Einstein: AI in the Salesforce Platform | Salesforce

Saving you do magic, magic when you.

Good afternoon, everyone. Welcome to the Salesforce Einstein breakout session. My name is Ali Witherspoon and I am on the product marketing team of Einstein. So I could not be more thrilled to be here today and to be the one to introduce you all to Salesforce Einstein. It's my favorite time of the year. You might notice some of the decorations around. I think it's putting everyone in the holiday spirit and it's also my favorite city in the world and I don't say that everywhere I go. New York really is my favorite city, but you can't tell San Francisco that. And I get to talk about one of my favorite topics, which is artificial intelligence. So some of you may have been at Dreamforce, so you might have had a little taste of Salesforce Einstein, and know a little bit about what we're up to. Others of you are seeing a lot of this stuff today for the first time. But regardless of where you fall on that spectrum, we, we will have something for everyone today. So for the next 40 minutes, we are going to talk about why Salesforce got into this space in the first place and why we think it's really the perfect time for artificial intelligence. And we're going to introduce you to Einstein. So we'll kind of break down what this product is and we'll show you a couple great customer vignettes and product demos. And then we're even lucky enough to have a real life customer, Katie from DonorsChoose here today to tell you her story and how they're thinking about what artificial intelligence in a completely new way. But before we can get into all of that fun, I have to deliver the absolute most riveting slide of them all, the Safe harbor slide. You will see some forward looking features and functionality in the demos that we show you today. So just make sure you make your purchasing decisions based on the features that are generally available. So I want to start by just talking a little bit about why Salesforce got into this space and what we're doing with artificial intelligence. So since our inception in the late 90s, we have been about taking consumer technologies that are really prevalent in the consumer world, bringing them to the enterprise. So we did this first with the cloud and we wanted to make Salesforce automation as easy as buying a book on Amazon. That was our mission. And then we did it again with social. So we drew inspiration from some of the great social networking sites like Facebook and Twitter and YouTube and Instagram and we wanted to bring that same engagement model to the enterprise and we did it with Chatter, which we know many of you guys use today. And then we did it again with mobile. So we knew that there was a huge mobile revolution going on. Six billion mobile phones, more and more time spent than ever on these devices and in mobile applications specifically. And so we really changed our product philosophy and the way that we build products to make everything mobile. First with Salesforce One, with Wave analytics, with every product that's come out. Subsequently, it's all about making it work across any device. And then when we launched Thunder, our Internet of Things cloud, once again, it was about taking this trend of connected products and connected devices that we were seeing in the consumer world and bringing it to the enterprise so that companies could start making sense of all of this data coming from these devices. And we're doing it one more time with artificial intelligence. Well, maybe not one more time. We might have some tricks left, but we're doing it again with artificial intelligence. So we know that in the consumer world, AI is already all around us. We use it every day in a lot of these key applications that, that we're using in our personal lives. But the thing is, a lot of us don't even know that we're using artificial intelligence. So when I open my iPhone and I ask Siri for a weather update or I ask for help defining a word, because we all do it, that's an example of natural language processing, which is one technology under the artificial intelligence umbrella. And then when you log into Amazon and you see product recommendations instantly served up to you based on your browsing history, based on your click history, and based on other custom with similar patterns, that's machine learning. So Amazon is actually learning based on all of those clicks and all of those interactions, and they're tailoring and customizing the experience for every customer. And then one more example is what Facebook is doing with deep learning. So when you upload a photo to Facebook and that photo is automatically tagged for you, that's deep learning at work. So Facebook is taking all of that unstructured image data and recognizing patterns in it. And a lot of these examples, what they have in common is they help us make better decisions as consumers, they help us make smarter decisions, they help us work faster, they help us be more productive. But when we contrast this with the workplace, it's just a different story. AI is very complex, and a lot of businesses have just been slower to adopt a lot of these trends. But it's not really for a lack of wanting AI. We run focus groups all the time. We talk to hundreds of customers before we launch Einstein. And all of the customers. Customers really wanted more intelligence. They wanted to get smarter about their customers. They wanted more predictions in Salesforce, but they also recognized that with that comes a lot of complexity below the surface. So that first item of complexity is really all around your data. So when we talk to these customers, they tell us a lot of times that their number one pain point is just the sheer amount of data that their business has to deal with on a daily basis. A lot of it lives in siloed, disparate sources and it's not integrated and it doesn't talk to one another. And then even if you are lucky enough to have a team of data scientists to kind of try to integrate that data and make sense of it, it's really expensive to hire data scientists today to train them, to keep them at your company. They're a hot commodity. And those data scientists have to model, they have to have the infrastructure and then they have to put it all back in the context of your business. Because you can't just do AI for AI sake. You have to then make those predictions visible to all of your business users. But as Salesforce, we love a good challenge. And we wanted to take this problem that we saw. We wanted to reduce the complexity for our customers and make AI really accessible to every business user. And so that's what we've done with Salesforce. Einstein, it's all about baking AI into the Salesforce platform that you guys use every single day. So taking a lot of those consumer technologies like machine learning, deep learning, predictive analytics, taking