How to setup an Einstein Chatbot in Salesforce? : Sparkybit

How to setup an Einstein Chatbot in Salesforce?
by: Sparkybit
blow post content copied from  Forcetalks
click here to view original post


In this article, we will discuss Einstein Chatbot and consider Omni-Channel, Lightning Knowledge Base, Embedded Service Deployment, and more.

The first point to note: this functionality is available for Service Cloud User and Chat User feature licenses in Enterprise, Performance, Unlimited, and Developer editions. Chat User feature license can be obtained by purchasing Digital Engagement, an add-on for Service Cloud. Also, a significant limitation is that 25 bot conversations per month are available for each Digital Engagement license. If they are not enough, you can purchase Additional Einstein Bot Conversations, which provides 100 additional monthly conversations for $50. Prices are current at the time of writing.

First of all, set up a regular chat without the Einstein bot. Go to Setup > Chat Settings > Enable Chat > Save to do this. The settings, which appear after activating the chat, leave as default.

Salesforce Einstein Chatbot

After that, go to Service Setup and select View All on the “Recommended Setup” component.

Salesforce einstein chatbot tutorial
Find Chat with Customers.
How to create einstein chatbot in Salesforce

Skip the first step, and enter the name of a queue and a group in the second. For example, Queue Name = Cases Queue, Name This Group = Support Group. Next, select the desired user from below.

Create a chat queue Salesforce Einstein chatbot

Now let's configure the workload of agents in the chat.

There are two parameters to configure:

  1. Work Item Size - describes the workload of a record per agent.

  2. Agent Capacity - describes the total workload of an agent.

In this case, one record will have 5 conventional load units, and the total will have 20 conventional units. Accordingly, up to 4 records can be assigned to one agent at the same time. Later, to change these parameters, go to Setup > Routing Configurations and change the Units of Capacity field in the required configuration to change the load of one record per agent. Also, Setup > Presence Configurations, and in the desired configuration, change the Capacity field to change the total workload per agent.

einstein chatbot configuration tutorial
Next, specify a link to the site where this chat will be used. If you have an existing site, you can include this link. The link to the Public Site is provided in the example. You can choose any available domain by searching through the options.
einstein chatbot on website

In the next step, choose the chat range of use: what records will be created when the user initiates the chat. Available options:

  • Creating Leads

  • Creation of Cases and Contacts

  • Creating Contacts

einstein chatbot for service

Do you need offline support? Let’s say you don't need that kind of functionality right now, so skip it. If necessary, it can be added later in Setup > Embedded Service Deployments > View > Chat Settings > Offline Support.

dont miss out iconDon't forget to check out: Salesforce Chatbot: Streamlining Customer Service and Support

einstein chatbot offline support feature

Salesforce will then start the setup process, which will take a few seconds.

Salesforce einstein chatbot step-by-step guide

As a result, you will get a piece of script that should be inserted on every web page of the site where you want to use the chat. It is recommended to insert this script before the HTML tag.

salesforce einstein chat code snippet

Copy the script and try to paste it on the site, and check the performance of the newly configured functionality. If you don't have your own site, you can create a Visualforce page with this code. The Public Site is used for the example, the created page is a Home Page. To do this, go to Setup > Sites > Edit. Click on the search icon near the Active Site Home Page, choosing the desired page. Then save the changes.

salesforce einstein liveagentsetupflow

Open a new tab and go to App Launcher. Select Service Console. On the Utility Bar component, click on Omni Channel > Available - Chat.

Einstein chatbot app launcher
Let's act as a client. Go to the website where you will see the Chat with Agent button in the lower right corner. Click on it and enter your data.
einstein chatbot client registration

From the agent support side, you can see that a Chat Transcript record has been created.

salesforce einstein chatbot chat transcript
On record you will have the opportunity to communicate with the client in real time.
einstein chatbot real-time

However, this approach has a number of critical points, in particular:

  • A large number of agents to process all requests on time;

  • Agents must work around the clock;

  • Low CSAT (Customer Satifaction Score) due to a long wait for problem resolution;

  • Some requests have the same type of problems.

