Salesforce, Inc. (CRM) Earnings Call Transcript & Summary
July 31, 2024
Earnings Call Speaker Segments
Kim Gledhill
executiveGood morning, everyone, and welcome to our Good to Great: 7 Digital Support Best Practices webinar. Let's get started. Firstly, I'd like to introduce myself. I'll be your facilitator for today's webinar. My name is Kim Gledhill. I'm part of our local Salesforce service cloud specialist team. Our team's focus is all about the customer experience and the agent and team members' experience, improving productivity and efficiencies and always keeping customers at the forefront of everything that we do. We have an exciting 30 minutes ahead of us. But before we kick off, let me just -- firstly, we would like to do our Welcome to Country. So we would like to begin with -- to acknowledge -- acknowledging our traditional owners of the land on which we meet today and recognizing their continuing connection to land, waters and culture. We pay our respects to Elders, past and present. Just a quick reminder that Salesforce is a publicly traded company. Our customers should base their purchasing decisions on products and services that are currently available. A little bit of housekeeping before we kick off this morning. So let me take you through our console. So we've got some great things that you can do on our console today. You can look at -- on the left-hand side, you can see and collect some of our resources that we have for you. We've got a demo there around the omnichannel solution. We've got some great tips from Matt, who is our speaker today, and his blog. So I encourage you to join that blog as well. And we also have our State of Service report and that is a great report for you guys to have a look at. [Operator Instructions] And also we have a great self-assessment tool there. The self-assessment tool is a great way for you to understand where you're at, at this point of time in your journey, and it will provide you with some great ideas to kick start your journey as well. So one of the things -- one of the questions that we get asked a lot is how do we create a great customer service program. And the Head of Customer Service is consistently challenged these days with doing more with less. But the teams need to continue to deliver that exceptional customer experience. This is where our 7 Digital Support Best Practices can help you start on that service excellence pathway. I'm going to hand over to Matt, who's going to take you through our 7 Digital Support Best Practices to help you transform your service experience. Over to you, Matt.
Matthew Kravitz
executiveOkay. And good evening. I'm actually in Los Angeles. So I'm on the other side of the world. And for once, I'm the one that has a funny accent on the call today. So hopefully, everybody can understand me. And I'm on our product team. Let me just really quickly introduce myself over on Service Cloud. And I have the coolest job in the world because I get to have conversations like this about -- centered around how do you get more value out of your Service Cloud investment and your relationship with Salesforce. And it's really cool to be on the product team doing that, and it's very much informed by a lot of the experience that I had actually before joining Salesforce. Prior to Salesforce, I was on the other end of these types of webinars, and I was a customer trying to figure out what to do next. So I spent 5 years running the Service Cloud program and the digital experience over at Disney streaming. And Disney streaming is sort of a consortia of Disney Plus and ESPN Plus and Star Plus and a company in Los Angeles called Hulu, which is previously where I started. And I learned a ton during that experience, and that is very much informed about like how do you get this technology to work. So I have somewhere around 22, 23 minutes, and I'm going to go through 7 best practices. This is a speed round. So let's try to get through this, and we have a few minutes for questions at the end as well. This slide is not terribly helpful. This is like the introduction slide. I'm not sure even why I created it because each of these 7 are going to have a section underneath it, but it's sort of like the table of contents that we all typically skip through. So let's get right into the first one. I had to pick 7 of my favorite best practices. This one I chose because it's pretty much the origin story of like why this Good to Great framework even got created in the 7 Best Practices because I realized that I had to redefine the definition of what maturity is. For many years in my career, the first 90% of my career until I started over at Hulu, I define maturity by like what technology I have. I have a but. Therefore, I mature. And December 5, 2022, I was in New York. I was working with the WeWork team who, I think, had done a nice job. They're using Experience Cloud. But I do this test. And the test that I do is like I take the third or fourth contact driver, and you'll see those yellow arrows pointing to a contact driver called like. What is your dog policy? Meaning, if you get office space through WeWork, can you bring your dog? And that's something obviously that as we look at our channel strategy, contact strategy, we want to deflect this one. Great, there's a knowledge article. And then I asked the chatbot, said, chatbot, can I bring my dog? And of course, what's the answer that we get? The answer is, I'm sorry, I don't understand. At that moment, I was like, oh, maturity is not