Retail fraud and dispute prebuilt component

The retail fraud and dispute prebuilt component checks a user's authentication level. Once the user has been authenticated, the component provides information about the user's most recent or past orders. The user may then submit a dispute related to an order or product. This component uses the Retail authentication prebuilt component to authenticate users. The component also uses Generators for post-call summarization.

Limitations

In addition to the common prebuilt component limitations, this prebuilt component has the following limitations:

  • This component does not encrypt any data.

Sample dialogs

The following sample dialogs demonstrate common scenarios handled by this prebuilt component. The sample dialogs demonstrate behavior when the component is called using telephony, which triggers the Default Welcome Intent and the first agent response at the start of the conversation.

Dispute for individual product

End-user has not been previously authenticated. The agent collects details about previous orders to verify the user's identity. Once the user has been authenticated, the agent collects information about the dispute including dispute reason and the specific product the user would like to dispute. After collecting required details, the agent then submits the dispute and provides the user with the confirmation number.

User input Agent response
Hello John. In order to verify yourself, we'll need your order ID, transaction date, and amount before filing your fraud or dispute case.
Could you please provide the last 5 digits of your order ID?
12345
The last five digits of the order ID that we got are 12345, is that correct?
yes
Great, we are done collecting the order ID.

Please provide your date of order. Please say your order month, followed by your order day, and then your order year.
2023-10-12
Okay, your order date is October 12, 2023, is that correct?
yes
Great, we have collected your order date.

Please tell us the dollar amount of your order. For example, if your order transaction amount was $100.30, say "100".
97
Okay, your order transaction amount is $97, is that correct?
yes
Please specify the nature of fraud or dispute you want to report. You may say something like: payments or billing, product, returns & refund, data security or privacy concern, delivery or shipping, packaging, or say "others".
product
Thank you for letting us know. To confirm, the reason for fraud or dispute is product. Is this correct?
yes
Please tell us whether the selected issue is related to an entire order, an individual product, or any other concerns.
product
Are you calling about your most recent order number ending with 12345 containing 2 lamp, 1 shirt and 1 other product placed on 2023-11-10 at 06:00 PM?
yes
We've found some products in your order. Please tell us which product you would like to report. Product 1: Google lamp, Product 2: Google shirt, Product 3: Google jeans.

Please say or enter the product number of the item you want to select, or say "next" to show more products.
3
To verify, you are interested in raising fraud or dispute case for the Google jeans. Is this correct?
yes
In a few words, please tell us the detailed reason for this fraud or dispute.
poor quality
To confirm, the fraud or dispute reason provided is poor quality. Is this correct?
yes
Thanks for the confirmation. We have raised a ticket to our Fraud or Dispute team. You should hear from them shortly. Your ticket number is 6595513.

Would you like to hear the confirmation number again?
yes
Your ticket number is 6595513.

Would you like us to send a confirmation message containing the ticket number?
yes
Would you like to receive your confirmation message via your registered phone or email?
phone
Just to confirm, would you like to receive the confirmation message on your phone? Please say yes or no.
yes
Thanks. A confirmation message has been sent to your registered phone. This may take some time.

Do you wish to raise fraud or dispute ticket for another order?
no
Thank you for calling us. Our team will follow up on your ticket.

Naming conventions

This prebuilt component uses the following naming conventions:

Feature Format Example
Flow [Component Name] Fraud and Dispute
Entity Type prebuilt_components_[component_name]_[entity_type] prebuilt_components_product_dispute_reason
Webhook prebuilt_components_[component_name]:[webhook_action] prebuilt_components_fraud_and_dispute_order:get_security_answers

Input parameters

Input parameters are parameters that are used to configure certain behaviors of the component. The parameters will be used by one or more conditions within the flow to determine how the component should behave. Flow-scoped parameters must be set on the start page of the component. Session-scoped parameters may be set either by a calling flow, or on the start page of this component.

This prebuilt component accepts the following input parameters:

Parameter Name Description Input Format
$session.params.phone_number (optional) User's local phone number, without the country code, used to identify the user. string
$session.params.all_recent_order_id (optional) This parameter lists order IDs associated with the authenticated user, sorted by the order placement date in descending order. The first element of the array represents the most recent order. If a user has no orders, the value of this parameter is null. List (string)
$session.params.email (optional) The email registered with the account. string
$session.params.user_first_name (optional) User's first name, used to greet and confirm ownership of the account. string

This prebuilt component allows the following optional parameters to configure the behavior of the component, including retries.

