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

Hello, everyone, and welcome to today's session, Delivering Real-Time Customer Support with Data Cloud. And thank you all so much for joining us today. My name is Ariana, and I'm on the marketing team here at Salesforce. And before we begin, I'd like to cover a few quick notes with you about our webinar platform. Today's webinar will be available on demand after we wrap up, and it will be accessible through the URL that you are on right now. It will also be e-mailed out to you tomorrow. Please note the slides will advance automatically throughout this presentation, and you can enlarge the slides or any other widget on your screen by dragging the bottom right-hand corner. Should you need technical assistance, click on the help widget located on the bottom left corner of your console. We've also added a few additional resources, which are available through the resources window to the right of the slides. There you can find some additional related content. And lastly, we encourage you to submit your questions at any time throughout our presentation today using the Ask A Question widget at the bottom of your console. We'll do our best to answer as many questions as we can at the end of the presentation. And with that, I am turning things over to Kuber to get us started.

Thank you, Ariana, and thank you, everyone, for joining us today. I'm really excited to be talking to you today about how to deliver real-time customer support with the power of Data Cloud and, not just Data Cloud, the entire power of Salesforce behind you. I know there's a lot going on, so thank you for those who are joining us live. [indiscernible] watching a recorded view. Thank you again for that. All right, let's get started. At Salesforce, we are very proud of all the technologies we are always investing in, building, creating. Though we will be talking about a lot of forward-looking things today as well, we do want to mention that please do look at the details of the actual product before you make any purchase decisions. Now with that out of the way, let's talk a little bit about your team today. We have with you an incredibly amazing panel of experts and me to talk about real-time data. We have with us Vishal Chordia. He's a Director of Product Management. We have Nikhil Agarwal, who's also a Director of Product Management. Both of them are very close colleagues of mine, and we all, combinedly, work together on Data Cloud's real-time capabilities. They'll be joining us later in the session. For now, it's me, Kuber Sharma. I'm Director of Product Marketing at Salesforce, focused [indiscernible]. So let's talk about data. [indiscernible] data is not really a differentiator. Now I've been a data professional as an analyst, as a product manager, as a consultant, as a marketing manager for almost 2 decades. And the biggest thing I've seen in my 20 years in this business is data is no longer a big deal. It's no longer a differentiator. We all know that data and insights, they help companies differentiate themselves and also outperform the competition. But what happens when everybody is data-driven? Then it's difficult to differentiate yourself. And it's not just about that. It's also the fact that our customers, they are providing us more data than ever before, but in return they are asking for so much more in return, because customers are also used to being delighted by incredible data experiences. So as we get more data, it makes job of enterprises like ourselves more difficult because of the higher complexities involved and higher expectations. We can easily address today's business needs with real-time data. As customers ask for more expectations, we can actually serve them with personalized customer engagement powered by real-time data. Operational efficiency is a very increasing focus for a lot of organizations, and we believe real-time data can help with operational insight acceleration. Then, there is the need for agility, need to make things fast. Founder mode is now default really. And with that, real-time data is on hand to help with instant adaptability. So as I said, data is no longer a differentiator; real-time data, that's where the [indiscernible] is. But doing real-time data is never easy. Thankfully, we have a new product from Salesforce, the Salesforce Data Cloud, which will help you do just that. What is Data Cloud? It's a hyperscale data engine native to Salesforce. It will help you make sense of all your data easily and at scale. So it creates a trusted data layer which will power Agentforce agents across your platform, whatever Salesforce product you might be using, like Service Cloud, [ Sales Cloud ]. It helps you create that canonical Customer 360 view by bringing together all kinds of data. And then what