Salesforce, Inc. (CRM) Earnings Call Transcript & Summary

Welcome to the webinar on how Salesforce sales leaders use GPT to accelerate growth. We are so excited to share with you the latest insights on GPT on the Salesforce platform and super excited about today's announcements. But first up, an exciting announcement from our legal team, a quick reminder that Salesforce is a publicly traded company, and customers should base purchasing decisions on products and services that are generally available. Now before we get into the content, I wanted to say a huge thank you to everyone joining us here today, our customers, our partners and those considering Salesforce. We know your time is valuable, and I appreciate that you're choosing to spend it with us. My name is Frank Fillmann. I'm the revenue leader and country manager. I am thrilled to welcome Ketan, and our EVP and GM, of Sales Cloud globally here in his 12th year in Salesforce. Good morning from 10,000 miles away, and thank you for joining us Ketan.

Thank you so much for having me, Frank, and a big hello from Palo Alto, California.

It's great to have you. And so for everyone dialing in, we've got such an exciting agenda for you today. We cannot wait to share with you how we've created the world's #1 AI CRM. Now there's a changing landscape around us. Everything has changed seemingly overnight. And we've seen the fastest tech adoption to 100 million users in history with GPT and it happened in weeks. And frankly, it's pretty hard to keep up with. Now the next piece of the agenda is the new business formula. How do we react and how do we respond to that? Well, expectations have changed. Consumer-grade tech and the enterprise just got a whole lot more critical. Your highest grade experience becomes your new expectation. Generative AI will fundamentally change the way that we engage with technology, it's going to impact every industry. It might even transform and disrupt some industries and companies. So how do we respond to this? It starts out with real-time data. Now it's tempting to jump right into GPT as a topic, but it's important to take a pause and evaluate your data strategy holistically. Data is going to be key and foundational to make sure that AI and GPT is successful. At Salesforce, we believe the right data strategy is going to lead to better decision-making and will power greater creativity and boost productivity. Now from a predictive and GAI standpoint, Ketan is going to take us through a conversation, a demo and then drive some Q&A. But before we get into that, I want to set the table. Now when you think about AI, there's so much more than that. It's about trust, it's about compliance, it's about safety. On the left, we've got consumers. We're impatient. We want all these capabilities now. But on the right, we've got companies. We want to understand the future. Maybe it's about driving growth or efficiency or getting into new markets. But naturally, there's a gap between the expectations and our ability to deliver. Customers are embracing generative AI faster than ever before. And as the world changes, customers are trying to keep up. Many customers today are looking for guidance as they develop and enhance their AI strategy. Today, we want to give you guidance on how we think about this and how we use AI here at Salesforce. So as you think about your AI strategy, you think about what's important, we want to show you how we believe we should bridge the gap. There's 4 things. So first is data. Good data in, good data out. The second is talent. We need to make AI super seamless and easy for our selling community. The third is regulations. Over 2/3 of IT leaders believe generative AI is going to introduce new security rest of companies and customer data. And finally, integrations, this is key to integrate AI and your customer workflow and into the data. AI is in a side show as part of the Main Act, and that's when we're at Salesforce. We want to introduce to you the new business formula. Right. So what does that mean? Sales GPT brings AI, data and CRM together to accelerate your growth. From a CRM standpoint, my team uses Salesforce as a single source of truth every day for visibility and action. From a data perspective, they're effective because they've got a unified harmonized data in a way to really connect customers and businesses together so they can act in real time based on visualization and insights. From an AI standpoint, it's really important. We've got trusted built-in AI that runs trusted customer data, and on the gold standard public sources to give them relevant personalized recommendations or assistance, helping them to work harder, faster, smarter. You can think about these 3 elements in a silo, but the reality is, it's really a formula working together into a continuous and enriching loop. So look, I've spent a lot of time with customers over the last few weeks. And I'm really seeing 3 things. One is intense curiosity, whether it's competition or disruption or just tinkering with the technology, everyone's really inquisitive about what does this mean? The second thing I'm seeing is how do customers look at accelerating their tech stack. Most firms have a very material tech debt. So how does it fit within their existing operations, their existing process, workflow and the tech stack, right? And the third thing that we see continuously is -- some customers just say, I just want to start using it. And I'm so excited that we now have generally available products that Ketan is going to talk about today that we're -- customers have an opportunity to actually get their hands on it. I like to think about AI like water. When you're thirsty, what do you want? You want 2 things. The first you want it when and where you want it. If you're running a marathon, and you need hydration, you don't want to go find it. You want someone to give you a cup of water at the point of need. The second thing is you need to trust that it's purified, distilled and safe to drink. Now we take this for granted, but it's critical. You're less interested in where it came from. Where did you pull it out of the ground? What's the filtration process? How did it get distributed to me? You don't care about that. You just expect for it to be safe to drink. When you think about generative AI. Don't you want to say 2 things. You want it in the moment in the flow of work when you want it, and it's got to be trusted. With that, Ketan, I'll turn it over to you to start the conversation around how we bring the new business formula to life.

