女子做绝育手术后再次怀孕 状告医院索赔被驳回

Einstein Bots interact with your customers quickly and accurately with automation and artificial intelligence (AI) capability. In terms of AI initiatives, the biggest challenge is measuring business impact. One reason most AI projects fail is that people are looking at “model performance” instead of business value, such as how much money, in either additional revenue or cost-savings, can be created. Measuring value is part of an operational challenge when implementing a chatbot. So, let’s explore what you need to know to make Einstein Bots a success at your company.

What is Einstein Bots?

Einstein Bots is simply a chatbot which is an application that simulates human conversation via text message. Instead of speaking with a person, like a sales rep or support agent, a customer can have a conversation with the bot. Einstein Bots has the distinct advantage of connecting to your CRM that stores your customer data, and leveraging AI technology called Natural Language Processing (NLP) that allows it to understand the intent behind a conversation.

Learn what Einstein Bots can do

Here’s a list of common use cases for Einstein Bots.

• Answer FAQs
• Issue Reporting
• Information Management
• Scheduling
• Qualification, Triage, and Routing
• Customer Identification and Authentication
• Marketing and Lead Generation

As you can probably tell by now, Einstein Bots is highly customizable. You can tailor it to your company’s specific needs and build it to any industry standard. Here are some of the customer use cases our industry clients have adopted.

• Healthcare & Life Sciences: Appointment Scheduling, Check-in, and Provider Search
• Financial Services: Insurance Quotes, Credit Card Replacement, and Branch/ATM Location Finder
• Communications & Media: Contract Renewals, Up-Sell, Troubleshooting, and Service Availability
• Retail & Consumer Goods: Contactless Pickups and Personalized Recommendations
• Energy & Utilities: Next Best Offer, Service Activation, and Bill Updates & Payment
• Public Sector: Information Routing, Nonemergency Assistance, and Alert

There are three ways Einstein Bots performs an action to complete the preceding use cases.

A rule-based bot is the first approach Einstein Bots can leverage. A rule-based bot provides guidance as a menu, where a customer can select a path where they want to ask or resolve a question. It’s the most effective way to understand what the bot can do for a customer. You can consider putting a rule-based bot on your common inquiries.

The second approach is an NLP-based bot. Using NLP through intent management is uncharted territory for most admins. NLP bots allow the end user to enter text about what they want to ask first. Also, NLP bots can interpret multiple pieces of information within one phrase sent by the customer, which can reduce the number of steps required to complete a task.

You can have a hybrid bot as well, which uses rule-based methods and NLP methods at the right time to create a blended experience. Including a menu can help customers who need an idea of what the bot can do. Also, integrating NLP helps the bot sound smarter and more conversational.

What’s the typical timeline to roll out Einstein Bots?

It’s important to remember that building bots is iterative—starting off with a small bot and then building it into a complex machine is part of the journey. As a best practice, there are five steps to implementing bots.

1. Plan your bot

To start, implement your bot before configuring settings. Careful planning is essential to making your bot effective and your customers happy. You must consider a couple of plans from different perspectives.

• Business Planning
  • What are the common inquiries your customers have?
  • Determine whether or not a bot can resolve them.
  • Which business key performance indicator (KPI) do you want to achieve with a bot?
• Technical Planning
  • Select channels from Chat, SMS, Facebook Messenger, WhatsApp, and Slack.
  • Integrate with Knowledge articles or CRM.
  • Will you use a rule-based, NLP-based, or hybrid that leverages both rule and NLP capabilities?
• Tone and Manner Planning
  • What’s the name of your bot?
  • Think about how the bot will say a welcome greeting.
  • What is the tone and style of your bot?

When preparing business planning, typically, you have the following KPIs, which are the cost of operation savings due to Einstein Bot.

• Return on investment (ROI): Percentage return on investment for a bot
• Cost Savings: Reduction in cost of agent operations
• Deflection Rate: Conversations not transferred to a human agent
• Goal Completion Rate: Conversations that have completed a goal your customer wants to achieve
• Average Handle Time: Average time an agent spends on a case

After you’ve defined the business KPIs, the Einstein Readiness website will help you estimate ROI with some assumptions. On the website, you can easily estimate outcomes of how Einstein contributes to your business across other Salesforce AI products including Sales, Service, and Bots.