is data unless you are using it in decision-making and activating it in near-real time? And that's what [ Data Cloud ] [indiscernible]. Now Data Cloud is a pretty complex product, but let me talk a little bit about how it works. If I have to oversimplify it, there are 3 things at Data Cloud helps Data Cloud do the magic that it is. Number one, connect any data from anywhere at any scale. Our friends at Apple like to say, "If you have a problem, there's an app for that." Riffing on that, if you have a data source, we have a connector for that. With more than 250 native connectors now, Data Cloud helps capture any data that you have live or at batch easily and make sense of it. And then it brings together all this data. It harmonizes the data. It preps, transforms and resolve identities so that data from disparate sources can become unified. But we don't stop at just the unification part. Next process is the real magic happens, where we act. With the power of Agentforce, we activate this data inside your applications natively for all Salesforce applications and also through APIs and integrations in other third-party applications. I'll talk a little bit about how this works. With Salesforce Data Cloud, we harmonize by the power of metadata framework for a unified view. So your external data sources, they may be structured or semi-structured, like online engagement data, billing data, subscription data that comes in. Then we start enriching this and combining it with the Salesforce data that already exists. So depending on what your use cases are, depending on what your installation are, you may be using Sales Cloud, Service Cloud, Marketing Cloud or Commerce Cloud, we are able to enrich this data with the external data sources that are structured and semi-structured. But thanks to our latest innovations and the recent acquisition of Zoomin, we can also bring in data from external data, unstructured sources such as PDFs, knowledge graphs, financial reports, customer conversation, chat descriptions, et cetera. So this helps you create a unique view of your customers, something that has never been easy in the past. But the great thing about this is this identity and the way it's being [ resoluted ] is not just about batch data. Now, for the first time in the history of data, you can also do this in sub-second real time. I'm excited to talk to you guys about a relatively new feature from Salesforce called Sub-Sec E2E Real-time. It helps you maximize the value of all your real-time data. You don't have to create complex pipelines to make sense of your data. With the power of this feature, you can ingest data in real time, under 500 milliseconds. Then you can take this data and enrich the identities that you already may have or create new identities. Imagine resolving identities in real time. We don't just stop there. As I said earlier, the magic is when you activate data at scale. And generating these insights from this real-time data, segmenting it, bringing out insights and then activating it back into Marketing Cloud-like experiences, that's where Data Cloud is truly differentiated, and nobody else in the market can do that today. Let's talk a little bit about how these applications have a huge impact. Let's talk a little bit about what the impact for a service agent may have. Three common things that we've seen already from our customers who are using these products. Number one, we can easily, with real-time data, boost a service agent's productivity. By embedding real-time insights for frontline servicepeople, we can speed up resolution time with the power of AI and a unified view of each customer. This helps you drive efficiency. Efficiency is part of your success metric. Maybe a second success metric is around growing revenue, and this is where real-time data will really help your frontline people. They can get near real-time recommendation on cross-selling, on upselling, so that they can surface these opportunities with the big customer profile being enriched in real time. And finally, a sticky customer is a happy customer, and a happy customer is a growing customer. So we can also help, with the power of real-time data, you can do proactive service. You anticipate any issues your customer may have and reduce their cost, potentially your cost as well, and thereby increasing customer satisfaction. This leads to reduction in cost from data, from your products, websites and customer conversations. Now I talked a lot about what is real-time data, how Salesforce [indiscernible] value of real-time data and what are some of the obvious impact to your business. But this is easy to say. Now I would love to invite my colleague, Nikhil, to actually show in action a live real demo to showcase what is Data Cloud real-time data in action. Over to you, Nikhil.

