Localizing chatbot conversations by using real-time translation
Administrators can enable real-time translation in chatbot conversations so that end users can interact with the chatbot and the support agent in a language of their choice. You can leverage a real-time translation provider such as Google Cloud Translation Services version 2 or Microsoft Azure Translator Speech API version 3 to localize chatbot conversations.
Real-time translation is supported by BMC Helix Chatbot, Skype for Business on-premises, Microsoft Teams, and Slack.
You can enable localization in chatbot conversations in the following ways:
- By creating multiple localized Skills for each language. For more information, see Localizing-chatbot-conversations-by-using-localized-Chatbot-Skills-for-each-language
- By enabling real-time translation. For more information, read through the content in this topic.
The following video (3:13) provides an overview on real-time translation:
Locales supported by BMC Helix Platform
The following locales are supported by BMC Helix Platform. Locales that are not listed here are currently not supported.
- English (default)—en
- French—fr
- German—de
- Italian—it
- Japanese—ja
- Korean—ko
- Portuguese—pt_BR
- Spanish—es
- Chinese—zh-CN
- Dutch—nl
- Swedish—sv
- Russian—ru
- Danish—da
- Norwegian—no
Locales supported by auto-categorization
leverages IBM Watson Assistant as the classification service provider for auto-categorization. To see the list of supported languages, see Supported languages in the online documentation.
Scenario: Enabling real-time translation for a chatbot
An organization - ABC has implemented an IT chatbot to resolve all IT-related queries. The primary language of the IT chatbot is English. The organization wants to use real-time translation so that its employees can interact with the chatbot in Japanese and French. The ABC company uses Google Cloud Translation Services for automatic language translation. In BMC Helix Chatbot, the administrator selects the translator provider (Google Cloud Translation Services) and provides the Google Translate API key. Before enabling real-time translation for the chatbots, administrators can test whether real-time translation is working correctly by using the BMC Helix Innovation Studio UI. After confirming that the translation works correctly, you can enable real-time translation in the IT chatbots.
To ensure that localization is enabled for live chat with support agents, the administrator at ABC uses the same Google Translate API service in BMC Virtual Chat.
Localization workflow by using real-time translation
The following image shows the workflow for enabling real-time translation in a chatbot:
The following table describes the tasks to be performed in the process of enabling real-time translation in chatbot conversations:
Task | Product | Action | Reference |
|---|---|---|---|
1 | Real-time translation provider (Google Cloud or Microsoft Azure) | Enable the translation service. | |
2 | BMC Helix Innovation Studio | In the centralized tenant configuration settings, configure the translation service provider. | |
3 | BMC Helix Innovation Studio | Configure the translation provider credentials. | |
4 | (Optional) BMC Helix Innovation Studio | (If you want to test real-time translation) Test whether the real-time translation works correctly. | |
5 | BMC Helix Innovation Studio | In each chatbot, enable real-time translation and set the default language for translation. | |
6 | (Optional) BMC Virtual Chat | (If you want to configure localization for live chat) Configure automatic translation in a live chat so that the support agent's messages are automatically translated to the user's locale. | Setting language options and translation in BMC Virtual Chat documentation |
Result of enabling real-time translation for chatbot conversations
Employees of ABC can now interact with the IT chatbot in English, Japanese, and French. When a French employee opens the IT chatbot and requests for an additional monitor, the IT chatbot informs the employee that the request requires manager approval. The employee then requests to chat with a support agent. The agent asks the purpose of requirement. The employee types the message in French, which is translated for the agent in English. The agent types the reply in English, which is translated to French for the employee. The interaction continues until the employee's request is resolved.
- What happens if the user's language is not supported by real-time translation
If the user initiates a conversation in a language that is not supported by real-time translation, by default, the following message is displayed:
I have not been trained in your language but am able to translate on the fly. This may impact the quality of my responses.
- Standard error messages
If the real-time translation configuration is incorrect, the following message is displayed:
An error occurred while connecting to the translation service call.
If the real-time translation configuration is correct, but the translation provider service generates an error, the following message is displayed:
An error occurred during translation <additional description from the service> .
- Live chat localization
If the IBM Watson Assistant Skills configured with a chatbot is non-English, for example, in Japanese, then the agent receives messages in Japanese. To ensure that the agent receives the messages in English, you must enable translation in BMC Virtual Chat.
Example of a chatbot conversation in a supported locale
The following image shows a sample conversation in French after enabling real-time translation:
Benefits of enabling real-time translation for localizing chatbot conversations
- Support for more number of locales when using real-time translation.
- Eliminate the need to create multiple chatbot Skills for each language. If you have localized chatbot Skills, you can continue using them. To add new locales, you can enable real-time translation.
- Ability to select chatbots for which you want to enable real-time translation.
- Easily test real-time translation for chatbot before implementing it in the production environment.