2. Build your bot

You can choose multiple options to build a bot based on the requirements outlined above. In accordance with your business purpose, you can also choose to build a bot from a prebuilt template, or Template Bot.
Here are five steps to build a bot.

• Create an Einstein Bot
• Understand Einstein Bot Dialogs
• How to Enable Natural Language Processing in Your Einstein Bot
• Use Intents to Understand Your Customers
• Improve Conversation Design

3. Test, test, test

After you’ve built your bot, small-scale testing is key to its success, since most AI projects must be iteratively improved step by step. Here are four aspects to testing a whole bot experience.

1. Testing in multiple channels, browsers, and devices
2. Testing the agent experience
3. Testing the natural conversation experience
4. Testing the customer experience

First of all, you must test your bot with multiple channels, browsers, and devices from a technical point of view. Using Events Log is helpful to view a history of all your bot conversations when you run into issues.

• Troubleshooting Your Bot

After you’ve confirmed your bot works technically, it’s time to deploy it in production where an agent can test it. Since a chatbot can handle a case without an agent, why test an agent experience? Because Einstein Bots can provide a seamless experience to transfer the conversation when your customer wants to talk with a human agent or another bot that can handle different scenarios. So, a chat agent can handle a case without asking the customer for their issue and information.

You must refine the conversation experience to see what it looks like as a customer. Natural conversation experience is strong, however, it’s tricky because a perfect and fluent natural conversation can’t be done without iteration. You must provide data, train a model, check model performance, test it as a customer, and iterate repeatedly.

• Understand Customer Intent with Natural Language Processing

Lastly, the most important thing is to test the customer experience by putting your bot in a place where your customer can interact with it. It’s not enough to test it one time; you must keep testing the customer experience repeatedly. Keep reading for information on how to do this.

4. Deploy

To launch your bot, perform the following steps.

• Connect Your Bot to Your Channels
• Preview Your Bot
• Activate Your Bot

After you’ve deployed one bot, it’s a starting point for a whole bot journey. Consider the following plans after deployment.

• Create an internal communications plan. Why is a chatbot needed? What is the business purpose?
• Create an external marketing message to show how your bot is helpful for your customer. How can we support a customer with a bot? What is a benefit for a customer?
• Plan a training bot strategy to make a better bot experience. What is a KPI? Who runs/owns this?
• Run baseline reports and create a performance track strategy on a regular basis. What is a baseline of KPI? How often do you track performance?

The last two items are needed iteratively. Let’s explore how you can do that with a useful tool.

5. Iterate and measure value

It’s important to keep tracking bot performance and adjust it to market and business changes in terms of business value. Use the Einstein Bots Value Dashboard to visualize and monitor bot performance and business KPIs, such as reduction in agent handling time, case deflection rate, and goal completion rate.

The value dashboard is very useful to track performance and how your bot brings business value over time. You can easily get the dashboard package through AppExchange (for free).

A customer in the travel industry got a result with the dashboard showing 198.15% ROI and 7,000 cases completed by Einstein Bots in 6 months. They are leveraging chatbot capability with both rule and NLP to handle an inquiry to check reservations, change an appointment, and handle common questions.

Other useful resources to improve bot performance:

• Evaluate How Well Your Bots Understand Your Customers
• Monitor, Analyze, and Refine Bot Activity
• Troubleshoot Your Bot
• Improve Your Bot
• Navigating Bot Conversations

The future is more efficient with Einstein Bots

AI and automation are efficient ways to improve the service experience and satisfaction for customers. But simply looking at this technology isn’t enough to get success in terms of business value on the whole chatbot journey. Now, you can grasp the importance of trying it, seeing how it works, measuring the value, and iterating it. Your chatbot journey is just starting!

For even more details, check out the Einstein Readiness website. And learn more about Einstein Bots on Salesforce Help.

Resources

• Salesforce Admins Blog: AI for Admins: What You Need to Know to Make Einstein Case Classification a Success