The function of BOT Manager is to perform BOT entity management, i.e :
- Create a BOT entity.
- Integrate BOT with FAQ Knowledge.
- Integrate BOT with Dialog.
- Integrate BOT with Auto Suggestion.
Login as ‘Trainer‘, ‘Supervisor’ or ‘Administrator’, for the 3Dolphins SRM login mechanism, do the steps described on the ‘Login‘ page.
Go to Bot Settings menu and click the bot manager menus, then you will see a bot manager page such as images below.
Create New Bot
Bots can only do what you program them to do in a specific channel. So, someone can’t sign in as a bot and do things that other members can do in 3Dolphins. Bots also can’t be set as trainer or Supervisors.
In the upper right corner, click the +new bot button. Then, you will see a pop-up of new bot, configure your bot user with some basic info (bot name, greeting text, leaving text and what bot says (invitation)). Once you’ve completed these fields, click Save button.
Field component explanation :
|Bot Name||Name of Bot to be created.|
|Greeting Text||Speech from the bot when the first incoming message arrives.|
|Leaving Text||Text from bot after bot or customer ends chat.|
|What Bot Says (Invitation)||What bot says to invite customer.|
Dialog is a branching conversation flow that defines responses to the defined intents and entities. You can attach dialog skill of Bot to response question from customer.
After creating the Bot, to attach the dialog to the Bot, click the dialog button. Then, you will see a pop up of dialog skill, select dialog and click attach button. If dialog skill change to green, it’s indicated the dialog success attach in bot.
Module of FAQ to answer frequently asked questions (FAQs) submitted by the customer. You can attach knowledge of Bot to response question from customer.
After create Bot, to attach the FAQ to the Bot, click the FAQ button. Then, you will see a pop up of FAQ Module, select module and click attach button. If FAQ module change to green, it’s indicated the module success attach in bot.
To deploy bots on the channel, select Bot and click the channel deployment icon ‘‘, Then, you will see a pop-up of channel deployment. Select channel on the ‘Available Channels‘ and click plus button ().
Messages that displayed by Bot when the customer does not response message from Bot after some time. (timeout period varies). Select Bot and click the timeout icon. Then, you will see a pop up of timeout message. Set the timeout message and click ‘‘ button. And then, type what bot says and choose language dialog timeout. Click “” button.
To set the message to be displayed when the bot unrecognized message from customer. Select Bot and click the unrecognized setting icon. Then, you will see a pop up of unrecognized setting. Type what bot says and choose language unrecognized message. Click plus button ().
To set the message to be displayed when the message from customer not yet known by bots (Knowledge is not yet available on bots). Select Bot and click the fallback message icon. Then, you will see a pop up of fallback message. Enter what bot says, choose language and choose type of fallback message.
3Dolphins Think Integration
Select Bot and click the integration icon. Then, you will see a pop up bot integration parameter.
Field component explanation :
|Client ID||Is used to identify the application.|
|Client Secret||This is the true secret key, which is stored on server side securely & not available to public.|
|API Endpoint||The point of entry in a communication channel when two systems are interacting. It refers to touchpoints of the communication between an API and a server. An API endpoint is basically a fancy word for a URL of a server or service.|
|Think News Endpoint||URL of a 3Dolphins Think.|
|Whitelist Keywords||Allow you to add words that you would like the system to add to the keyword cloud.|
Bot Option Configuration
Bot options configuration allows the users to configure multi-tenant bots. By using the multi-tenant bot feature, each bot has the bot option configuration and the users can configure each bot as needed. When the user does not configure the bot in the bot option, the bot will read the configuration in the system settings.
Select Bot and click the option settings icon, then you will see a pop up of bot option configuration,
Field component explanation :
|Field Name||Description||Example Images|
|Option List Button Image||When the image in the digital library is empty or even damaged, the system will read and replace it using the link contained in the button image list option.|
|Option List Button Action||You can configure label on the button.|
Example button action : Tanya
|Option List Button Message||You can configure message suggestions from bots.|
For example :
When there are several answers to a customer’s question, the bot will give a suggestion message and answer that comes close to the customer’s question.