Hey guys, this is Nikhil Agarwal, and I'm going to walk you through a real-time Data Cloud demo and then also the setup of this in Salesforce Data Cloud. So right now, I'm at Northern Trail Outfitters website as an end user. And what you see is some generic content which is being shown to me, right? Now this is because the website doesn't know anything about me, any information about me, and so it doesn't really know what to show. Now as soon as I start interacting with this website -- let's say I go into men's, and there I go into, let's say, rainwear -- what's happening behind the scenes is that this data is flowing into Data Cloud, and it's actually capturing this activity as an anonymous user for me and trying to build a profile and trying to determine what I like. So if I click on this rain jacket here, it understands that I like rain jackets. And instantly, it will create an anonymous profile where, for me, the unknown user, this interest of rain jackets will be locked. And so now if I go back to the homepage, you see that the homepage banner has now changed to rain jackets. And this is how unknown user personalization happens as the user is browsing, as real-time Data Cloud captures and processes this information in real time and acts on it in milliseconds. Now let's say, as a user, I sign in. So I go to the sign-in page and log in. Now as I logged in, right, Salesforce Data Cloud merged this unknown user's cookie ID and associated it with Felix, the known user, who I signed in as, and so now when I go back to the homepage, what it's going to do is instantly identify that, hey, this is Felix, who has a lot of other interests. He's a silver member, he likes outdoors, et cetera. And so it's going to change that rain jacket banner and give me a recommendation which suits my needs based on my rich known profile. Okay. Let's go back to the homepage and see it instantly tells me that you're a community member, you are a VIP, et cetera. And since I like outdoors, that's exactly the recommendation I'm getting, including the product recommendations which are given to me. Okay. Awesome. Now next, I'm going to walk you through how all of this gets powered. So on Data Cloud, we have web and mobile SDKs. And as part of the connector, we have a schema defined. The schema is how we are going to know what are things to -- that you want us to collect from a website or app and how to map it into our schema. And then below this, as you'll see, we have something called a site map. The site map is nothing but a tracking code, which allows us to know what are the things we want to track from a website, example, clicks and other interactions. And now we also have a module for Einstein personalization built into the site map. It's called the PSP module. Let's find this PSP module. Okay. Here we go, right? Now in addition to this site map, we also have the CDN script here, which you are going to host on your website. Okay. So once you have done the SDK setup, the next thing you need to do is map the data streams. Data streams are basically all the data that come through to Data Cloud in the form of web events or something else that we ingest. And how is this data mapped in Data Cloud? For example, we have contact point e-mail address, right, e-mail, behavioral events, IDs which you want to send out, and then other personalization decisions, et cetera, and a number of other data streams here. Let's click into maybe behavioral events, right, and see what it looks like. So now this is a stream where all events that come in through the website are ingested. And so if you go here -- let's go there, and you see all the IDs which are mapped. The fields which will come in are mapped to events, to our internal Salesforce standard data model, right? For example, you can see shopping cart, you can see order engagement, product engagements and a host of other standard or custom model objects, which are part of our data model. And this mapping helps us understand how these fields are going to be stored in our objects, or DMOs as we call them, so that Data Cloud can understand them. Once this mapping is done, the next piece we want to focus on is identity resolution. So what identity resolution does is it allows us to identify a user when they land on the website and understand if they are a new unknown user or an existing known user, and if an existing user, then which existing user, so that all their past and current data can be merged together to form a comprehensive unified 360 profile, right? Now in real time, we look at basically exact matches only. So you can go ahead and build any rules on any ID that you prefer, but we will run every match as exact match even if you have configured fuzzy matches when you define the rules. For example, if you have created fuzzy matches on name or e-mail, they will be run as exact matches in real time to identify if you can find the same user with those names and e-mails. Now this is useful because you don't have to update your existing rule set in which you defined already in the lakehouse for it to run differently in real time. All you have to do is create a data graph using an existing rule set already defined, which is how it works, okay? Okay. The next piece from here we look at is data graphs, which is nothing but a denormalized view of all the fields of the user's data that is [ pieced ] together so that it can be accessed very quickly. A data graph can be created for a user profile or for a car, which would be a car profile or for a product, anything like that, right? Effectively, you're defining the fields and attributes for this profile or user that you want to access in real time and you want processed in real time. Okay. Here, we have a bunch of different profiles like we have auto profiles. This can be a car. Or you have product category down below, right? And then obviously you have, finally, the real-time profile, which is the consumer profile, right? Let's say, if we click into the consumer profile, this is what it will look like. And as you can see, it starts with the unified individual, where we want to have all the IDs of those users, which will include their cookie IDs, their CRM IDs, et cetera, which are all mapped into this unified key, which we call as a key ring. And then you have all the browsing data, category browsing data. You have contact point details. You have product browse engagement below that. And you have product data, e-mail engagements and a whole lot more which you can add here. As you can see, you can also add calculated insights, which you can see there are 3 of them here, right? And basically, these are insights which can run and get updated in real time just as events happen. And they can answer questions like how many times did a user interact with a specific product category or what is the total value of products that a user has purchased, et cetera. You can calculate these metrics in milliseconds. Okay. All right. Now let's jump next to the goods product data graph. So what we've been looking so far is what's the consumer data graph, but now we are looking at the goods product data graph, which has all information about products, their SKUs, the categories that these products belong to, et cetera. And you can even add calculated insights, but at a product level, right, built at a product level. And this is a data graph which is used for personalization to decide which product recommendations to give when you're on a website. Okay. Now just one thing I want to repeat. Like I said earlier, a data graph doesn't have to be about users. It can be about cars. It can be about products. And of course it can be about users, too. All right. Next, I want to talk about real-time CIs, or calculated insights. We also call them real-time insights. So these basically allow you to calculate simple sum, count, et cetera, aggregations as the events come in. An example of this would be if you view the product for the fifth time. Now I want to do something or change recommendations based on that. And so the calculation on how many times a product is viewed is happening, being done in real time, using a real-time CI. Here's an example of this, an existing CI where we are aggregating the total incoming events ingested for a user. Okay. Next, we have real-time segments. Segments are very similar to CIs in terms of how they are built. But what it does is it allows you to build a group of users, right? Or in other words, it allows you to build an audience segment using event attributes or even CIs that you may have defined. All of that processing, remember, happens -- and segment building happens -- in real time, in milliseconds. An example would be that if you clicked on red shoes, then you will be added to the red shoes segment, which will be updated in milliseconds, in real time. Okay. That was all we had to cover. Thank you so much for this. Finally, I want to mention that real-time Data Cloud performs all of this processing that we saw end-to-end in sub-1 second, actually in 300 to 500 milliseconds. Thank you so much.