Parameter Name Description Format
$flow.max_retry_another_account_counter Specifies the number of retries allowed when selecting whether to continue with an alternate account. The default value is 2. integer
$flow.max_retry_telephone_number Specifies the number of retries allowed when providing a registered telephone number. The default value is 2. integer
$flow.max_retry_order_number Specifies the number of retries allowed when providing the last five digits of an order number. The default value is 2. integer
$flow.max_retry_another_order Specifies the number of retries allowed when selecting another order for return in cases where the order ID is not found or the order is not eligible for return. The default value is 2. integer
$flow.max_retry_select_order_number Specifies the number of retries allowed when providing the full order number in cases where multiple orders share the same last five digits. The default value is 2. integer
$flow.max_product_display_counter Specifies the maximum number of products to present in a single turn. The default value is 2. integer
$flow.max_product_display Specifies the maximum number of products to present over the entire conversation. The default value is 9. integer
$flow.max_retry_product_number_counter Specifies the maximum number of retries allowed when selecting a product to return. The default value is 2. integer
$flow.max_product_report_confirm Specifies the maximum number of retries allowed when confirming the selected product. The default value is 2. integer
$flow.max_retry_dispute_reason Specifies the number of retries allowed when confirming the reason for dispute. The default value is 2. integer

To configure the input parameters for this component, expand for instructions.

  1. Open the Dialogflow CX console.
  2. Choose your Google Cloud project.
  3. Select your agent.
  4. Select the Build tab.
  5. Click the imported component in the Flows section.
  6. Click the Start Page in the Pages section.
  7. Click the true Route in the Start Page.
  8. In the Route window, edit the Parameter Presets values as needed.
  9. Click Save.

Output parameters

Output parameters are session parameters that will remain active after exiting the component. These parameters contain important information collected by the component. This prebuilt component provides values for the following output parameters:

Parameter Name Description Output Format
phone_number User's local phone number, without the country code, used to identify the user. string
all_recent_order_id This parameter lists order IDs associated with the authenticated user, sorted by the order placement date in descending order. The first element of the array represents the most recent order. If a user has no orders, the value of this parameter is null. List (string)
email The email registered with the account. string
user_first_name User's first name, used to greet and confirm ownership of the account. string
dispute_ticket_id Ticket ID for the submitted dispute. string
transfer_reason This parameter indicates the reason the flow exited, if it was not successful. The returned value is one of:

agent: the end-user requested a human agent at some point during the conversation.

denial_of_information: the end-user declined to share information requested by the component.

max_no_input: the conversation reached the maximum number of retries for no-input events. See no-input built-in events.

max_no_match: the conversation reached the maximum number of retries for no-match events. See no-match built-in events.

webhook_error: a webhook error occurred. See webhook.error built-in event.

webhook_not_found: a webhook URL was unreachable. See webhook.error.not-found built-in event.
string

Generators

This component uses Generators to enable specific parts of the conversation.

Summarization

The Summarization generator is used to generate a post-call summary describing the conversation handled by the component.

Basic setup

To set up this prebuilt component:

  1. Import the prebuilt component.
  2. Configure the provided flexible webhooks with configuration describing your external services, see Webhook setup.

Webhook setup

In order to use this component, you will need to configure the included flexible webhooks to retrieve needed information from your external services.

Authentication

If you have not previously configured external services required for authentication, you will need to configure them to enable authentication of users for this component. See Verify user details webhook setup for detailed instructions to set up the prebuilt_components_retail_authentication:verify_user_details webhook.

Get order details

The prebuilt_components_fraud_and_dispute_order:get_order_details webhook is used by the component to retrieve order details for a specified order.

API request parameters

The following parameters are provided by the component as inputs to the API request.

Parameter Name Description Input Format
$flow.order_id Order ID to retrieve details for. string

API response parameters

The following parameters are pulled from the API response to be used by the component.

Parameter Name Description Output Format
order_date Date the order was placed, in YYYY-MM-DD format. string
order_time Time the order was placed, in HH:MM format. string
order_product_count The number of unique products within an order. integer
order_amount Dollar amount for the order. integer
order_payment_mode Indicates the payment mode used when placing the order. string
all_product_id List of product IDs for all products within the order. List (string)
all_product_name List of product names for all products within the order. List (string)
all_product_brand_name List of brand names for all products within the order. List (string)
all_product_quantity List of quantities for all products within the order. List (string)

To configure the Get order details webhook for this component, expand for instructions.

  1. Open the Dialogflow CX console.
  2. Choose your Google Cloud project.
  3. Select your agent.
  4. Select the Manage tab.
  5. Click Webhooks.
  6. Select the prebuilt_components_fraud_and_dispute_order:get_order_details webhook.
  7. Replace the URL in the Conversational Agents (Dialogflow CX) Webhook URL field with the endpoint for the service you want to integrate with. Select the proper Method in the dropdown.
  8. Review and update the Request body to form the proper request format for your webhook.
  9. Review and update the Response configuration to extract specific fields from your webhook's response. Do not modify the parameter names, as these are required by the component to access the returned field values.
  10. Review and update the Authentication settings as needed.
  11. Click Save.