Example type button message : Mungkin ini bisa membantu 🤔
|Bot To Agent||Some customers may require human attention. In these cases, you can program the bot to automatically transfer these cases to your customer support team that will use their agents to resolve the problem.|
For example :
When customers transfer to a live agent, the bot will automatically send message ‘bot to agent’ to indicate that they are now interacting with a live agent.
|Agent To Bot||Each agent solves the issue and closed ticket from the customer, the system will automatically transfer the message to your bot and send a notification message to indicate that they are now interacting with a virtual assistant (Bot).|
|Inactive Bot||You can configure a message from bot incase the customer service (agents) is currently inactive.|
For example :
When the customer tries to contact customer service, but no agent is available, the system will automatically send inactive bot message.
The assistant bot serves to display suggestion on the chat conversation of agent. For example, when Agent chit chat with customer and the message from the customer is contained in the knowledge, the response will appear on the chat conversation of agent as a suggestion for the agent to answer question from customer. This assistant bot feature is optional, you can enabled or disabled as needed.
Select Bot and activate the ‘Assistant‘ bot by clicking the ‘On‘ toggle.
To view step by step using the bot assistant, please refer to the FAQ Suggesstion documents.
Select bot and activate the contextual bot by clicking the ‘On‘ toggle.
To view step by step using the bot contextual, please refer to the FAQ Knowledge document and see in the FAQ knowledge contextual section of the FAQ.
You can input sentences that have the same meaning. This section contains instructions on how to add general synonym.
In the upper right corner, click the synonym button. Then, you will see a pop up form of add term and general synonym. Enter the term and synonym and then click plus button. In this section, you can enable, disable or delete of synonym. If synonym change to green, it’s indicates the synonym enable to Bot.
The function of Evaluate Bot is to testing and analyzing the output of Bots (to find out the output of the bot before changing FAQ and after changing FAQ).
Create validation script
The Validation Script is used as a Bot Parent reference to check the desired Bot response results. If the bot response results are still not appropriate, then the Bot Parent must manually correct the FAQ and Dialog until the Bot response results are as expected.
In the upper right corner, click the evaluate bot button. Then, you will see a evaluate bot page, in the upper right corner click the new script button.
You will see a pop-up to create new script. Fill in the fields and after you’re done, you can click the save button to add a new script.
Select evaluation script has been created, in the lower left corner click script button ‘” and enter a script “User Say” that will be tested.
Select Bot and click the play button “” . Wait for the test evaluation to succeed 100%. The results of the response bot will be the basis for subsequent validation, if the results of the bot response are still not appropriate, then it needs to be done adjustment in the FAQ or Dialog that is test.
Validate the script
Validation of the script is done when there is an addition to the FAQ or dialog, to help the parent bot in ensuring the FAQ or Dialog that is added does not cause the validation of the script that has been made to be invalid.
In the upper right corner, click the evaluate bot button. Then you will see aevaluate bot page, select the script to be validated.
Click ‘‘ button and wait until the final validate of the evaluation is successful 100%. If the evaluation icon changes to green it’s indicate the bot response results before and after the addition of the FAQ or Dialog are still valid (the same). and if the evaluation icon is red it’s indicate the script is invalid.
You can download the results of the bot evaluation that has been done, click “Download XLS” or “Download CSV“,
Upload Script Validation
In the upper right corner, click the evaluate bot button. Then you will see a evaluate bot page, in the upper left corner, click the download template button ‘‘.
Put the script in the template, and in the upper right corner, click new script button. Then, you will see a pop up form of new script. Enter script name and then click ‘Save‘ button,
Select script name has been created, in the upper right corner of script click the upload evaluation script button ‘‘.
On the pop up of upload script, click the choose button to find the file of evaluator template that will be upload.
Wait the process of upload is success.
If, upload file of evaluator template is success, then 3Dolphins will displaying success notification such as images below.
Automatic Language Detection
Currently 3Dolphins can detect the language of the incoming message, i.e ‘Indonesian‘ and ‘English‘. The Bot will respond the message in the same language as the input text of customer. If the customer message is not ‘Indonesian‘ or ‘English‘ language, the Bot will response the message of customer use unrecognized message.
3Dolphins can even change language during the conversation and the Bot will understand, such as images below.