inspiration from the consumer world and baking it into the Salesforce platform platform. So the first thing that this does, it allows the world's number one CRM to now become the world's smartest CRM with insights and predictions and recommendations. Right in Salesforce, Right in the CRM that you guys use every single day. And then secondly, it's empowering all of our business users. So no matter what industry you're in, which function you sit in, you can take advantage of artificial intelligence in your day to day workflow. And then lastly, we love a good platform at Salesforce, so we built Einstein to be platform first so that everyone on the development spectrum, whether you're a business user, an admin, a data scientist, a developer, you can all start building apps that are powered by artificial intelligence. So let's just take a look under the hood now and see how this works kind of in practice. So it all starts with your data, the customer data that you've been putting into Salesforce since that first time you logged in. So it's a lot of your structured data, your email data, your calendar data, but it's a lot of unstructured data, like text from your emails, image data, and even Internet of things data from connected products and connected devices. So we're taking all of that data, we're pumping it into our machine learning system, which is Einstein. And Einstein is learning from that data. It's training itself on your data, and it's creating recommendations and predictions for every single business user based on patterns in your data, so that you can discover insights about what's happened in the past and why some of these things have happened. And then you can predict what's going to happen in the future. And then based on those predictions, Einstein will recommend the next best business action that you should take, whether you're in sales, service marketing, or it. And then we heard from a lot of you that you wanted more automation in Salesforce. So Einstein automates a lot of routine tasks for you, like typing emails and data entry and things that we know that you're bogged down with. And we'll show you those kind of four pillars as we walk through the demos today, from Discover, to Predict, to Recommend, to Automate. And what this allows you to do is it really frees up time for every business user so that you can spend less time on the mundane, less time guessing and checking, and more time engaging with your customers. And so what this looks like, when we look at it in the context of the greater Salesforce kind of product stack, it's not something separate. It's not off to the side. It's not a new cloud, it's not a new application. It is an intelligent layer in the Salesforce platform. So just like Thunder is our layer of data and Lightning is our new ui, Einstein is our new layer of intelligence. That's all about taking those consumer AI technologies like predictive analytics, machine learning, deep learning, NLP and nlg. We're baking it into the Salesforce platform, and we're making our platform smarter. We're making all those applications that you see along the top smarter. And so we're making you guys smarter and more predictive about your customers as a result. So for the rest of the presentation, we're going to kind of divide it into two chapters. We're going to start by talking about what we're doing with the platform and how we're kind of building AI into the Salesforce platform. And we are going to do that through the Lens of a company called Square. I'm sure some of you are familiar with them. And then we'll switch and we'll kind of show you some of the applications that you might be more familiar with, like sales cloud, service cloud, marketing cloud. And Willie will show you how AI is manifesting itself in some of those applications that you guys use every day through the lens of fanatics. And then DonorsChoose will tell their story as well. So we'll go ahead and get started with the platform. So, as you know, clearly you've probably seen Einstein running around a lot. We drew inspiration from Albert Einstein when creating this platform product. And he was probably most well known for his equation E MC squared, which is all about taking the complex and making it simple. And so we came up with a similar equation for AI. And that is data plus models plus business context will result in artificial intelligence. We'll kind of unpack each one of these for you. So the data, as I mentioned, it all starts with your customer data. So whether you're using sales cloud, service, cloud, marketing cloud, it's all of the data that you've been putting into our platform since day one. We are analyzing all of that data and we're building predictive models based on your specific data. Because every company has their own data. They have their own use cases, their own processes, their own workflow. And there's just not enough data scientists in the world to create and hand tune models for all of these companies. And so Einstein now does automated machine learning for you. So it trains on your data and is able to create the best predictions for every single business. But we don't just do this, like I said, for AI sake. It's all about putting it back into the context of your business, for your business users in a really seamless way so that you're not learning a new ui, you're not taking your data out and putting it back in somewhere else. These insights are all served up right in the same UI that you guys use every single day in a very seamless way. So you're not training yourself on a new tool. It's all right there in Salesforce. And the result is really AI in the Salesforce platform. And like I said, whether you are a business user or an admin or a data scientist and a developer, you can all build AI powered apps faster than ever with Einstein. So on one end of the spectrum, we have our business users and our admins. And in our Square demo today, you will see how you can embed AI into your applications with Lightning So if you're a process builder fan, you will see how you can take those predictions from Einstein and treat them like any field in Salesforce. You can report on them, you can create workflow around them. And then for our data scientists and developers, we have a few APIs that you guys can leverage that we have Predictive Vision service and Predictive Sentiment service so that you can start making sense of all of that text and image data out there. And then for those of you who are more interested in building custom predictive models, we are offering Prediction IO in Heroku private spaces. So if any of these pique your interest, you can learn more at our zone in the campground or you can come talk to us after the session. So now to show you some of these features in action and show you AI in the Salesforce platform, it is my pleasure to introduce Megan York, Director of Product marketing.