Thank you, Frank, and a big hello to everybody. Good morning. Thank you for joining us. I'm Ketan Karkhanis, Executive Vice President and GM, for Sales Cloud. I was so grateful you could join us. And Frank, I love what you said is, you need it when you need it, and you need to trust it. That's such a great way of putting it. And that has been our focus. AI on the side is a side show. It needs to be woven into your business. It needs to be woven into the outcomes you are trying to optimize, and that's been our focus. But guess what I'm going to start with? I'm going to start with data. I'm going to start with data because it is the foundation. And you all know this intuitively. You don't need a slide to tell you that. You know this intuitively because you know today, what happens is buyers and seller connects everywhere, and there is data proliferation. It's no longer about a one-to-one interaction on a single channel. Buying is happening in an omnichannel scenario, buyers and sellers in the B2B are interacting on web, on e-mail, on video, on social, on phone calls, on events and so on and so forth. How do you get all that data into your CRM? You can't rely on manual updates. You can't rely on physically assembling disparate deal activities. And then you just have to not -- limited visibility is always a problem. That's why one of the biggest foundations we set was how do we get better data to drive faster sales. Step one in the AI journey is automated activity capture is ensuring your CRM is hydrated with great quality insights and data for every digital interaction, your buyers and your customers are having, and your sellers are having. So the first thing is Activity 360. It's about automated activity capture. It's about real-time deal insights. Imagine going to an opportunity on Sales Cloud and knowing everything that ever happened on that opportunity, that's kind of core. And we are not even starting to talk about AI yet. Imagine having that visibility into every deal and imagine doing that without relying on manual data entry, that is key, no more data entry. And that's what we are focused on because once you've got all this data going, then you can start getting into how do I now get trusted AI to generate those not just generative experiences, but also predictive experiences to drive that intelligence into your CRM workflow. And listen, when it comes to AI, you all know we've been at this for a long time back in 2014, we started our journey when we started our AI research lab. And then in 2016, all of you heard of Einstein, and we've been doing a lot of work democratizing predictive machine learning for the enterprise, for the B2B work for software and the journey has just been accelerating exponentially, whether it's our work around search, whether it's our work around conversational AI or more recently, all the announcements you are seeing around GPT and taking you into the generative experience world. Before I go into -- deeper into the generative world, it's really important to step back and understand that sales AI has never been more compelling. It's -- we're going to spend all our time talking about generative and GPT. But it's important to note it's predictive plus generative. And let me kind of break it down to you in a simple form. Now you all have heard about predictive analytics and predictive ML. You probably have heard about LEED score. These are ways in which you make predictions based on data that is in your CRM. So for example, you have a long list of LEEDs, your VDR, the SDR, inside sellers are working, how do you know which LEED to work on first. Well, you need LEED scoring to give you a ranking of high propensity leads so that you're not spending time with tire kickers, you're spending time with people who are going to have a higher probability to close. But predictive stops there. Predictive tells you, "Hey, this is a LEED with a high propensity to close." Generator picks up the baton from there