Awesome. Thank you, Nikhil. What you guys saw was a real product, which ran all the way from ingesting the data to updating the profile, running the CIs. Let me walk you through how all these things work. I'm going to go and explain to you how the data flows. For everything in real time, the key thing is the real-time data graph. Real-time data graph is that foundation layer, which is what the whole real-time layer is built on and the stack is built on. Real-time data graph could be built based on your business need and your business logic. You can select which objects and which fields which are the most valuable for your business, you bring in those. And you build this graph with all those fields and even the history of the customer which you want to keep, which you think is the most valuable, will be available in milliseconds. And you want to process those in milliseconds. And think of like this, right, you have a user who's browsing through your website. And on the website, you have an SDK. The SDK actually just collects the web data immediately as the browsing event happens. As soon as the browsing event lands, the first thing we do is we ingest that data. We ingest it so that we know how to ingest it directly into data graph. And then we run an identity resolution. We first go -- I mean, Nikhil just spoke about that -- we go and look if there's a profile already exists for this user. Yes, if it exists, then we will retrieve that profile and start updating that in milliseconds. If there's no profile, we actually go create a new profile and start putting all your behavior attributes to that profile. And we update the data graph. The next piece which comes is the calculated insights. Think of this, you're looking at that red shoes. Every time you look at that red shoes, the calculated insights, the real-time calculated insights, is processing, oh, view 1 time, view 2 times, view 3 times. It's a simple sum and count ability to increment the sum or the count values and provide that values for segmentation or for personalization, right, ability to know that, oh, I've viewed 5 times this product. Basically, I'm giving a signal that I'm super interested in this product, right? How do you get that signal? That's the calculated insights which comes in and helps you do that. Segmentation. We allow real-time segmentation, every event which happens in your website. We evaluate every event in real-time segmentation. You can build segmentation rules based on a bunch of things, all the profile attributes which you have on the data graph. Or you could include like CIs, if someone has seen the product more than 5 times, then put in that, yes, he's highly interested in the red shoe category, right? And we run this, the whole evaluation, within milliseconds and add the user as the user browses through the website and make it available so that the next interaction or the next event, that segment is updated and is available for personalization immediately. Actions. Actions are coming soon. They're coming very soon. We are in the process of finalizing the product of action and delivering it. It allows you to do real-time actions. Real-time actions uses all the real-time data graph profile attributes, which you can act on. Or you can also use the CIs, the CI values to act on. Or you can also use the segment values to act on and you can send that truth flow to an e-mail activation or to an SMS activation. Or it doesn't matter, any activation which flow works on, this will work. We will be able to send the action out in milliseconds via the flow. Real-time actions is natively integrated with the flow, enabling you to tap into all the power of flow to enable all actions. Furthermore, we also update the data graph with all these events. That data graph is now your profile API, and that profile API can be called through an Einstein personalization. This is our Salesforce product, which actually does the personalization on the website. Our Einstein personalization product is natively built on data graph. So there is no more these 2 different profiles sitting at 2 different places, and you have to sync those profiles. You don't have to do that, because it is that one single profile which is giving you all that insight. And Einstein Personalization can deliver those personalized content immediately based on all the actions [indiscernible]. If you look at our architecture here, from end-to-end, by the time we take the SDK, the interaction from the web, by the time we update and the profile API call can happen, it's under 500 milliseconds. We deliver that in sub-second. That is really, really fast. That said, given the real-time layer is focused on processing data really, really fast, a lot of times not every data is going to come into the real-time layer, nor you want to push it, because sometimes the actions are not real time. It's actually much slower. We also send the same SDK data into a lakehouse layer, which is what our streaming layer is there, that is where our batch layer is there. Think of this when you're bringing AWS large batch data from S3 or you're bringing Snowflake data or you're bringing any other data into lakehouse or your off-line processing data. That data, plus the real-time data, lakehouse will process the same identity rules which are there in the real-time layer, the CIs, plus with all other information of the lakehouse, all the segmentation, and run this in the regular cadence. And it runs with all the accuracy. A lakehouse layer is a truth layer. It's a source of truth because it gets every data for the Customer 360. And then there's also a lakehouse data graph, which normally mimics a real-time data graph, which process this lakehouse data with the real-time data and pushes that data back into the real-time data graph. That's how we keep both the real-time layer and the lakehouse layer, which is basically one is your core layer or semi-warm layer and an extremely fast, hot layer processing in sync. We do this syncing on a regular cadence so that we make sure that your zero copy data is also available in the real-time layer, so that when you do the profile API call, you are able to get not only the real-time profile attributes, but also your lakehouse or your zero copy data also back into Einstein Personalization or a third-party personalization or any other system via the profile API call. This is what powers everything. We have two layers, one at extremely fast and one is pretty fast. All right. Let's go to the next slide. What you saw here is the end-to-end real-time capabilities. Our unified profile is where Data Cloud real-time capabilities shine, right? We get all these continuous profile data, both from structured, unstructured, from zero copy. Through all these different sources, we are able to process all this