All right, thank you, Ali. And I want to introduce you to my buddy James over here in the demo desk. So today, James and I are excited to tell you this story of Square. So just like Salesforce is democratizing AI for all of our customers, Square is democratizing financial services for thousands of customers in the form of payment options, marketing support, even small business loans. Now, with thousands of customers, it sometimes poses a challenge for some of those Square sales reps to know what contacts and what leads should they be following up with first. Well, enter Salesforce Einstein to give them some help. And we're going to show you a demo of how they use that. Today we're going to do our demo through the lens of one of our rock star sales reps named Jim. So Jim wakes up in the morning, he opens his email and he sees he gets an alert from Salesforce Einstein. But you know what, this is not exactly what Jim wanted to wake up to today. Einstein is actually predicting that he's going to miss his quota by pretty significant amount this quarter and that is just not going to get him to president's club and he is eyeing that for sure. So what he's going to do now is going to click in into our sales wave dashboard and get a little bit more insight into what is going on here. And under the hood Wave Einstein is helping Jim understand how he can close that quota gap. So we're doing a lot of analysis on all of Jim's history to understand how things like industry and close rate, how those play into how he can close that gap. So he's doing some experimenting here and he sees that if he focuses on the restaurant industry, that's his greatest chance of closing the gap in his quota. So what does he do? Like any good sales rep, he goes back into the office and he pulls up his lead list of all of those leads that he has in the restaurant industry. Well, when we pull that up, you will see that the marketing team at Square has done an awesome job of recruiting leads for Jim in the restaurant space. But again, this poses a problem. How does Jim know who to follow up with first? So he then goes into his home screen and again, Salesforce Einstein is there to help him with these Einstein insights. He's seeing who is his top leads to follow up with through their lead score. And so what Einstein is doing is it's taking all of your custom and standard fields and understanding how things like role and industry and geography, even things like the last web page visited, how all of those things play into a lead's likelihood to create close. And then we're assigning that lead a score to understand where they should fall in our priority list. So now Jim can go back into his lead list. He can sort this list by that lead score, and he can see who are the top people he should be following up with first. Okay, now, but that's not the only thing we can do with this lead score. Jim is really lucky because he has some awesome trailblazers at Square who have built him some great processes that he can use to understand which leads he should be following up with. So when those leads come in the door, process builder is actually looking at that lead score. And if that is the lead has kind of a low score, we're going to automatically put that lead into a part out nurture campaign so our marketing department can send that lead all the information they need to warm them up a little bit. But if it's a hot lead, then Jim is going to be notified right away that he has somebody to follow up with through maybe an email or a text alert so he knows that that one deserves his attention right away. And then for all those leads in the middle, we're going to route those to maybe our junior sales representative and let them follow up with those people in person. All right, so now Jim is rocking and rolling on his restaurant leads. Now what he wants to do is check out his opportunities. And James here is going to pull up Jim's opportunity board. And you'll see this looks a little different than it used to. Thanks to Einstein. We've got some green arrows, some red arrows, some gray areas. Those arrows are showing Jim what is going on with those opportunities. And we see here that Sock Depot has a red arrow associated with it. Well, Jim's a little bit bummed because he thought things were really going well with Sock Depot, but he clicks in and what Salesforce Einstein has done is uncover that there is an issue with this account. And it's done that through combing through all of Jim's email and his calendar and through natural language processing is understanding that there's an issue going on here. So because Jim's email is embedded right here into the sales cloud, he can send Julian from Stock Depot an email. Right. And he suggests that maybe they meet up at Dreamforce and talk about this issue in person. And again, because of that vetted email, after Jim sends the email, we will see that Julian follows up right away. And you know what? This has a good ending, this story. Julian and Jim, they meet up at Dreamforce, they get their issue all worked out and they're closing the deal. So if we go back into Jim's dashboard, then we can see that what started out as a problem with the help of Salesforce Einstein is really a good solution. Jim has been able to close the gap in that quota based on all of the insights that he has. So again, through this demo, what we wanted to show you was how the sales reps at Square use Salesforce Einstein to know which customers they should connect with at just the right time. Back to you, Ali.