It would be better to avoid such negative consequences. So, let's add a chatbot that will work 24/7, will be able to serve an infinite number of customers simultaneously, and will prepare answers for similar routine problems.

Preliminary preparation before implementing the bot requires only going to Setup > Einstein Bots > enable Einstein Bots.

Einstein bots

Let's create a new instance of the Einstein chatbot by clicking “Create a New Bot” below.

 

There are two types of bots: standard and enhanced. Let's not focus on their differences. If you are interested in this topic, you can read about it here: https://help.salesforce.com/s/articleView?id=sf.bots_service_enhanced.htm&type=5 

Currently, choose the standard one.

eistein standard chatbot
Choose the Intro Template and modify it for your own needs, instead of doing everything from scratch.
Einstein chatbot template

Name the bot and choose the language.

einstein chatbot personalization

The bot will specialize in customer support, so exclude unnecessary options.

salesforce einstein chatbot personalization

Let's skip this step; we'll talk about it a little later.

einstein chatbot deployment

At the end, click on the Proceed button to complete the settings.

einstein chatbot configuration example

If everything is configured correctly, then a green checkmark will appear near each item, and the Finish button will be available.

einstein chatbot deployment check-list

You will be redirected to the Einstein Bot Builder page. Familiarize yourself with the standard flow. To do this, select Map instead of Detail in the picklist.

Einstein bot builder page

The bot works according to the following principle:

einstein bot builder configuration

At first glance, the scheme may seem cumbersome and confusing but don’t worry because it is not. You will learn everything easily after you start working with it. Also, it is similar to the diagrams in Flow Builder, so if you have experience working with Flow, it will be a good advantage and will facilitate the process of exploring this diagram.

One step was skipped during the bot setup, and as promised, we're coming back to it. Let's connect the bot to the chat. To do this, go to the Overview tab.

einstein bot builder example

Find the Connections section and click Add.

Salesforce einstein chatbot builder

Select Connection = Chat, Deployment = the group name you initially specified. In this case, it's the Support Group. Leave Require Agent Online disabled.

In the Bot Response Delay section, you can configure the bot's delay before each message is sent, which simulates the typing speed.

There is an option to save conversation data. To get this option, you must enable Store Einstein Bots conversation data in the Log Conversation section. This feature comes in handy when you have problems with unexpected behavior, so we strongly recommend that you include it during development. However, it is worth remembering that this feature can store some sensitive data, so it is recommended, even required, to turn it off in a production organization.

Also, let's configure the user on whose behalf the transaction will be launched. To do this, click on the pencil icon in the Bot User section, and in the "Bot performs actions as" section, select Custom Chatbot User (Recommended) and find your user.

Following these steps, the bot should have the following settings:

einstein chatbot deployment

Finally, activate the bot in the upper right corner.

activate einstein chatbot
Perfect! The basic setup is complete. We suggest checking the bot's performance at this stage. Go to the resource (website or public site) and start a conversation.
einsitein chatbot test

Unlike regular chat, as a support agent, you won't see a new record in Omni-Channel because now the bot is actually taking the chat for us.

einstein chatbot record in omni-channel

The chat starts automatically, and at this stage, you have two options: “Report an issue” and “Check the status of an existing issue”.

einstein chatbot chat example

Let's act as a client who has encountered a certain problem that they want to report. To do this, select Report an issue and describe the problem.

dont miss out iconCheck out another amazing blog here by Sparkybit here: Need to Hire Salesforce Developers-in-House? 5 reasons Why it is a Bad Idea!

einstein chatbot report an issue

Specify data for feedback

einstein chatbot digital assistant

and the company name.

einstein chatbot digital assistant example

After that, you will see that the case was successfully created, and you also received a number to track the progress of your request processing.

einstein chatbot digital assistant case created

If you return to the main menu, select “Check the status of existing issue”, and specify the data so that the bot finds exactly our case,

Check the status of existing issue in digital assistant

then, as a result, you will receive the status of our case and an offer to add a comment to it.

einstein chatbot digital assistant case status

When the support service agent comes to the workplace, they will see the created case without their participation. The agent will solve the problem and call the client back or reply to the email. Also, the agent will be able to see the bot's correspondence with the client and related records that were created during the conversation.

client data in salesforce chatbot einstein

Let's imagine a situation where the client has a specific question for which they need to contact an agent instead of creating a case. To provide this opportunity to clients, go to Setup > Einstein Bots > select a bot > Main Menu > add “Transfer to Agent” to the menu.

einstein chatbot transfer agent

Save and activate the bot. It is important to understand that transferring to an agent is only available if the agent is online.