data, integrate it and process that in milliseconds. We saw that previous slide, that from the time we ingest to the time we actually deliver actions or bring insights, it's sub-second, around 500 milliseconds. We allow you to make that fast decisions so that you can deliver that customer experience immediately. The key thing is all comes down to building that unified profile, consolidating the data from all these different sources so that you have that one single view of the customer. It doesn't matter where that data comes from, it comes from web or it comes from a store or it comes from a batch file or it comes from Snowflake via zero copy, we unify all the data and provide you that holistic view of the customer to know what's all going on with the customer. We have built this whole infrastructure in this scalable and flexible architecture from a B2C scale, like from a consumer scale perspective. Not only that, all this real-time data is ready for your Agentforce, is ready to be integrated into the AI, so that they know the context of the user so they can respond back with just the accurate information. They know the context of what the user is doing on the website, what they have clicked. Users are sending us signals by clicking in those things and our AIs can read those signals and respond and cater to customers' needs. So if you see one of the top examples, if you want to talk about, it's actually the personalized services. Customers are not just looking for answers or solving the problem, they are actually expecting a proactive experience. They want the brands to know that, "Hey, this is my problem. This is what I've been trying to figure out and I'm stuck." The agent should already know that information because they tried this on the website and the consumer doesn't have to explain all those things, right? That is what the whole proactive personalized services is. Let's walk through a few of them here. So before I talk about that, one key thing, to enable any of those scenarios, the key thing which I keep talking about is about bringing the data together across multiple different touch points, across multiple different data sources. We have over 250 prebuilt connectors, which we seamlessly bring, unify that data across all of those, so that you can have the real-time calculated insights, so that you can enable this real-time calculated insight, can enable you these key insights to do cross-sell and upsell. Or you can say that this is the proactive service, which is -- this is the challenge the customer is trying to face. It can actually help the agents boost productivity, to tell that this is what the user is trying to solve and help them with this. Not only that, they will be able to also capture this unstructured data in all different ways so that the agents can also respond to this. We are unlocking a lot of things with the combination of unifying the data with real-time and Agentforce. Let's talk through an example. Think of it like this: you are browsing on the website, you are unable to log in and you've tried different things. On the website, the web has SDK, it's sending that information and you are speaking to an agent, saying, "Hey, I have challenges with my log-in." I think you shouldn't even have to tell that. The agent themselves should automatically know that you have challenges with your log-ins and the ability to tell you how it can help. This is the key from a proactive service side, ability to learn the different signals which the customer is already telling to the brand, getting that web data so that the agent can know that this is what the problem is, and even before the customer is asking, "Hey, this is my problem," the agent is able to tell them, "Do you have troubles with log-in? I can help you here to reset your password." That would be a faster way to solve the problem. This is where the end users are expecting from the brand, that not just having the data, but being proactive with that data, using that data, which I'm giving you in signals by clicking on the brands, using that to help me resolve my problems without me describing the problem. Let's go to the next one, right? We spoke about how to solve these problems, but even all the service agents can also start thinking about taking out all the insights of all the engagement the user is doing. Think of this: you are a retail customer and you're helping the customer on identifying a specific product, you could actually look at all the browsing behavior of the user immediately as they browse, and you could see that. Furthermore, you can know that specifically, they are looking at a shoe number, size 7, but they're actually looking for that wide, because previous history tells me that they always buy the wide-angled shoes, they're unable to help that. Ability to get that insight, ability to respond that faster instead of that user telling, helps the service agent now not only be a service agent, but also be a little more -- help sell them and close the deal for the customer and help them increase revenue, right? The next piece which I specifically want to talk about is the cross-selling and upsell, and how, given that as the users interact with the brand, it doesn't matter if they speak to a service agent or speak to a sales agent. For them, they're interacting with the brand. That means the service and the sales agents are all trying to collide themselves and they're expecting a personalized experience, not just in solving the problem, but also telling, "Hey, these are your additional offers, this is what you can unlock." The service rep ability to cross-sell on that call or on that chatbot specifically on the current interest and what they have done and the ability to upgrade them or give them expedited shipping, it doesn't matter what that is, helping them save. [indiscernible] [ sales lens ] also is one of the key things which we will be able to unlock because they can understand what the current behavior and the past behavior, all normalized and providing for the user, not looking at silo of like, "Hey, this is my case history," but rather, "Hey, you reached out about this, but I can also help you unlock this," by giving them an upgrade path or saying that, "This upgrade path will save you additional 10%," which is much cheaper, right? This is the power of real-time in Agentforce. It will unlock all kinds of capabilities across your sales and service agents. It's going to enable you to increase revenue. It's going to enable you to be faster in solving issues. It's also going to help you just increasing customer satisfaction, building that long-term relationship because you're able to see all the insights of the customer. That said, I want to invite back Kuber to talk about the next few slides of our customer success. Thank you so much. Kuber?

