Dydubox
Last updated
Last updated
Tous droits réservés @ 2023 dydu.
Dydu provides access through the BMS to the Dydubox, an interface to make its customers even more autonomous in their chatbot's deployment for their intranet or website.The design and functionalities offered by the chatbot to the end-user are entirely customizable without using a single line of code. You can quickly try different styles, labels, or chatbot features and adjust them whenever you want.
Dydubox is available through the Integration > Channels menu.
Three choices are available for the creation of a chatbot :
create a new chatbot from scratch
import the style of another chatbot to your chatbot
create a chatbot using dydu's Github repository and then import and publish it through Dydubox (for developers)
You can also create an integration by importing a configuration file in the format ZIP.
There are 2 types of integration:
Dydu integration: a dydu integration creates a standard version of the chatbot. Via this type of integration, users can personalize what their chatbot look like.
Github integration: it applies to a rare case when certain clients who have non-standard chatbots have to re-publish their chatbot on the Internet. By creating a Github integration, users can only publish the chatbot, but can not take advantage of our personalization options.
Once the Channels menu is open, to start customizing your chatbot you just have to:
Click on the Create an integration button.
Choose the Multiplatform option.
Click on Create from scratch option.
Throughout the customization of your chatbot, a preview of the final result is available on the right.
From the beginning of the customization of your chatbot interface, you can choose two principal elements of the design of your chatbot: the color and the avatar.
By default, we propose some predefined colors for your chatbot. If none of these proposals suit you, you can pick a color from the color palette or indicate a hexadecimal color code.
For the sake of consistency, the chosen color applies to the following elements:
the chatbot's header,
action buttons:
main actions: the background color corresponds to the chosen color,
secondary actions: the button's outline corresponds to the chosen color.
the user's question bubble with a slight opacity (transparency),
links applied to text content,
the background of the text next to the teaser (the teaser corresponds to the chatbot in its "reduced" state).
The uploaded avatar will be displayed on three elements :
the teaser
the header
the bot answer bubble (optional)
In this section it is possible to enable or disable features of the chatbot dialog display.
Allows to display the avatar of the bot in the dialog next to the bot's response bubble.
You can enter a text to be displayed above the bot's response bubbles.
It is possible to customize the avatar according to the matching result. The avatar can thus change according to 4 different states. You can customize the images of each of these different states:
by default
when the bot understands users request and has a direct match
when the bot suggests a reword,
when the bot has not understood.
Within the header several options for managing the display of the chatbot can be added :
This option allows the end user to minimize the chatbox (“teaser” mode).
the chatbot closes and does not appear on the web page anymore. To display it again, the user must refresh the page. Before closing, a window specifies these conditions and asks the end user for confirmation. The text in the displayed window can be customized.
This option allows the end user to extend the chatbox and thus switch it to full screen. There is no preview available for this option within the dyduBox.
It is possible to configure the maximum number of characters allowed for user questions. By default, the value is 100. It is also possible to choose whether or not to display the character counter. If displayed, the character counter appears in the input field and will be counted down as the end user types.
It is possible to deactivate a language as long as the corresponding knowledge is not finalized. This feature is especially used when you add a new language to your bot: although the bot understands this language directly, the content on the knowledge base side is not ready/finalized so you don't want to offer it to end users. When the chatbot offers several languages, the user has a drop-down list within the chatbot to select the flag corresponding to the language of his choice.
It is possible to choose whether or not to display another tab within the chatbot. This tab is called contact tab. It allows you to highlight the different ways to contact you, whether it is your social networks, emails or telephone. The texts in this tab are customisable.
If you need to display less then the three sections for the contact tab, please use the CSS editor. Here
You can choose whether or not to display an information banner. This is a text box that is positioned at the top of the chatbox. It is permanently present unless the end user decides to close it. It allows you to focus on important information. The content of the information banner is customizable. In the banner you can choose to add a "more information" button with an external link.
You can export the ongoing conversation with the chatbox via email. To do this, activate the "exportConversation" option in the "moreOptions" category of the "Debug" menu found in Channels. An icon resembling a cogwheel will then appear in your chatbox toolbar, providing access to the "Export conversation via email" option. This feature allows you to receive a PDF file containing a summary of the conversation you had with the bot via email.
You can print the ongoing conversation with the chatbox. To do this, go to the "Debug" menu in Channels and navigate to the "moreOptions" category to enable the "printConversation" option. An icon resembling a cogwheel will then appear in your chatbox toolbar, providing access to the "Print conversation" option. This feature allows you to access the print menu to print a summary of the conversation you had with your bot.
In this step you will be able to choose the features you want your chatbot to offer to the end user.
Example:
During the COVID period on booking sites the main question was about the possibility of a refund.
Instead of letting the user type in the question, the sites highlighted the relevant knowledge throughthe top knowledge feature.
This feature improves the user experience by presenting the most frequently asked questions at the beginning of the conversation. There are two possibilities:
let the bot automatically calculate the top most called knowledge over the last 30 days or
manually configure the top knowledge and thus choose the knowledge to be displayed in the top.