Thanks, Megan. So now that you've seen a little bit of a kind of sales use case, we are going to talk more about AI in our core applications and we're going to switch to a marketing use case through the eyes of customer fanatics and then donors choose as well. But first, I just want to show you this slide which kind of is a representation of all of the new features that we're launching with Einstein. The main takeaway here is that no matter which cloud you've invested in from Salesforce or which product you're using, every app is going to get smarter with artificial intelligence. And this isn't just a one time launch with every Salesforce release. So three times a year, we will be rolling out new Einstein features for every single cloud. Now, what this really allows for is allows everyone at Salesforce, every customer at Salesforce to become a data scientist, or at least to have a data scientist working for them inside of their CRM. So sales. Now, like you just saw in the Square demo, sales reps know exactly which opportunities and which leads to focus on at the right time so they can hit their quota every quarter. And service agents are now assisted with Einstein baked into the service console and so that they can resolve cases proactively and be assisted through every step of the case resolution process. Marketers now know the right time, the right channel, the right content for every single customer. No more AB testing, no more guessing and checking. Einstein will help you through that. And it can now easily embed artificial intelligence into every single application you build. Like I showed you, no matter which end of the spectrum you're on, from low code to no code to high code, we have solutions for every single app builder out there. So with that, I'm going to have Megan come back up and we're going to show you the Fanatics demo and a marketing use case of a company that's really transformed the way that they connect with their customers.

Awesome. Okay, if any of you have ever bought a jersey from NFL or the NBA or major League Baseball, there's a good chance that you've purchased that from a site that is powered by fanatics. And Fanatics knows that the relationship that we have with our sports teams that we choose is highly personal. At the risk of you guys throwing things at me today, I'm going to tell you, I'm from Indianapolis and I'm an Indianapolis Colts fan. And so Fanatics sends me an email that mentions anything about Tom Brady. I am not going to be pleased with that. Likewise, I'm guessing a lot of you do not want Andrew Luck gracing your inbox.

Go Colts.