Thank you, Vishal. Sorry, while I get my video in process, I hope you guys can see me now. So so far, you've learned about what is real-time data. You learned about how Data Cloud is uniquely positioned to help you make sense of all your real-time cloud and data. You also see the insights of it. You saw some use cases. You saw some potential benefits. Now let's talk about where the rubber meets the road. Let's talk about customers. Let's talk about our customers. And some of you may be in the room as well with us right now who are -- perhaps businesses with the power of data and Salesforce. I have 8 interesting stories for you. I'll go through them one by one in brief. Number one, let's talk about Heathrow, a well-known brand. It's the primary airport not just for the U.K., but also one of the biggest airports in the world. They really wanted to improve the experience of their customers. As an airport, they have an interesting spot. While they do serve millions and millions of customers, not a lot of them really choose them. People usually choose an airline which just happens to pass through Heathrow. So they have an interesting experience with their customers, but they wanted to learn more about their customers, their grievances, their delights, and help better in supporting them. With the power of Data Cloud, they were able to satisfy their customers a lot more by building clear profiles of their customers. They were able to divert more customers from real-time customer support agents to chat agents powered by Agentforce so that they can simplify and make things more efficient. So that's an experience in transport industry. Let's talk about banking now. First Horizon Bank, one of the big banks, one of the larger, I want to say, regional banks here in the U.S., I want to say, based out of Nashville, Tennessee. They wanted to connect their clients' experiences and enable their employees, their bankers, to grow these relationships, learn more about their customers and drive cross-sell and upsell. By creating new Customer 360 profiles for all their customers with the power of Data Cloud, First Horizon Bank was, in under a year, able to grow their deposit by millions of dollars. They were able to unify as many as 2 million profiles, anybody who's ever touched them through different digital channels as a direct customer or as a potential customer, with the power of Data Cloud. And the interesting thing was they were able to do all this profile unification in days, and not months or years. Similarly, let's talk about Ecolab. Not a very familiar brand to you, but one of the biggest sanitation support companies in the world, I want to say. A Fortune 100 company for sure, with $50 billion-plus in revenue. Now they sell mostly directly to businesses. And in that way, a direct selling model, they had their salespeople talking to their customers, taking orders and slowly growing their business. A few years back, Ecolab decided they want to do more of B2B e-commerce now, and they wanted to better drive their customers to that online experience. With the power of Data Cloud, they were able to unify this commerce experience so that when their salesperson was talking to the customer, they were able to also offer exactly the same experience on online as well. Historically, Ecolab had 2 different sales channels or commerce channels, so to say, a direct model and an e-commerce indirect model. With the power of Data Cloud, they were able to bring together those 2 commerce channels so that they were able to boost their field reps' efficiency and responsiveness. They optimized their schedules and gave more opportunities to customers to make their choices. The net impact, in less than 18 months, Ecolab grew their average online order value by as much as 300%. At a scale of a company like Ecolab, that's like millions and millions of dollars in upswing. Now let's talk about a different industry, VillageMD. They are a general -- GP practice, a pretty large GP practice, here in Northeast United States. And they wanted to serve their customers. They had a tricky position. Being a health care company, they had a lot of sensitive data. They have to make sure everything is compliant with HIPAA regulations or any other regulation the industry goes through. But despite that, with the power of Data Cloud, Service Cloud and Sales Cloud, they were able to achieve a 360-degree view of every customer that they had, integrating data from multiple applications and sources. This