and says, okay, now let me help you write a beautiful outreach e-mail to that particular lead so that you can work the lead faster. And you can extend this -- it's almost like a tango, predictive and generative working hand in hand, and I'll show you some of this in my demo, but that is a very important point. As you think about AI, you want to think about your complete AI platform. You don't want to think only about generative experiences, your predictive experiences here, and that's what our focus has been all along. So if you look at all the AI that is available today, and if you all are -- if some of you on the call are customers and you are on Sales Cloud Unlimited, guess what? You have this today. You may not know it. You may not have used it, but you have this today. Let me orient you around a couple of things. One is LEED scoring we already talked about. But in the prospect flow, we just released a buyer assistant. All of you have heard of web to LEED forms and you have on your digital experiences, wherein you're capturing needs, that then get worked by a sales organization. That's really monolithic, that's very static. Nobody likes to fill in a form anymore. But what if you could beat every visitor with a one-on-one personalized conversational experience, which is guiding them, giving them the information daily, giving them insights, and then when the time is right, when they have the information, connecting them to one of your sellers. That's buyer assistant. We've got conversational AI line. Imagine if Einstein could join you as an assistant on every meeting you're having. And imagine Einstein taking notes, transcribing the call, developing insight, generating sentiment detection, giving you that coaching while you're on the call. All these capabilities are available today, including the automated data capture, which I talked about. But that's like immediate. More AI is coming, even more innovation. In fact, this slide is a little outdated from 2 days ago, Frank, because the first one, sales e-mail is already GA, like this should have been moved to the previous slide. I apologize for that to my audience, but generative e-mail experiences is already GA. All of you pick up this is, on an average, how many e-mails do you think your sellers are sending every day, especially if they are inside sales roles, especially if they are BDRs, SDRs, they are pretty much sending 50, 70, 100 e-mails, maybe every day. What if you could save an hour a day and now multiply that across your organization. See AI is directly linked to productivity and ROI. The AI conversation is an ROI conversation. And sales generative e-mail is GA today. Buyer assistant also came out last week and beta. And then we're working on an incredible new technology sales assistant, that's going to be coming out, you all tune into Dreamforce, better still try and make it here. We would love to have you here for Dreamforce. But what you're going to see is giving every seller that assistant they need to take care of all the work which is the mundane administrative work. For example, I'm a seller, I inherited an account. I want to research that account. I want to find out how much business we've done with them. I want to find out data about that account. I want to find out closest connections to that account. I want to learn more about those people. I want to update my CRM. I want to update the LEED. Imagine all those things, I want to build a report, I want to generate new LEEDs. All those things are coming in our sales assistant. And then finally, call summarization. I told you about Einstein conversational insights, but imagine and transcript is one thing. But imagine, Einstein summarizing that call in 5 lines for you, and auto updating in CRM. Every rep wants their calls to be summarized and auto updated in CRM, so all these capabilities are coming. In fact, the first one is already here. So Frank, maybe I should just show a demo. Are you ready for one.