Go blue. Okay, so, but Fanatics knows this relationship is personal. And you know what? It's using Salesforce Einstein to understand all of our preferences and then automatically deliver content that is based on the relationship that we want to have with that sports team. And they use the marketing cloud to do that. So, James, let's go ahead and pull up the marketing cloud. And like any good marketer, we all begin our day in an analytics dashboard. Right? We start looking at what our customers have done in the past, and then we use that to help us make decisions about how we should reach our customers in the future. Well, this analytics dashboard is different. This is predictive scoring. What we're doing here is not looking at what our customers have done in the past. We're looking at what they're likely to do in the future. So what percentage of our customers is likely to make a purchase or likely to click in an email or even likely to unsubscribe? Using this kind of data, it helps us reach our customers where they are in their journey. And Salesforce Einstein is doing another cool Thing here, it's servicing some segments to fanatics that they might be interested in based on the data. And we see these window shoppers. So our windows shoppers are people who are highly engaged with our marketing, but they're not making a purchase. So what Fanatics wants to do is send these customers on a journey to help shepherd them to that coveted conversion. And we're going to do that through Journey Builder. So James, let's go ahead and add that audience to Journey Builder. Now, if you're not familiar with the marketing cloud, Journey Builder is our platform for one to one customer journeys. This means Journey Builder can listen to customer behavior. It allows marketers to drag and drop channels onto the canvas so that we can connect those experiences across channels and then we automate all of that in real time to meet our customers where they are. Now, as marketers, we're always thinking about four things. When we're planning a campaign or we're planning a journey, we're thinking about who should our audience be, what channels should we use to reach them, what timing should we reach them, and then what should that content be? And fanatics can use Einstein to help us determine the optimal component of all of those. So like we said, this predictive audience, our window shoppers, that is going to be our audience for our campaign or for our journey here. Now let's talk about channels. James is going to drag and drop over a smart split. And with the smart split, Einstein is actually recommending to us what channels we should use to reach our customers. And then we're also taking it a step further because we're automatically routing customers down the path that they're most likely to engage with. So if I'm most likely to open an email, I'm going to get an email from fanatics. James is more likely to interact with an ad in his Facebook feed. So you know what, he is going to get that Facebook ad again, making sure that the channel is highly personalized to a customer's preference and also being delivered the channel that we're most likely to engage with. So we've got our audience, we've got our channel. Now let's talk about timing. So using send time optimization, Einstein is able to understand what's the optimal time to send any message to an individual consumer. You know, I might open my email in the morning, so that's when I'm going to get the email from fanatics. James is most likely checking his Facebook account over lunch. He's totally not doing it during the workday and so he is Going to get that facebook ad sent to him at lunchtime. Okay, so now we have our audience, We've got our channels, and we've got our timing. Let's look at what personalized content looks like from a customer's perspective. So Ali indicated to you that she is from San Francisco. She's decided to jump on the warrior bandwagon. Who doesn't love a little Steph curry? And so she is going to get an email from fanatics. Now, Ali's been on the website checking things out, but she hasn't made a purchase. She's one of those window shoppers. And what does she get? She gets a personalized product recommendation just for her. And so she likes what she sees. So allie is going to go ahead and buy that jersey through the fanatics website. So she makes the purchase and then the next message is going to be a follow up from fanatics, actually encouraging her to come join the fanatics community. Now, communities are great places for you to connect with your customers and also to let customers connect with each other. And fanatics is running a really cool campaign. If you post your picture to the community site, you can be entered to win season tickets to your favorite sports team. Ali here is kind of a sucker for a really good selfie. And so she decides to participate in the contest and she uploads her picture to the fanatics community. Now, all this is great for customer engagement, right? It's a really fun customer experience, but that's not where it stops. Phanatics can get some great information from these photos using Salesforce, Einstein and our predictive vision services. So what Einstein is doing is looking all those photos and understanding the context of each one of those photos and seeing in those photos what are those customer attributes based on those pictures that were submitted. But it's not enough just to get the insights around the customer attributes. What can we do with those? We know all those attributes are then appended to our customer profile. And with that customer profile, we can then use that to make even more personalized product recommendations to that individual. And so then when allie goes back into the community site, what she sees is that personalized product recommendation just for her. And we're showing you now how fanatics can use Salesforce Einstein to do all those things. Determine the right audience, the right channel, the right time with the right content. Now, we've seen how this can play out in a retail setting, but what about a nonprofit? How could a nonprofit use Salesforce Einstein to connect to their donors in all new ways? I'm really excited today to have Katie Bisbee from DonorsChoose, one of my favorite nonprofits in the world, to talk to us about that. Thank you so much for joining us, Katie.

Thanks for having me.

Can you tell us a little bit about DonorsChoose and what you guys are.

Trying to do there?

Yes.

So public school teachers on average spend about $600 a year out of their own pockets to get critical resources they need to teach their students. And so our mission is to make sure our teachers have exactly what it is that they need to give their kids a great education, whether that's basic supplies like books or microscopes or technology for their class, or whether it's experiences like a field trip to the zoo. Teachers can go on our site and request whatever it is they need. And then when a project is fully funded, we ship those supplies directly to the school. And then our secondary mission, or at least something we spend a lot of time thinking about, is how to make for donors they're giving the best philanthropic experience that exists on the web. So, so whether a donor gives $5 or $500, they get to pick the exact classroom project they care about. They get to see exactly how their money is going to be spent. And then they hear back from the classroom with photos of the kids going to the zoo or using the microscopes, and they also get a letter from the teacher explaining exactly how their supplies helped. So it's incredibly direct and personal and we always strive to make it better.