technology. Maturity is the ability to use technology to serve and to resolve a use case. And in this example, it didn't give me the answer, right? And there's a lot of operational reasons why this happens and a lot of our bots, traditional dialogue bots, which is essentially what created this issue, have this problem. So my first lesson to really impart upon this group is like use your own technology, test it based on your top contact drivers, use your search component, start with Google, use your chatbot, but make sure your stuff works because having technology itself is not mature, it's got to just serve the use case. So that's our start. There's a lot of examples of this and a lot of themes or flavors of maturity. Another one is this example. This is another bot. So this bot is like an insurance claims that you should be able to go through and file a claim online. Unfortunately, what the bot said is, hey, here's a link at the bottom below. Please go through and fill out this form. Well, then what's the point of the bot? Bot should action. They should resolve things, not refer me to another page, not triage, but do something for me. And I thought that, yes, that's another perfect example of a lack of maturity. Speaking of a lack of maturity, that's a really -- it's this webinar framework, this really funny image that we had here and distorted to some weird things. So another example of maturity is this webinar platform potentially. Another flavor. It's like imagine you're going through and you're going into a pre-chatform, which a lot of us use. I think they're actually pretty good techniques, especially when your chatbot or your help site isn't authenticated. And I put in the contact driver here, which is, "are my batteries covered by insurance?" And then I even selected the category, which is a reimbursement or insurance. And of course, you'd predict like when I start chatting, just please don't ask me what you can help me with. And of course, that's what's happened. How can I help you today? Well, I just told you twice. So these are all flavors of maturity. Yes, we have pre-chatforms, we have bot, but if your stuff isn't creating a contextual automated experience, we've got some work to do. The second lesson is you have to realize at minimum 65% of the time. The journey actually starts at Google. So for any of you that are users of Salesforce and if anything goes wrong, the first thing you do, which is pretty crazy, is you go to Google, you're going to type in Salesforce customer support or Salesforce, I -- my permission set isn't working, right? You're going to do the contact driver. This is an example of what not to do, which is here's a phone number. So you shouldn't show your phone number in these Google results, okay? You also should show your Contact Us page, which, again, is just going to show you a phone number or an e-mail. So what should you show? You should show your home page of your help site. Bring us through the front door. So we could show you our knowledge articles, and we can give you a search component or our chatbot that does wonderful things, but you're not going to be able to deflect if you're showcasing your phone number. So Salesforce, we have a little bit of work to do. We're not perfect. So we'll fix that. So Google your own stuff at the end of this. Take out your phone, Google your own stuff. And search also takes another interesting flavor. Search has got to work on your help site. This is an example of a very large retail brand that is sharing their help property in with their dot-com site. This is extremely common in retail as an example. But that search component very rarely understands the language of support. So in this example, where the top contact drivers, track my order, and it's absolutely no idea what track my order even means. It's showing you track suits. So you got to test this stuff, test and experiment with Google, use your own search components because the journey really does very often start at search. Lesson number 3 is, think about channel strategy. All channels are not created equal. Sometimes voice is amazing. Sometimes chatbots are great. So the exercise that I'm encouraging you to do is in this very first column here, list out your top contact drivers, and if available, the percentage of volume in this very simple example, I'm showing you a retail company. The goal is to lock yourselves in a room with all of your peers and figure out for each of these top contact drivers. What is the best and the worst experience? In these examples, put a 1 where it's a terrific use of that channel and put a 7 or 6 obviously where you see, boy, I don't want people in this example. Look at the top row. If somebody wants to track their order, we don't want to showcase our voice or video channels because it's a poor use of our resources. We much prefer to have a website flow or knowledge articles or bots. So ultimately, and I created in this next slide, I'll show you in a moment, a type of a graph that you would use, and you'd adapt to, and that -- you put your contact drivers. This is just the sample, right? Put your contact drivers in this first thing, and you'll do the 1 through 6. But in this format in your own that you would change some of these values, right? So if you're not using Web-to-Case, you would change this over to SMS or if you are using WhatsApp, you would change it -- or Line. So certainly, this is a framework, but it's not meant to be set in stone. This is for you all to adopt