Configuring Einstein Bots

Configuring Einstein Bots

What’s all about Einstein bots?

1. Nothing but chat bot

2. Digital Support Agent

3. Only designed to support specific limited queries like order status, password reset, Book an Appointment

4. This will avoid long waiting queue for simple queries

5. Not built to replace the Agents / Human

6. Also not designed for long conversational engagements.

7. Any time you can connect with agent by clicking ‘Transfer to Agent’ from your chat window.

What are the Perquisites?

1. A Service Cloud License

2. Chat Agent License

3. Chat agent must be enabled in your org to use Bot’s

4. Chat button , Deployment must be enabled

5. Lightning Experience should be enabled

6. Einstein Bot setup only available in LEX

Benefits of Einstein Bots:

1. Chatbots can quickly answer specific questions for customers to make them happier. It will reduce workload for support agents to resolve.

2. By Using Chatbots we can avoid long waiting queue for simple queries .Customers spend less time waiting “in the queue.” They get answers to common questions quickly in a chat window.

3. Since Chatbots can deflect easy cases, agents can spend more time to complex issues.

4. Bots can instantly welcome customers with a branded greeting in a chat window and can connect to agent as and when needed.

5. When bots are connected to NLU technologies, they can learn how to respond to customers appropriately and let agents do more complex work.

Enable Chatbots for your org:

1. Follow below path to enable chatbots in your org

From Lightning --> Setup --> Service --> Service Cloud Einstein --> Einstein Bots

>> Make sure to enable settings as per the requirement

Log conversations: If you enable this will store chat transcripts, including customer data in Event logs

Bot options Menu: It will display persistent option menu in Einstein chat window 

Create an Einstein Bot:

1. Click New from the Bots Page

2. Provide name of your Bot

3. Click next and enter Einstein Bot Greetings Message

4. If you want to add menu Item  as optional

5. Click next to Create your First bot

>> Close the screen and launch your newly created Bot

What is use of - Overview, Dialogs, Variables, Entities, performance, Intents, Model Management?

Overview: Is the place where you can add channels, set the bot response time and Bot context user

 

1. Channel - Here you have an option to add three type of channels

i). Apple Business Chat --> It allows companies to communicate directly with customers via iMessage. In order to use it, companies must login via a customer service platform, with Salesforce among those supported.

ii). Text  --> Lightning Messaging Service

iii). Chat --> Chat Agent (Formerly Live Agent)

2. Bot Response Delay – Delay bot response to simulate typing

3. Bot User – Each bot must have bot user. This will determine what bot can access or do. If your bot require additional features like field service, create a custom chat bot user and assign required permissions to user.

Dialogs: Is the place where you can perform different operations like Message, Questions, Actions, and Rules

1. Messages -- The Message that you are displaying to the visitor.

2. Questions -- Bot can place a question to visitor.

>> Also you can provide menu/buttons predefined options to your customers

3. Actions -- Where it will run some apex code/flow to get status or create appointments/orders, example order/ Appointment. Also you can send email and object search (Currently you can only search knowledge Articles as this is Beta now)

4. Rules -- It will check the condition like our simple if - else condition. What it supposed to do is defined in Rule actions. We have different Rule actions (Like Call Dialog, Clear Variable Value, End Chat, Redirect to Dialog, Set Routing Type, Set Variable, and Transfer) available currently.

Variables:

1.Variable that can be used to store customer inputs to Bot Questions

2.We can reference this variable as input to Apex class/Flow by using syntax like below or while printing output. Example {!aNumber}

Entities:

1.Any Case / Appointment / Order management system typically follow the certain format. In our system Appointment starts with A and '-' then 7 numbers. Entities helps chat bot to recognize when customer enter an Appointment number in a free text conversation.

2.Create an entity name "AppointmentNum: with type Pattern and RegEx mask ^A\-\d{7}$.

3.If user enters any appointment number to know status like 'A-0000001', system will read the input and pass it to your Apex class / Flow to get the status.

Performance: This is the place where you can track your bot performance like total sessions, average interaction time Escalation Rate, Event logs, historical performance of bot 

1. Dashboard – You can track your bot performance like total sessions, average interaction time Escalation Rate, historical performance of bot.

2. Event Logs – Here you can see your bot conversation logs for debugging

>> Click on each Session to see bot logs

Model management: It have two section below

1.Model – This will give the average accuracy of your all trained intents and also gives the accuracy of individual intents

2.Bot training – From here you can train your Bot. In order to train the bot first you need to create intents with minimum 20 utterances

Follow below path to create Intents:

Setup --> Einstein --> Einstein Platforms --> Einstein Intent Sets

>> Click new create to create new intent Sets

>> After creating Intent Sets --> Click Edit to create your intents

>> After creating intents add minimum of 20 utterances to train your bot but for better results we suggest you to add 150 utterances

>> After adding minimum utterances train your bot from bots ‘Model Management’, follow below path to train your bot

From Lightning --> Setup --> Service --> Service Cloud Einstein --> Einstein Bots --> Open Your Bot --> Model management --> Bot Training

>> Last but not least, make sure to activate your Bot.

You are all set to test your Bot.

Test Einstein Bot Experience:

>> Add your chat agent Button to Bot from Bot Over view page and set disable ‘Require Agent online’ check box.

>> Create Embedded Service Deployment to test this as Preview from your Bot, select same chat button and deployment to Embedded Service Deployments

>> Go to bot Dialog page and click ‘Preview’ button --> Click Chat with Expert button.

>> Fill Pre-chat details and click ‘Start Chatting’ to chat with bot

>> It will connect to your bot and conversation