An opportunity to talk to an agent causes the problem of a bigger workload on the employee. We suggest setting up the Lightning Knowledge Base so that customers can first try to find a solution to their problem on their own, and if that doesn't solve the problem, then contact an agent.

Again, go to Service Setup > Recommended Setup > View All > Knowledge Setup.

einstein chatbot knowledge setup

Note that once you enable Lightning Knowledge, you cannot disable it.

Select a user who will have access to creating articles.

einstein chatbot lightning knowledge setup

There is an option to create categories for Knowledge Articles records, but now let’s skip this step.

einstein chatbot knowledge setup example

You need to fill a database with Knowledge Articles records and generate an artificial intelligence model that will search for relevant articles based on a question entered by the user. The minimum number for model generation is 10 records. To start the learning process, go to Setup > Einstein Bots > Smart Features > Get Started.

einstein bots smart features

After that, map the required fields and start the training process, which can last quite a significant period of time. The minimum time for training is 30 minutes.

einstein chatbot cofigure article answers

After completing the training, go to Overview in Einstein Bot Builder. In the Settings section, find Article Answers. Enable this option.

einstein bot builder overview

Return to the Dialogs tab. You will see that a new Article Answers dialog has appeared.

einstein chatbot article answers dialogue tab

Optionally, you can change the process that occurs when this dialogue is initiated. Now let's leave it unchanged.

After asking a question, the bot will trigger the search of the knowledge base, and the system will find the most relevant articles according to the given question. If there is no answer to the question in the database, in this case, the bot offers to try again. Also, you can customize the dialogues. For example, create a flow that will collect information about unsuccessful attempts to find an answer and store this data in SF so that in the future, you can improve the knowledge base and increase the level of customer satisfaction.

If you are interested in this topic, we suggest you continue an independent study with such topics as:

  • Enhanced Einstein Bots - improved chatbots;
  • Dialog Intents – for routing the client to different dialogs based on the entered text;
  • Multiple Languages Knowledge Base – to localize the Lightning Knowledge Base into different languages;
  • Multiple Languages Einstein Bot – for localization of bot dialogues in different languages;
  • Flow/Apex - to expand the capabilities of the bot;
  • Omni-Channel - a deeper exploration of this topic for effective distribution of chats between help desk agents;
  • Supervisor Configurations – to create a system of control over support service agents and help with solving customer issues.

To sum up

So, here is a summary of the work we done:

  1. Chats with the user are configured.

  2. Added a script to the Visualforce page.

  3. The page with the script has been moved to the Public Site.

  4. The bot is enabled and configured.

  5. Lightning Knowledge Base has been set up.

  6. Articles with frequently asked questions have been added.

  7. We trained an AI model to enable users to find answers to frequently asked questions independently.

Sparkybit, Salesforce consulting and development company. We make Salesforce perform at its best by fitting it to Clients’ custom business needs. 9 years of sharpened expertise, 25+ successful long-term projects globally.

[email protected]

Copyright belongs to Sparkybit GmbH, Im Dörener Feld 3, 33100, Paderborn, Germany.
2023 © Sparkybit. All rights reserved.

The post How to setup an Einstein Chatbot in Salesforce? appeared first on Forcetalks.


September 20, 2023 at 11:13PM
Click here for more details...

=============================
The original post is available in Forcetalks by Sparkybit
this post has been published as it is through automation. Automation script brings all the top bloggers post under a single umbrella.
The purpose of this blog, Follow the top Salesforce bloggers and collect all blogs in a single place through automation.
============================

Salesforce