helped them create personalized health care journeys using segmentation based on factors such as race, age, location, and lifestyle. And they were able to make actual life-changing [ decision ] impact for their patients/customers. They were able to reduce high-risk patient admission by as much as 30%. At the scale of VillageMD, that's like saving lives of hundreds of people every year. And not just that, they were also able to increase first contact resolution by as much as 15%, making their operations more efficient. So we've talked about Heathrow, First Horizon, Ecolab, VillageMD. Let's talk a few more customers in different industries. I'm going to pick up with Turtle Bay. It's a small, unique, boutique experience hotel in Hawaii. Since this slide was made, they've actually already rebranded recently into a Ritz-Carlton, but let's just call them Turtle Bay because they'll always be Turtle Bay to us. Now Turtle Bay is very unique and very boutique. They have less than 50 rooms, and they get very few customers every year, which makes their challenge interesting. Their customers are super diverse. Even though they get thousands of customers only every year, they come from all over the world. They believe less than 50% of their customers are actually from the U.S. They do see a lot of people from Japan, China, from Europe, which made segmentation and personalization a big challenge because geography is a very, very big differentiator. With the power of Data Cloud, they built really rich segmentation so that they can truly personalize the experience and the offers that they were giving all their customers. And they really had an impact. They saw 50% improvement in concierge efficiency and 20% increase in booking conversion because they were able to understand their customers much better and then serve them so much better. Similarly, Mascoma Bank, a big bank out of New Hampshire in the U.S., they also created a single source of truth for all their client data, including core banking, to allow their bankers to be very proactive in their recommendation. Now the next example, it's very near and dear to my heart, Formula 1. As a Formula 1 fan for -- I don't want to age myself, but for at least 25 years, since the days of Mika Hakkinen and David Coulthard and Gerhard Berger, I know from firsthand experience, Formula 1 has hundreds of millions of fans everywhere. And all that goes into Data Cloud. Think about the scale of driving these experiences. With the power of Data Cloud, Formula 1 has been able to really focus on fan satisfaction, fan happiness, things like e-mail delivery rates, e-mail open rates, customization, personalization, and doing this at the scale of an organization like Formula 1, delighting hundreds and millions of end users almost every week. Finally, a travel experience. Air India, India's leading airline, recently brought together all their customers' data to Data Cloud. Coming from a country of 1.4 billion people, they do serve tens of millions of customers every year, and all their data is served in the omnichannel experiences with the power of Data Cloud in near-real time. Air India, Formula 1, Mascoma Bank, Turtle Bay, and earlier we talked about Heathrow, First Horizon, Ecolab, VillageMD, those are some of our customers who are able to leverage the power of Data Cloud at scale. Now we recognize not all of you are using Data Cloud today. It still is a relatively new product. It has a long way to go. Feel free to use this QR code to get more resources about data. You can learn about Data Cloud on Salesforce Trailhead for completely free, at zero cost to you. You can download the Data Cloud guide to learn more. And if you are advanced enough in your journey for Data Cloud, I always encourage people to take the Data Cloud consultant certification exam so that you can be a certified expert on Data Cloud. Now this has been the end of our presentation. As we promised at the beginning, we do want to leave the last 10 minutes for any questions that this group may have. Now if you have a direct Salesforce rep, do talk to them about this. But right now, Vishal, Nikhil and I are here live for you to answer any questions that you may have. Slowly and steadily, we actually have been receiving a bunch of questions in expert chat section. So I'm going to start picking up some questions. And Vishal and Nikhil, how about I start picking up these questions and invite you guys as the product experts to answer them. [ Does that work ]?