While you're getting set up, just 2 things. One is it's funny you mentioned, hey, this slide is 2 days out dated. That's like the whole theme of GPT. There's things going so fast -- it's like it's hard to keep up with. But I just saw a stat from a study that Salesforce put out here in the first half, it says the average account exec is spending 72% of their time not selling. And it just resonated with everything you're saying, which is, how do you get more time to actually go spend selling, which is freeing up the admin piece? So everything you're saying really resonates with what my team is telling me.

Perfect. Just to confirm, you all see my screen okay. It's a live sales cloud product. Frank, you see this good.

Good to go.

All right. Great. So I'm going to walk you through a couple of small demos. The idea is not to like take all the time, but to give you a flavor. This is a live core demo. So it's going to be fun always doing live demos. But we'll leave a lot of time for Q&A. So now it's a good time for you to start prepping your questions, which you want, Frank and I to answer. And everything I'm showing you is used by sales teams in Salesforce today. So we have already -- all our sales teams are using this. So let's start with something very simpler. I'm a seller. I'm logged into Sales Cloud, the #1 core SFA. You can see this prioritized to-do list. This is available today for everybody. And it kind of is auto surfacing to me all the actions that I need to perform based on various cadences, based on various sales stages, based on various engagement activities and so on and so forth. Looks like I need to send an e-mail to Hanna. Right here, I can click on this e-mail button. Nothing's changed. I'm in the flow of work, right, Frank? I'm running that marathon and now I need water served to me to take your example. I'm not going to leave Salesforce. Let me just make the screen bigger. And now I could start typing the e-mail, which is maybe -- it's going to be very antiquated in a few years' time? Or GPT is right here with me. Now in this case, what I'm showing you is a prompted experience. Now this is not a template experience. This is a prompted experience. So at this point, what we are doing is invoking the LLM but grounding it based on CRM context. So who [indiscernible]? What have they purchased? What products they have? Where are they in the sales journey? Or whatever all that. In this case, it's a LEED. So I just want to have a simple introductory e-mail. And Einstein is right now working hard, this is live, going to the LLM. As you can see, picking all the relevant information, all the product information and typing up the e-mail. Of course, at this point, just to be sure, it's not all done for you because Einstein is not connected to your calendar, so doesn't know what in the future, maybe you could, but doesn't know what date and time to project out for your meeting, but very quickly, this is where you can then quickly say, "Okay, why don't you just send them a scheduling link? Normally takes 6 to 7 e-mails to schedule a link, and now this is all. So this is an example of how accessible AI is now no longer for the few, AIs for everyone. And it's almost like you don't even know this is AI, is baked into the workflow you are in. Let's keep the demo going, let's say, I am on this lead here and I want to pull up Dan. Now you can see our sales engagement, and I'll draw your attention to the right out here. Now you're seeing predictive plus generative. Remember that slide I showed you, predictive plus generative, working as a team to give you that next step. So first, predictive is telling me, hey, the LEED score is high, maybe you should reach out, and if I send e-mail, it's the same experience, and I can maybe nudge them for a follow-up. Now these prompts you see, you might be wondering, Ketan, there are only 3 come on. I have to communicate with my customers on various other topics. You're absolutely right. That's why what we are also building in the prom studio, which allows you to take the proms we are shipping out of the box, customize them, change them with clicks, not code. You don't need to learn Python or data science or machine learning, it's clicks, not code, using our prom studio and add your own proms or change the proms we have done. One of the other things we have done is when we -- when we -- these proms are also context-sensitive. That doesn't mean if I'm sending an e-mail here, I see the same prom. And if I say somewhere else, I see the same prom. They are contact sensitive. In this case, I want to nudge for a follow-up. So this e-mail is kind of generated here. But let's say, I'm on a contact and let's pick up this contact. Okay, great. And I want to send an e-mail. Again, the Einstein button is right here with me, but now pay attention. Now I'm seeing new proms, because they are a contact, maybe you want to reconnect with them, and so I can type the reconnect prom and now that e-mail is getting generated. Hopefully, you understand 2 things. Number one is the CRM context. Number 2 is it's baked in the workflow. And number 3 is, it's customizable. You can make it fit like a glove to your business, your organization. But let's keep the demo going out, but a couple of more things for you to show. You might be wondering, Ketan, that's great, like Frank, probably will tell me, but Frank's team, they are not only in