3.8.5 Live Agent Escalation
Live Agent Escalation
- 1 Live Agent Escalation
- 2 Solution Customization
- 3 Video link
Why was it created?
Though Orbita helps you create intelligent chatbots that can address simple and complex queries of the users, there is always a situation where the patients need additional support that the chatbot cannot provide (i.e. support from a call center rep or nurse to answer clinical questions). The Live Agent Escalation solution gives the ability to transfer the conversation to a human agent.
What is its purpose?
The staffed call centers will have longer wait times and limited support hours. Oftentimes the questions/issues go unresolved because the user doesn’t know who to call.
The purpose of the Live Agent Escalation solution is to provide you with a chatbot that offers 24/7 real-time convenience to customers and improves customer service productivity, conversions, and sales.
What does it mean to the business?
Improved customer experience and fast customer inquiry resolution equate to better brand perception.
In-depth reporting and transcripts to support data analysis and decision-making.
Scalability (Agent can work with multiple customers at the same time).
How does it work?
The Live Agent Escalation solution provides a more engaging, pleasant, and constant experience and helps the users to access a Human Agent on-demand, or when swearing/frustrated words are used. It also Automates triaging the issues and helps connect users to certain agents based on user needs and the agent’s expertise.
How do you install it?
Login to the Experience Manager and select a project
Navigate to your project > Project side navigation menu > Create > Solution Center> Solutions.
Click the Add button against the Live Agent Escalation solution to install it to your project.
You will see the below pop-up message after the solution is installed successfully.
You can click on the Remove button to uninstall the Program Registration Solution from your project.
Post Installation
Configure Settings
After installing the Live Agent Escalation solution, click on the Configure Button.
This will open a window as shown below.
You can configure the bot name, MS teams settings to be used here.
Sample code{ "user_name": "Bot User", "tenant_id": "sample-82f8-4e96-b39d-33dd498cf4a1", "teams_id": "sample-5f03-4237-a0a5-25ad44950e9e", "channel_id": "sample-0FbnLHPOb8-YRWt0xqOZ5PnBS5VoogCLxWy6c1%40thread.tacv2", "callback_url": "https://sample.orbita-eng.cloud:8443/oeapi/v1/microsoft/token", "client_id": "sample-dbb6-44db-97f8-c7d7b04b55cc", "secret": "sample-Myt6y.L-5mfSLMn-SX1MvyED7o4", "scope": "https://graph.microsoft.com/ChannelMessage.Send Group.ReadWrite.All ChannelMessage.Read.All offline_access" }
Change the config settings as required and click Save.
The settings mentioned above are needed to integrate the chatbot with the O365/MS Graph API which is utilized for communication with the Microsoft Teams environment. The details of these settings are mentioned below.
A Microsoft Account to act as a ‘System User’ that posts messages in teams. It needs to be scoped with the following permissions.
ChannelMessage.Send
Group.ReadWrite.All
ChannelMessage.Read.All
offline_access
For clarity, we recommend naming this user ‘Bot User’ - but any name will work.
The following settings also need to be obtained, and placed within the LiveAgent configuration:
tenantId
clientId
secret
Refer to this documentation to get the tenantId. https://docs.microsoft.com/en-us/onedrive/find-your-office-365-tenant-id
Refer to this documentation to get the clientId and secret.
https://docs.microsoft.com/en-us/office/dev/store/create-or-update-client-ids-and-secrets
These configurations and information will need to be obtained from the Sys Admin that is responsible for your O365 administration.
The below graph APIs are used and they require OAUTH O365 authentication.
https://graph.microsoft.com/v1.0/teams/[tenantId]/channels/[channel_id]/messages
Post a message to a channel, this returns the [thread_id] which we associate to the user’s session/socketID.
https://graph.microsoft.com/v1.0/teams/[tenantId]/channels/[channel_id]/messages/[thread_id]/replies
[GET] Poll for replies from the agent, this response array is pushed via socket to the end-user on changes.
[POST] Send a message from the bot to the thread.
Chatbot User’s and Agent’s view
After all the configurations are done and when the Live agent intent is called in the chatbot by the chatbot user, the conversation is transferred to the Live Agent.
Chatbot user view
Before the Live Agent is connected
After the Live Agent is connected
Agent view
The chat will get escalated to a Live Agent with the transcript of the conversation between the chatbot and the chatbot user.
The Live Agent can then take over the conversation and reply to the user.
Dashboards
After the solution installation, submit a request to support@orbita.ai to enable the pre-built dashboards from the Project side navigation menu > Report > Dashboards.
What dashboards & visualizations are available and what purpose do they serve?
You will have a dashboard that will capture the Chatbot analytics such as Escalated users, Escalation Path, Reasons for Escalation, and so on.
Can I customize the dashboard?
You can create your own visualizations for the Live Agent Escalation in Insights. Refer, Guide to Orbita Insights
Live Agent Escalation solution data is captured in the data-<environment>-backend.escalationdata
index pattern/schema.
Solution Customization
How can I update content?
You can add the utterances the chatbot user might say when he/she wanted to talk to a human agent.
Go to Project Side navigation bar > Create > Intent Library > LiveAgent
Click on the vertical ellipses button and click on Edit
Add the utterances.
Refer to 3.2.3 Intents, Utterances, Slots, List
How to update the branding of the chatbot?
You can customize the chatbot appearance using CSS and JS. For more information on customizing the appearances of your chatbot, refer 3.9.1 Orbita's Chatbot -- Coming Soon | Customizing your chatbot
Integrations
The current solution integrates with LiveChat through which the human agents can take over the conversation. You can choose to integrate the solution with the customer care software of your choice such as Microsoft Teams, Salesforce, and so on.