Yes. I think Nikhil has already answered and I've answered a few of them, but we can ask those questions [ system ]..

But because all the answers can't be seen by everybody else, let's just try to get a couple of high-level questions. I'll try to get at least 5 questions, so let's just spend a minute or so in each answer. So [ Bonnie ] asks, how does the new feature Sub-Sec E2E Real-Time differ from the streaming data that already exists in Data Cloud? Vishal, do you want to take this?

Yes, that's a great question. The streaming right now actually process end-to-end in minutes. It takes a few minutes from the time we ingest data and goes through the IR and others. It's much longer. Compared to specifically the sub-second end-to-end, we actually process that in sub-500 milliseconds. So that's why there's a big difference between the streaming and the real time. So that's the answer.

Great. Thank you, Vishal. I have a great question from [ Daniel ]. He is asking, is the goal for real-time personalization with Data Cloud and Einstein to replace Marketing Cloud Personalization? I think it's a pretty straightforward answer. Do you want to go for it?

Well, I want to say that that's the eventual goal. I don't know if I'm going to say that's the goal right now. The Einstein Personalization is natively built on Data Cloud. Because it's natively built in Data Cloud, it gets all the benefits of Agentforce, your governance and the unified profile and IR. And there's no copy of data between what goes from Data Cloud to the current Marketing Cloud Personalization or Marketing Cloud Personalization into Data Cloud. You have to send data and copy, so there's a sync because that was an acquisition. Yes, over time, we want to make sure that the Einstein Personalization or the new personalization has all the features and capabilities of MCP and more, to deliver not just personalization but also decisions via agents. So that's the eventual goal.

Great. Thank you, Vishal. The next question is from [ Aaron ]. How does [indiscernible] Data Cloud performance of Einstein 1 tools? I'm going to take a shot at answering this. The simplest answer to this is, the more data you have, the better AI and agents will work. What Data Cloud does is bring together all your data, structured, unstructured, at rest, batched, streaming, and enhances the Einstein experiences. The more data you feed AI and agents, the better they will perform. So we always recommend power, leveraging the power of Data Cloud to power your AI agents. All right. I have an interesting question here from [ Mahindra ]. He says, the workflow that you just described, ingestion, action, et cetera -- he must have asked while Vishal was going through the workflow process -- is that a standard Data Cloud workflow or something that's built from scratch? Vishal or Nikhil, do you want to answer that?

Nikhil, do you want to answer or I can answer? I'm going to take it. So the whole pipeline of what we said, from ingestion, we have transformation, the CI and a bunch of those, that workflow actually exists in the lakehouse architecture right now, which is in the batch processing and others. We have actually simplified the same workflow in the real-time stack. The reason we did that is to make sure that the exact same workflow and the same logics can also be run in lakehouse. So there's no difference in how it runs. It runs exactly what it is. If the data is fed in, it goes through the same workflow in lakehouse as it goes in real time. It's just that there are differences between those 2 workflows where the real-time ones, like Nikhil talked about, identity resolution, we just look for exact matches, whereas the lakehouse will do normalized, fuzzy and exact. It has a lot more power because it has a lot more time to do, compared to the real-time layer; it has less time. So we need to pick the most impactful workload and do that. I hope that answers the question.

That was a great answer there. Thank you, Vishal. I have a great question here from [ Amy ]. She is asking about how do we surface knowledge content on an Experience Cloud site consumed into Data Cloud.

Right now, I would have to get back to you specifically on this one to connect with the Experience Cloud team to understand what they are doing on unstructured data. However, what I do want to say is slightly different. I do want to say that all unstructured data which comes into Data Cloud, is available to Experience Cloud. If you have Experience Cloud, which is connected to Data Cloud, then your both unstructured and your knowledge graph data is available in Experience Cloud. That is in process. In your settings screen, in your Experience Builder, there's settings and integrate, and there's something called Data Cloud. And as soon as you enable that, you will be able to pull all the information from Data Cloud into Experience Cloud. If it's going to be without Data Cloud, then there's something specifically I would have to get back to you speaking with an Experience Cloud expert.

Thank you, Vishal. Yes, we do have experts on this call, but we can't answer every question. I feel like this is the last question for the day now. [ Tarun Bhatia ], he's asking, do we have proactive service use case in B2B industry like high tech? And the quick answer is yes. I don't have a customer example in front of me, but I do encourage you to go and check out our website. It is salesforce.com/data/usecase. Specifically I'm saying /usecase because we have hundreds of use cases that customers are using Data Cloud today for. So you can learn from our website what are the possibilities there. All right. We are almost at time, and you guys have been so generous with it so far. I love the questions. Thank you, Vishal and Nikhil, for joining us today. I feel like if there are no questions coming in, it's the end of our session. Thank you again. I don't know if you guys are watching this live or this is a recorded version you're watching. Thank you. Thank you either ways for joining us. I hope this was worth your time. I hope that you will be able to leverage the power of real-time data in your workload sometime very soon if you're not doing it already. Thanks, Vishal; thanks, Ariana; thanks, Nikhil, for all your support. And see you sometime soon. Bye for now.

Thank you all.