And even thank you letters from the kids sometimes. I'm telling you guys, it's the most amazing experience. So you guys have been a marketing cloud customer for a long time, and when we started talking to you about these AI capabilities, you were excited right away. Can you tell us a little bit how you're going to use them to reach your donors?

Yeah, and I'd love to explain how it is that we connect with donors, which may not be. Which may not be obvious given that we're a non profit. So when someone comes to our site to make a donation, they come with a specific idea or interest they have in mind. So it may be something they're really passionate about, like chess, or it may be a cause they really care about, like helping the environment, or it may be that they want to support a school down the street from their home. And so they can find any of those projects on our site, whether it's a chess set for a third grade chess club or whether it's recycling bins for an environmental effort at the school. And that chess club May even be in their neighborhood, who knows? So once they donate, in order for us to really build a long term relationship with them, we need to show them that we care about them by showing them that we've listened to what it is they care about. And so we need to serve up classroom projects to them. And for us, the best channel is email that we know that they all care about. And if we do that right, we spark that deep bond with our organization and with the classroom and we can get them to. To come back and give again and again. So it's actually quite, it turns out to be quite similar to e commerce companies, just like fanatics, has to send you the right jersey or has to send Allie the right jersey. And if they get it wrong, you won't purchase. We need to be personal and relevant and on point with our recommendations. And so hopefully that explanation makes it really obvious how artificial intelligence can help us, but we believe that it can make us understand our customers a lot better and be a lot smarter in our targeting and ultimately drive more donations to public school classrooms. And then as a nonprofit with our incredibly lean team, my two email marketers and I are here today. The ability for us to automate the data, gathering, the intelligence, the content, serving up is especially exciting to us.

I love it. Okay, so we have a special thing we're doing today in honor of the world tour. Can you tell the audience what we're doing with DonorsChoose today? Yes.

I get to announce it.

Yes.

Okay, so 100 of you here, thanks to Salesforce, are going to be getting a $50 donorschoose.org gift card to go find a project that you feel passionate about funding and you'll be able to, like on Salesforce, actually fund a classroom project. And I challenge you to search around and find something you really care about. And then feel free to reach out to me and let me know how we can improve our targeting and make it even better. And thank you so much for being so generous.

Thank you for being here. Katie, round of applause for our donors.

Choose.

Okay, so your assignment, speaking of teachers, your assignment today is go into the campground and visit us at the Einstein Zone where we'll scan your badge and then post World Tour. 100 people will be receiving an opportunity, like Katie said, to donate to DonorsChoose. All right, let's wrap it up.

Allie, awesome. Thank you so much, Megan, for those two amazing demos, for showing us AI in the platform, AI in our applications, and to Katie, for being, being here and bringing your team as well and sharing your story. So hopefully by this point, you all feel a little bit more acquainted with Einstein. You feel like, you know, kind of what we're doing with AI. You feel inspired by some of the customer stories that you've heard. And hopefully we've got you thinking about how you can start bringing AI into your unique businesses and how artificial intelligence might help you solve some of the challenges that you see in sales or service or marketing. So a couple call to action. You can visit us, like Megan said, in the campground, right when you walk in, you'll see Einstein front and center. We have a theater presentation going on there all day long. So you can see different use cases from sales to service to marketing, and really see Einstein in action. If you go to any one of the product areas, sales, cloud, service, cloud, marketing cloud, you can ask them for Einstein specific demos and they will show you where Einstein is appearing in those products. And then, last but not least, I'm sure you guys are all familiar or hopefully you are all familiar with trailhead by this point. We have two new trails on trailhead for Einstein. The first one is an introduction to Artificial intelligence. So you can take that and instantly become a data scientist and get your Ph.D. i wish. And then there's a second trail that's all about Einstein features so you can really go deep with Einstein and see how Einstein is appearing in all of our clouds. So with that, I just want to thank you all for being a tremendous audience for your time for your attention today. I hope you guys all enjoy the rest of the afternoon and stay warm out there. Happy holidays.