Get duplicate order details

The prebuilt_components_retail_services:get_duplicate_order_details webhook is used by the component to retrieve detailed information in order to distinguish between orders when multiple order numbers have the same last five digits.

API request parameters

The following parameters are provided by the component as inputs to the API request.

Parameter Name Description Input Format
$flow.duplicate_order_id List containing the last five digits of order numbers shared between one or more orders for the user. List (string)

API response parameters

The following parameters are pulled from the API response to be used by the component.

Parameter Name Description Output Format
all_order_date List of order dates for duplicate orders with the specified last five digits in YYYY-MM-DD format. Each element in this array corresponds to the same element in the all_order_time array. List (string)
all_order_time List of order times for duplicate orders with the specified last five digits in HH:MM format. Each element in this array corresponds to the same element in the all_order_date array. List (string)

To configure the Get duplicate order details webhook for this component, expand for instructions.

  1. Open the Dialogflow CX console.
  2. Choose your Google Cloud project.
  3. Select your agent.
  4. Select the Manage tab.
  5. Click Webhooks.
  6. Select the prebuilt_components_retail_services:get_duplicate_order_details webhook.
  7. Replace the URL in the Conversational Agents (Dialogflow CX) Webhook URL field with the endpoint for the service you want to integrate with. Select the proper Method in the dropdown.
  8. Review and update the Request body to form the proper request format for your webhook.
  9. Review and update the Response configuration to extract specific fields from your webhook's response. Do not modify the parameter names, as these are required by the component to access the returned field values.
  10. Review and update the Authentication settings as needed.
  11. Click Save.

Post reporting details

The prebuilt_components_fraud_and_dispute_order:post_reporting_details webhook is used by the component to submit a fraud or dispute ticket.

API request parameters

The following parameters are provided by the component as inputs to the API request.

Parameter Name Description Input Format
$session.params.phone_number User's local phone number, without the country code, used to identify the user. string
$flow.order_id Order ID to submit a ticket for. string
$flow.dispute_product_id Product ID that the user wants to file a dispute for. string
$flow.order_product_dispute_reason Dispute reasons for the disputed product. Valid values are defined by the prebuilt_components_product_dispute_reason entity. string
$flow.order_product_dispute_reason_other If the user selected "others" as the dispute reason, this contains the specific dispute reason the user has provided. string
$flow.order_product_dispute_reason_detail Detailed dispute reason. string

API response parameters

The following parameters are pulled from the API response to be used by the component.

Parameter Name Description Output Format
order_product_dispute_ticket_number Ticket ID for the submitted dispute. string

To configure the Post reporting details webhook for this component, expand for instructions.

  1. Open the Dialogflow CX console.
  2. Choose your Google Cloud project.
  3. Select your agent.
  4. Select the Manage tab.
  5. Click Webhooks.
  6. Select the prebuilt_components_fraud_and_dispute_order:post_reporting_details webhook.
  7. Replace the URL in the Conversational Agents (Dialogflow CX) Webhook URL field with the endpoint for the service you want to integrate with. Select the proper Method in the dropdown.
  8. Review and update the Request body to form the proper request format for your webhook.
  9. Review and update the Response configuration to extract specific fields from your webhook's response. Do not modify the parameter names, as these are required by the component to access the returned field values.
  10. Review and update the Authentication settings as needed.
  11. Click Save.

Send message

The prebuilt_components_fraud_and_dispute_order:send_message webhook is used by the component to send confirmation details about the dispute to a user.

API request parameters

The following parameters are provided by the component as inputs to the API request.

Parameter Name Description Input Format
$session.params.phone_number User's local phone number, without the country code, used to identify the user. string
$flow.order_product_dispute_ticket_number Confirmation number for the submitted dispute. string
$flow.confirmation_channel The specific channel to send information to. Valid values are defined by the prebuilt_components_channel_email_phone entity. By default phone and email are supported. string

To configure the Send message webhook for this component, expand for instructions.

  1. Open the Dialogflow CX console.
  2. Choose your Google Cloud project.
  3. Select your agent.
  4. Select the Manage tab.
  5. Click Webhooks.
  6. Select the prebuilt_components_fraud_and_dispute_order:send_message webhook.
  7. Replace the URL in the Conversational Agents (Dialogflow CX) Webhook URL field with the endpoint for the service you want to integrate with. Select the proper Method in the dropdown.
  8. Review and update the Request body to form the proper request format for your webhook.
  9. Review and update the Response configuration to extract specific fields from your webhook's response. Do not modify the parameter names, as these are required by the component to access the returned field values.
  10. Review and update the Authentication settings as needed.
  11. Click Save.

Complete

Your agent and its webhooks should now be set up and ready to test.