Sales Cloud. As Salesforce, we are a Gmail user, we use Gmail for our e-mail. So here I am in Gmail, and Frank's team has extension, you can have it too. This is my test account. This is not my real e-mail account, as you can see. But I want to send an e-mail to Hanna and I want you to pay attention to the right here. What happens at this point is the extension is connecting to Sales Cloud, is the same durative experience is now following me wherever I go and typing up that e-mail for me in Gmail. So no matter where I'm doing my outreach, sales GPT follows you. Now you must be wondering, Ketan, that's great, but we don't use Gmail. We are an outlook, Microsoft user when it comes to e-mail. That's fantastic, the same experience as you may have rightfully [indiscernible] in Gmail too. So let's go here and say I want to send an e-mail to Hanna, right here is the sales cloud extension that types that [indiscernible] up. And this extension is not just for e-mail. Hopefully, you all have seen are catching up on it's like the entire Sales Cloud experience is bottled up in a very simple intuitive side panel, and it follows you wherever you go. This is a game changer for adoption because now no longer are you telling your user, "Hey, go log into Sales Cloud." Sales Cloud is coming to you, wherever you go. And in this case, well, let's maybe send a meeting invite. And as you can see, it's the same experience. I can also schedule links. I can turn on e-mail tracking, I can say schedule for later or I can look up all the information around this particular contact and I can pull up their record without ever leaving the experience I am in. So this is kind of showing you the generative e-mail experience that just went GA 2 days ago. We already have customers who are using this. We're very excited with the uptake we are seeing. But let me just close out on one more quick thing for you. Our sellers are not only doing outreach. Frank, you and your team, you're on a lot of sales calls and a lot of your sales calls are virtual, right? Not all of them are in person anymore. So Einstein is joining with you on your calls, and let's pull up a phone call, a video call. Einstein is like your assistant, is on the call. The entire call is recorded, of course, with your permission and with your customers' permission. But let me draw your attention to something more exciting. When you're on a call with your customer, you want -- you want to focus on your customer. You don't want to spend time taking notes, figuring out, sometimes what is spoken between the lines, not missing out on action items and then imagine updating all of that in your Sales Cloud, that's a lot of work. You probably spent 30 minutes after the call to work on the action items from the call. But what if Einstein was there as your buddy, as your mate on the call, diligently, typing up all the transcripts. Einstein generated this transcript based on the call and as you can see, it's not a dump transcript. It is actually sorted where it is. So if you want to review your call, you can go to a specific area and just go and listen to that piece of the video versus trying to forward and back and forward and back, which is great. But anybody can do transcription, look at the left. It's generating insights on what happened in this call. And then these insights are reportable, whether -- and you can set -- that are out of the box in sites, obviously. But again, you can be like my organization, this particular world is very important to us. We want to ensure our sellers are saying this. You can have that as an insight. But here additional pool thing. It auto captures all the action items from your call and it updates you saying, you know how when we are talking on the phone frank, we say, let me take that as an action item or I will follow up with you on that. Einstein picks up on those signals and creates follow-ups and action items for you. One of them, of course, is send e-mail. And then again, that handy little generated button follows you wherever you go. So I could keep doing demos all day, Frank, but hopefully, that gave us a little bit of flavor on what's available today for everybody to use. Hopefully, you're seeing my screen again, but let's continue. How do you get ready? This is our last slide, and then we are going to open up for Q&A, Frank, I'm hoping you're ready for Q&A. All the questions are going to go to you, Frank, which is going to be fun. And number one is accelerating -- it's late in the evening for me. I'm allowed 1 joke. And I appreciate the latitude. I really appreciate the latitude, but look, number one, I mean it when I say it. You can accelerate your data and AI foundation today. This is not a -- this all technology is not something that's going to come 2 years from now. Ladies and gentlemen, it's here today, and you can accelerate that journey in a simplified fashion, talk to your account executive, ask them about Sales Cloud UE. If you're already on UE, unlimited edition Sales Cloud, great. You can work with our Premier Success team because it's not about more technology, it's about more success and work with our Premier Success team to build your AI plan. And then ask the team, ask Frank and his team about GPT e-mails because that's also GA today. So with that, I think I'll go to pause here because we want to be sure we left a lot of time for Q&A. Frank, should we open it up?