for your businesses. And you will define your best and worst. And there's something pretty interesting that we could do with this channel strategy because now you can't just live in this Google Doc or wherever you're going to complete this. We need some way to institute this in our help site, and I'll show you something really cool that we can do. Number four, let's start with the basics of information architecture. And you sort of have to realize the analogy here is like you walk into a library. You have something that you want to do. For example, what you just saw in that last slide are jobs to be done. Like I want to check the status of my order. I need product information. Imagine if library didn't have a Dewey Decimal System, meaning it was hard to know the organizational structure of a library that analogy, that metaphor is very, very similar here, which is like you need some way to group your top contact drivers. And we call these Level 1, Level 2 information architectures. And what you see here in bold are examples of Level 1. Underneath of them are the Level 2s. Level 2 is like I need to log in. I need to manage my account. I need to check my claim status. So you've got to make sure the stuff makes sense to your customers. And when you define it, it's got to be consistent everywhere. I mean look at this example here. This is a Web-to-Case form where you choose why I'm contacting you and look out dramatically different it is from what was defined on this page. It says log-in assistance. This is 2 words. It's 2 words up here but It's 1 word in the chatbot. So the chatbot has a complete different Level 1 architecture, then the pre-chatform, then the help site homepage. Looking at this chatbot, it's like, oh, okay, find a dentist. Find a dentist. It has nothing to do with where I see these 2 areas here. So to me, you sort of -- you know behind the scenes. What's happening is there's organizational barriers that are happening that these 3 teams, essentially the chatbot team, the operations team that owns the pre-chatform and the help site team just didn't sit in room together. So we just don't want to expose organizational disparity into our customers. So this is an opportunity. Get everybody in a room. You guys figure out what's Level 1, Level 2, what are the top contact drivers and just make sure that you have continuity and provides just tremendous benefit. Number five is to really focus on knowledge. Knowledge has been sort of under the radar for the last couple of years, and we used to just take team matters and give them, "all right, you've been a great team manager. Now you could write knowledge articles." Well, now more than ever, it's important because of a lot of reasons. One is you'll see here on the left-hand side, is it's critical for search results, and it's also critical for generative AI. Here's what I mean. So what you find here, and this is an example of a Disney article using Experience Cloud and Salesforce knowledge. And down below, you see the snippet here that's actually really useful of what the best practice is. What you'll notice here in the Google Search is that has been literally showcased or copy and pasted within the Google result. How the heck do you do that? The answer is using meta description, meta tagging, is that you can decide within your knowledge article, what you want to show, what you not want to show, what you want to suppress. This is why we're encouraging all of our customers to get away from that sort of one large FAQ page because you can't do any of this stuff. You can't do the meta description and tagging. You can't do the suppression. And it's also not very conducive to what Google is in the middle of showcasing right now, which is Google's generative results within search. And it's really interesting to see how wildly inaccurate some of the Google search results are looking at your data because of the way that sometimes it's even finding another company's best practices are showing up in your search results. There are some really interesting stuff happening. You should be aware of it, and you should start reading a lot of the, About articles of Google's generative AI and how they're growing your knowledge articles in order to get results. So experiment with Google, and I'm also providing for you 2 of these links. These are 2 links. The first link is for those of you that don't have knowledge, some helpful best practices in the first. And then the second one, it's all about, if you have knowledge or if you're offering knowledge, what are the considerations that you should be using as we're now thinking about, wow, generative AI and how I'm going to be able to use bots that are going to pull from these knowledge articles in a generative rounded fashion or that my agents can have grounded replies based on these knowledge. So there are some tips and best practices there as well. So it's pretty shocking and the fact that we have to even put this slide in here. But for around 95% of B2C, B2C specifically help sites, they are not authenticated. This is an example of when I was at Hulu. We did take the time and effort to take the SSO, the authentication that the Hulu.com property had used. And you've got to authenticate your site. Why? You'll see it here on the left-hand side. Number 1 is it's the only way to get true resolution. If you're not authenticated, all you're doing is telling customers how to do something. This is how to reset your password, how to cancel, how to change your