for language, press 2 for Spanish.

Step 3 — Triage Interaction (Spanish)

Test Script — Spanish

Provide your Patient ID - ID De Paciente when asked (e.g., 001)
Request an appointment: "Necesito agendar una cita con un cardiólogo"
Choose your preferred location from the options provided by the AI Agent
Select a date: Any weekday in June 2026, between 8am and 6pm**
Confirm the appointment details
(If calling from a valid cell phone) → Verify the SMS confirmation is received in Spanish ✅

Need help? See the full test script — Spanish

Role	Message
🤖 Prompt	Bienvenido a Cumulus Hospital. ¿En qué puedo ayudarle hoy?
👤 User	Necesito agendar una cita
🤖 Prompt	(Agent performs `customer_details` action — asks for Patient ID)
👤 User	Mi ID es 001
🤖 Prompt	Hola [NombrePaciente], ¿qué tipo de especialidad médica está buscando?
👤 User	Necesito un cardiólogo
🤖 Prompt	(Agent consults KB and offers available locations for cardiology in Spanish)
👤 User	Ciudad de México
🤖 Prompt	¿Qué fecha y hora prefiere?
👤 User	10 de junio a las 10am
🤖 Prompt	(Agent creates appointment and confirms SMS will be sent in Spanish)

Lab 1 Complete — Congratulations!

You have successfully completed Lab 1 — Orchestrating Multilingual Voice AI.

Component	Status
Autonomous AI Agent	Created from scratch in AI Agent Studio
Knowledge Base	Uploaded, processed and associated to AI Agent
`customer_details` Action	Flow created in Webex Connect and linked to AI Agent
`create_appointment` Action	Flow imported, published and linked to AI Agent
Voice Flow	Imported, configured and published with multilingual support
Channel Configuration	Routing flow mapped to your assigned DID
Welcome Message	Updated to use dynamic variable from Voice Flow
End-to-End Testing (EN + ES)	All scenarios validated — KB, scheduling & SMS confirmation

Field	Value
Flow Name	UseCase1_AIAgent_customer_details
Method	New Flow
Template	Start from Scratch

Field	Value
Notification Settings	Notify AI Agent
Define JSON	Enter JSON

Field	Value
Action Name	customer_details
Action Description	Ask Patient ID and use customer provided as "PersonaID" to get customer data like "PersonaName", "PersonaEmail".

Field	Value
Select Service	PodXXX_LTRCCT2001_CUMULUS (Replace XXX with your PodID)
Select Flow	UseCase1_AIAgent_customer_details

Node	Function
Node 1	Data passed by AI Agent to this flow
Node 2	Airtable PUT that creates the scheduling / appointment
Remaining nodes	Send SMS confirmation based on customer language

Field	Value
Routing Flow	PodXXX_UseCase1Cumulus
Music on Hold	defaultmusic_on_hold.wav
Version Label	Latest