A Top knowledge page is dedicated to this purpose and allows you to search and select the knowledge to be displayed in the top. The display of a top knowledge in the chatbot is managed within the dyduBox. This is an option to be activated. If no manual top knowledge is defined then the 3 most called knowledge over the last 30 days will be displayed.
This feature facilitates the work of the end user. Indeed, as soon as the end user starts to write his request, the dydu NLP starts and the bot dynamically suggests knowledge according to what is typed. By default, the number of autosuggestions is limited to 3 proposals. Knowledge that cannot be "reformulated" in the database will be excluded from the auto-suggestions.
Auto-suggestion uses matching to determine which knowledge to show. That is, if your knowledge uses a phrase "how to change my password" and the user starts typing "change" the knowledge "how to change my password" can be proposed.
In the context of an HR bot, a new employee wants to know the conditions relating to lunch and starts typing: "luncheon voucher".
Even before he sends his question, the dydu NLP brick starts up and performs an auto-suggestion.
The chatbot will immediately suggest the relevant knowledge it has in its database, such as:
Am I entitled to luncheon vouchers?
how to get my luncheon vouchers?
where to get my luncheon vouchers?
For your chatbox, it is possible to set up :
only one consultation space
more than one consultation space and let the user decide which one to select
A satisfaction questionnaire can be displayed for each response. In case of dissatisfaction, it is possible to ask other questions to the end-user to better understand his dissatisfaction. The configuration of the user path following a dissatisfaction is configured within the dyduBox. Two options can be independently activated:
suggest reasons for dissatisfaction,
suggest leaving a comment.
Statistics on dissatisfaction issues are available in the Analytics menu of the BMS.
If enabled, the reasons for dissatisfaction are displayed when the end-user clicks on the dissatisfaction icon. Three reasons are available:
We did not understand your question
The proposed solution did not satisfy you
Our answer was not clear enough
A fourth proposal can be implemented thanks to the global sentences of the bot.
It is not necessary to select a reason for dissatisfaction to continue the conversation.
When this option is activated, the end user has the possibility to leave a comment to explain his dissatisfaction through a "free text" field displayed in the chatbot. It is not mandatory to leave a comment to continue the conversation.
If both the reason for dissatisfaction and the comment are activated, the reasons for dissatisfaction are displayed first, followed by the comment field (once the reason for dissatisfaction has been entered).
In order to give best practices for using your chatbot to end users before they start to use it, you can display an onboarding for chatbot's. By default, this option is disabled. Thanks to this features, you can display 1 image and each images can have a related text explanation on three card maximum. By default dydu offers 3 images, you can modify them by the button "Modify" (accepted format : SVG, JPG, PNG). The first cards is complete with text by default in order to help, There is no character limit. Labels can be modified in step 2 in the "Label management" section.
All the texts (labels) present in the chatbot and not managed from the bot's knowledge base can be customized via a dedicated excel file, to be exported and re-imported into the dyduBox. For each language, a Configure button is available to download/import the excel file of the labels. The text changes made and imported are directly visible in the preview. More details : label management section.
When the user arrives on the chatbox for the first time, a text called GDPR disclaimer is displayed. Its purpose is to inform the end-user about the collection and processing of his/her personal data and to obtain his/her consent. The user must accept the GDPR clauses in order to access the chatbox. To date, the display of this GDPR disclaimer cannot be disabled.
Once accepted by the end user, an operational cookie, valid for 1 year, is deposited on the end user's device. This way, the RGPD disclaimer will not be displayed again when the user visits the site.
The content of the GDPR disclaimer can be customised using the label management section. In this page you can customize :
the text of the disclaimer
the text of the buttons
for the buttons, the primary color will be applied
When the personal data option is activated, the end-users can, directly from the chatbot, make their request for recovery and/or deletion of their personal data (which the chatbot has). Each request made will be directly recorded and sent to the BMS side in the RGPD menu dedicated to this purpose. You can thus consult the requests and process them individually.
The publication allows you to obtain the javascript script to be added in the pages where the chatbot must appear. When publishing, all the options chosen and configured for the chatbot are taken into account.
There are two possibilities for publication:
publish in staging,
publish in production.
Each publication mode (preproduction and production) provides dedicated JavaScript code. Thus, publishing in preproduction allows for testing the various choices made within the dyduBox without impacting the production chatbox. Publishing in production allows for directly pushing the changes made to the production chatbox.
For each publication, it is possible to:
access a preview (test) page: this is a web page that integrates only the chatbox;
download the entire chatbox code (using the Download button);
copy the provided script to transmit it to a third party or directly proceed with its integration;
send the integration code by email.
In case of an incident on your production site following a version upgrade or modifications, it is possible to revert to a previous publication version. You can revert up to 3 versions back. This feature is available starting from the version of November 28, 2023.
Reminder: It is important to test your configuration in preproduction before publishing it in production to validate your modifications and version upgrades.
Below is an example of restoring a previous edge version:
Click on the "Restore..." button: a modal opens;
Click on "Select a version to restore";
Select the desired version.
Click on restore.
A message appears to indicate that the restoration has been successfully completed.