subscription. And when you're authenticated, you can be like, okay, well, I could do that for you. I can reset your password. I could change your product entitlement. I can do these things. I can help you troubleshoot and detect what's going on. So it's very important to get from something that's like, oh, go do it yourself. Way to evolve is to authenticate and to actually help solve things. So if you have an authenticated help site, your bot can also inherit that authentication. The bot could be like, oh, the help page that I'm on, that's authenticated to site. And with a little bit of work, probably 6 to 8 hours of development work, you can actually get the cookie from the help site and your chatbot can say, oh, Matt, welcome back. I know how to help you, right? I know what you're trying to do. And it may cost you. I'm giving you very round numbers, and these are in U.S. dollars. Let's assume it's $0.25, $0.50 for each authenticated session, right? There is some costs associated to using Experience Cloud authenticating, but you get tremendous ROI on it because when the chatbot picks up, oh, welcome back, Matt, oh, you're going to transfer over to an agent. It's also that session, who you are. We'll get screen pops to the agent. You're going to save all that time on -- the handle time of being like, oh, let me pull up your account. All that will be fed to the agent when the chatbot is transferring to them because of this self authenticated session. So literally, around 95% of B2C help sites, certainly an opportunity here to differentiate yourselves. And lastly, let's get to my favorite slide, is let's talk about brand-new generative bots. This is going to be the story over at Dreamforce, is all around authenticated bots. And instead of explaining what is a generative bot, I just want to talk about a very simple contact driver, something like what is the status of my order. I'm going to let you -- I'm going to propose to you like this maturity model, this framework of how to start engineering and thinking about your bots. This is what we mean by Level 1 bot. You ready? Look at the answer to the question. The question is, what's the status of my order? And the bot says to you in a Level 1 bot, just go to My Account and then the Orders page. And essentially, the bot is just saying, go do it yourself. Doesn't matter if you have a menu bot, a dialogue bot, and that's one of those bots that you type in, or even if you have this brand-new generative bot you're going to see at Dreamforce. It's going to give you the same answer, which is, go do it yourself and how to do it. Now a generative bot has a couple of advantages. Specifically, number one is that it's no longer going to say what you saw in that first slide. It's not going to say, I'm sorry, I don't understand. A generative bot, because it hasn't been programmed with dialogues, is actually going to be able to go through and search knowledge articles and give you the answer. So that's pretty cool. But again, you still have to do it yourself. And the other thing is for some of your companies, especially software and hardware companies, there's really rich data that's sitting outside of knowledge articles and unstructured sources and actually a generative bot can go through and search all that rich product documentation and give you some answers as well. So Level 1 is good, but it's not really impressing me that much. And neither is Level 2. And I just told you in tip #6 that you need to authenticate, look at how the answer changes now in Level 2. Now it says, welcome back, Matt. But you still get that crammy answer like, go to My Account, go to the Orders page, and it's still my burden to do it. But sure, welcome back, Matt, you're a gold member. Okay, that's cool. I'm not terribly impressed yet. Let's get to Level 3, and all of a sudden everything changes. Level 3 is that we can actually access transactional data. What do I mean by transactional data? For airlines, it means flights. For retail, it means your orders. For an insurance company, it means your claims. Any sort of scaled transactional data. If you can pump that into the data cloud, the answer changes. And for the very first time, look at this answer. Your order has shipped. Welcome back, Matt, you're gold member. Your order has shipped. Now we've reduced contact volume. Now we're resolved for the very first time. Now I'm impressed. Not a lot of bots are doing Level 3. And trust me, there's a million reasons why they don't. It's back in integration. We've never been able to pump in hundreds of thousands or millions of rows into an architecture like Salesforce. Things have changed. It's really exciting. Level 4 is you can actually action against it. Yes, your order has shipped, but now look what I could do. I could return the order. I could change the order. This is a Level 4 bot. You could do something for me, not just tell me what to do. Amen this is amazing. Now we're talking, but it's rare. And we need to see more of this stuff. Let me show you some examples of this stuff. I pick this one because I love it because this is Apple messaging for business. Bots are not just chatbots. When you're at Salesforce, if you build a bot, you can put it in an SMS spot, Apple messaging for business. I love it. It looks gorgeous. The problem is I want to return my online order and look what it says to me, head to our website to complete your return online. Means go do it yourself, it's a Level 1 bot. Beautiful, but that wasn't real helpful. This