What is a Bot? with George Hu

Today on the Salesforce Admins Podcast we’ve got George Hu, Principal Product Designer of User Experience at Salesforce. This is the first episode of our new series, All About Bots, which will explore how we can use them on the Salesforce platform.

Join us as we talk about the difference between bots and AI, how to train a bot, and why bots can make both your agents and customers happier.

You should subscribe for the full episode, but here are a few takeaways from our conversation with George Hu.

The basics of bots.

“There are a lot of misconceptions about what bots are in the market today. Bots are not AI,” George says, “we need to train the bot and script it to make it learn.” Humans teach bots what to do and say, so it’s more useful to think of it as a form of automation. AI comes into it because a bot can use that to recognize what people are saying and figure out how to respond, based on what you’ve scripted.

A common scenario you’ve probably run into is a chatbot for customer service. “Bots can be the first line of engagement for your customers,” George says, “they’re there to great you, identify as a bot, answer basic questions, and hand you off to human if they can’t help you further.”

Bots and you.

OK, so we’ve seen bots on the Dreamforce stage, but what does that mean for the average Admin? For George, it’s the new technology that is going to make a big impact. “The benefit is both to the consumer and the agent. On the consumer side, it’s a terrible experience to get into chat and get a number and wait instead of maybe being able to help yourself,” he says. “On the other side, the agent gets more time— they don’t have to do the tedious or repetitive tasks and instead can spend more time with the customer.”

If you’re looking to try a bot in your org, the first thing you need to do is think about what your customer experience is like today. How do most customers reach out to you? What resources do they have to help themselves? “Put yourself in the shoes of the customer, think through their experience today, and see if bots can potentially improve that experience,” George says.

Building your own bot.

If you’re getting started making your own bot, the first place to start is with your case history to understand what issues your customers most commonly ask you about. Once you understand the top three or four issues, you can try to automate them with a bot. “The more complex those issues are, the harder they’re going to be to automate,” George says, so it’s important to have a good understanding of how automation could play a role.

For Adidas, the top issues that they solved with bots were returns, order status, and latest product information. These kinds of problems are great to solve with automation, provided you can get the bot the data it needs in a place it can access. “Building a bot is kind of like watching a child grow up,” George says, “you’re going to need to learn about what the bot will say to your customers and what the customers will say back. You’re going to have to do a lot of iteration.”

The best advice that George has is to get the bot out in the wild as quickly as possible. While you can do a lot of planning, it’s hard to predict what will happen when that X-factor of a live human being on the other end is involved. Once you get some data you can make tweaks to make your bot even more effective.

Resources

• Gear Up for Chatbots: https://admin.salesforce.com/gear-up-for-chatbots
• Einstein Bot Basics module on Trailhead: https://trailhead.salesforce.com/trails/service_cloud_move/modules/service_bots_basics
• Design Framework for Chatbots: https://chatbotsmagazine.com/design-framework-for-chatbots-aa27060c4ea3?gi=35d65bef0e10

Twitter

• George: @george_hu
• Gillian: @gilliankbruce
• Salesforce Admins: @SalesforceAdmns

We want to remind you that if you love what you hear, or even if you don’t head on over to iTunes and give us a review. It’s super easy to do, and it really helps more Admins find the podcast. Plus, we would really appreciate it.