Yes. I've got a quick 1 for you as we transition to Q&A. You think about having a human in the loop, right, that the systems are serving up data and the question is, well, what's it coming back with? And is it going to be accurate for my need at the time or for what my customers ask? or I didn't keep the human in loop just to make sure that we safeguard and we're complementing digital and human together, right?

That's such an important point. That's such an important point. If you saw in the entire demonstration, GPT didn't send the e-mail. GPT did not do the Einstein conversation follow-ups. It's like it helped you do the job and kept it ready for you for you to say, thumbs up. This is great or no, I need some changes. That's what we call the fundamental design principle for us at Salesforce in our AI work is human in the loop. We believe having the human in the loop is extremely critical for a trusted AI interface. And that's why you see all the technology we build is always going to have human in the loop. It's okay to speculate someday about AGI and the future 5 years, 10 years from now. But that would be a separate webinar you and I need to host, Frank.

Yes. Listen, there's one question that I just saw it come in, and it's so interesting. I think everyone wonders about these questions, but like over 100 million of us go to ChatGPT and start interacting. And then -- and now every company on average is almost 1,000 applications, right? So like there's a consumer grade and enterprise grade. So part of it's like what's the different experience, and I think you demoed and helped demystify that a little bit. But the other part is like what's the difference in safety and trust? Like do we remember what people are searching for. Does that live somewhere else? Can you take the scary out of the conversation for those of us who are curious.

That's a great question. Let's do it step by step. Number one, if you remember the slide, Frank showed you, the trusted business formula, CRM. CRM is where you have all your data, and that's what we call is within the trust boundary of your organization. We actually use the word trust boundary because you have it compliant, you have the regulation, you have the governance, you got to change management. You have all of that. It's your trusted Navan CRM platform. What we do is when we do generative experience and, let's say, to prompt the e-mail, we are sending the customer name to the LLM. At that time, one of the big innovations we have done is the Einstein GPT trust layer. Now this trust layer is like a sorry to use a super simple nontechnical word, it's like a firewall in between. What it does is things such as PII masking, data masking, anonymization and we also did a killer innovation, which we call 0 retention. What 0 retention means is your data is not our product. Your data is your data. It is within the trusted boundary of your data and it's not getting used to train the LLM or it's not leaking into the LLM for future training purposes. So I think so the trust element is really critical, and that's why you saw me show the demo within the context of CRM because it's baked into the trust boundary of your sales cloud, of your service cloud, of your marketing cloud and so on and so forth.

Here's a follow-up question to that one. It says, we're -- how is the model trained for Einstein GPT? And where is the data actually stored for training? Can you help address that one?

Definitely. One, there is more than 1 model. A key part of our strategy is we believe strongly in an ecosystem way for an LLM GPT model ecosystem. So we have built our own models. You probably have seen Mark a tweet about our GPT models and how they're performing, much better than anything else out there. So we have our own models that are trained in our labs with our data. We also leverage outside models like open AI or Cohere or Anthropic. These are what we call public hosted models. And then, of course, we allow customers to bring in your own model. I think to bring your own model, we are seeing a lot of customers who are making investments in bringing their own model. So we got to do that. Now in terms of training the model, if you look at a public LLM like Open AI, it's trained on publicly available data. It's not trained on CRM. If you look at our core Gen LLM, which is our core gen, which is used for Apex development, that is used on our APEX data. So the answer depends on what it is. But no matter what the answer is, the Einstein GPT trust layer is always there between you and the model. And that's what we focus our energies on. So I guess we're having a lot of more questions, Frank, from -- I don't know if Cat is going to walk us through those?

I was just letting you go, I mean, that was great. And I can -- a lot of things are coming up in our questions, so totally fine. Just a quick reminder [indiscernible] this webinar will be sent to the e-mail that [indiscernible] afterwards. So please, there will be a recording, we will be able to share it [indiscernible]. And if you haven't found it yet, the Q&A box is at the bottom of your screen. So one question that you guys haven't got to yet, and it's not super technical. But how do I get started with GTP? How do I make this start to happen inside my sales assistant. Frank, that might go to you.

I like this question. This is the Ketan answer. Call your team, right? You've got a dedicated team who's focused on your success. It's part of the subscription service here and Salesforce to success. The second you buy something. That's our job is to make you successful as an organization. So reach out to your account team and reach out to your Salesforce contact and ask these questions. You're on a journey. This technology is changing rapidly. We're first to market with a lot of really interesting capabilities. I was wondering, Ketan, does your team sleep or not? It doesn't seem like they do the last 4 months. But they're coming out with technology and releases and GA so quickly embedding it into our entire portfolio of applications, but the quick answer is call your team, ask the questions, express any concerns you have, and let us know how we can help you on this journey.