is a major retail brand not using Salesforce. So essentially, I just want to return my shoes online. And it says to me, you can return your shoes online, log in to Order History, blah, blah, blah, who's going to remember all this stuff? You're a retail company. You have a bot and I can't return my shoes online. Not impressed. Level 1. If you don't know this next story, please Google it. This is not to do. This is Klarna. So you'll end up Googling Klarna GenAI. You're going to see a ton of articles about a service leader that built his own bot and wrote all these articles. Look at me. I'm a hero. I built my own bot. Well, guess what he did? Laid off 60% of his workforce. First of all, I think that's a horrible example to show anybody that's working for you. This is the power of AI. We're now going to let you go. Not a fan. And when I use their bot to say, why did you let them go? First of all, I notice up here, there's an M. I authenticated into this bot, which is cool. I didn't see that in my last 2 examples. So I said to the bot, "Bot, when was my last order placed?" Here's the answer I got, "I'm sorry, Matthew, but as a chatbot, I don't have access to your specific order details." So congratulations, you knew my name. You're still Level 2 bot. How could you not have access to my transactional detail, which has to do with loans? This is a buy now, pay later service. So cautionary tails. So where is Salesforce heading? This is all where we're going to be going in our -- what we call our winter release. I guess it's your summer release. This is what's going on in the September, October frame, is we've got Level 3, Level 4 generative bots for you. Actually, this one on the right-hand side, I assume you guys have OpenTable, but it's a way you make dinner reservations, but we just released this -- by the time I created this slide a couple of days ago. The company name was confidential. It's now on LinkedIn and X. This is OpenTable. You could change the time of your dinner reservation. That is amazing. Not just when is my dinner reservation, which would be Level 3. That's what you're seeing on the left. But changing and updating. So think through a lot of your contact drivers and like, what are some of those contact drivers that we could get to Level 4 because it's pretty amazing. In closing, this is a super rich slide. This is like a screenshot slide or I assume we're going to get you a copy of this deck, but this is my handpicked greatest hits. We're always showing you a road map slide, which was like, okay, so here's what's coming next. This is what we've had for years. Some of the new stuff, but a lot of my favorite things are here. Please go through, Google these things, talk to your solution engineer, your account team, learn about them. But they're some of my greatest hits. And that's it. So I'd love to stop talking, so I can see if any questions came through. So Kim, I'm going to pass this back to you and let's see what the team came up with. That was a quick 20 minutes. I wish we had an hour, but hopefully, that helped.
Kim Gledhill
executiveGreat, and it was a quick 20-plus minutes. So one of the questions that I have -- that's come through is what will be the quickest win when it comes to all of those customer experiences?
Matthew Kravitz
executiveFantastic. What -- and I just ran out of time. Please go on to Paramount Plus's help site, okay? It's a non-authenticated help site. And go into their Contact Us page. What they have done is they've taken Experience Cloud in their Contact Us page, and they've taken that idea of channel strategy, which is what's the right channel at the right moment, and they built a flow driven Contact Us page to literally say, oh, this is why you're here. Here's the right channel. There's so many of our customers that say like, "oh, how do I get less phone calls about technical troubleshooting or resetting your password." Paramount Plus did it. So did Hulu. But it's authenticated. You may have access issues. But Paramount Plus has perfected this idea of like this progressive loading. It's a great example of using flow to perfect channel strategy. But you need to do check the section 3 first, but the way to operationalize it is through that Paramount Plus example.
Kim Gledhill
executiveWell, thanks, Matt. Look, we do have a couple of other questions, and we will get back to those to everybody. The session has been recorded. We would love to thank you for your time. We're right on time now. There's some great assets again for you to download here. I would encourage you to use a self-assessment tool. Please dive into the resources that we've shared. Once again, the session has been recorded, and it will be shared with you. And also, please feel free if you do want to start the journey, your service excellence program journey, please reach out to your local Salesforce team member, and we'd be happy to talk to you about how we can assist you with the 7 Best Digital Practices and the Good to Great framework. So again, thank you, Matt, for your time, and thanks, everybody, and have a great day. Or for you, Matt, have a great evening. And bye for now, everyone.
Matthew Kravitz
executiveThanks, all. Bye-bye.
This call discussed
For developers and AI pipelines
Programmatic access to Salesforce, Inc. earnings transcripts and 248,000+ others is available through the
EarningsCalls.dev REST API. Plans from $24.99/month — full transcripts, speaker segments,
full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.