FY18 The Year of Einstein - Ch. 2: Coca-Cola | Salesforce

With that, I want to bring up Shubha Nabar and she's going to talk about how Coke is using AI to transform how they connect with their customers.

Shubha, thank you. Parker. Let's talk about the intelligence transformation that's taking place in enterprise software. As consumers, we interact with AI all the time. When I talk to Siri on my phone, that natural language understanding capability is powered by AI. When Facebook automatically tags my pictures, that image recognition capability is powered by AI. But for the majority of business applications today, that level of intelligence is missing. What my team and I are doing is changing this. We built Salesforce Einstein to bring AI to everyone, so that now everyone has a data scientist. How does this work? Well, businesses are generating far more data than ever before. At Salesforce, we've been collecting this data from the very beginning. There's the CRM data of course, but also email activity, calendar data, social data. All of this data resides in Salesforce. And simultaneously, there have been many algorithmic and hardware advancements in AI technologies I.e. making it possible for us to build smarter and smarter software. Software that generates insights, makes predictions and recommendations, and even automates tasks for you. So with Salesforce Einstein, we're bringing all of these advanced AI capabilities into the Salesforce platform to make all of our apps across all of our clouds smarter. Now, every business is different and the power of the platform is that it enables businesses to customize these capabilities to their own unique use cases. The other nice thing about the platform is that it's enabled us to iterate really rapidly and build a whole bunch of new features. Since we introduced this just a few months ago at Dreamforce, I'm going to touch on just some of these features here. But we now have live sales cloud Einstein where Einstein is enabling sales reps to close deals faster with intelligence that's based on customer sentiment and engagement and competitor involvement. With service cloud, Einstein's supervisor is helping service agents predict customer satisfaction and also making concrete recommendations on how to better improve the customer experience. For marketers, Einstein is helping take the guesswork out of audience segmentation. Those are just a few of the examples, but the thing that really excites me the most as a developer is that we're now opening up these capabilities to external developers as well. One of the most amazing examples of this is the Einstein Vision service which was built by a world class team of researchers and it enables any app developer on the platform to embed state of the art image recognition capabilities into any app that they build. To give you an example, really a Mind blowing example of how these capabilities could transform your business. I wanted to introduce our chief scientist, Richard Socher.

Thanks Shubha. I am super excited today to talk to you a little bit about our research and, and how it impacts CRM and Einstein. We've already in a short time that Salesforce research has existed, published five papers at the top AI conferences. We hold state of the art benchmark results for complex natural language processing tasks like question answering, where we use dynamic co attention networks that we invented. We use reinforcement learning and deep learning to do summarization and machine translation and a whole host of other tasks. But research is only one of the important ingredients and knowing the right kinds of algorithms is only one of the ingredients to bring AI products to our customers. We also need to understand the data and the workflows that our customers have. And that's why we're uniquely positioned at Salesforce because we have access to the data, of course, in a very trusted way only for our customers. We now have the right kinds of algorithms through our research and great engineering and, and we uniquely understand the workflows that our customers have. And so I'm very excited to give you a first example of how one of our most amazing customers, Coca Cola, uses some breakthrough AI algorithms for their business. Coca Cola chose Salesforce for their sales, service and marketing. And they have millions of retailers, all of whom have big inventories and need to keep track of that. And now fortunately, Coca Cola can rely on Salesforce to help understand that inventory and automatically use AI to basically gauge its status. So let me walk you through an example. Here you see the app based on communities cloud that Coca Cola has. Every retailer, every store out there has the same kind of interface with Coca Cola branding. They can here understand the latest products and promotions. They can ask questions and get answers. They can trigger service cases. And what most of them have to do very regularly is to go and do their next deliveries. Now it used to take thousands of people to go and manually tag, sometimes with pencil and papers, how many bottles are in each cooler, and so on. Now this got a lot easier thanks to Salesforce, Einstein. So I can just, we have here a real cooler with Coca Cola and a couple of bottles. And for some of you, maybe Barry, it would be easy to know exactly how many there are from a quick glance, but it usually takes people a little bit of time. Now what I just need to do is I click here on next delivery, take upload and photo. And now I can go here, take a real shot and use this photo. And now with the power of Einstein. We use deep learning to do object detection. We don't just classify the whole image, but we identify every single bottle and every single object in that image. And we can get back to you and say, this is exactly the inventory that you have. You see here the little bounding boxes around each of these bottles. Now, this looks pretty good. So I confirm. Now, on top of that, we understand all the past purchases that this retailer made. And not just that, but we can correlate all of that to global marketing data and others. So you can click here on Einstein and you see recommendations that it made for this particular store. So let's look through them. First thing we notice is cooler purity 100%. That's great. Coca Cola does not want to see non Coca Cola products in a Coca Cola cooler. So Barry's happy. He has a big smile.

All right.

Now, we noticed also that there's no Coca Cola life left in this case. So we actually know from correlating this with demographic data and other things that that would actually sell really well. So we added six cases to the next order. And then we also adjusted your normal monthly order based on several other factors, such as March Madness. March Madness. Great promotional opportunity. Right now we're in the middle of March. Basketball fans all around the United States, really excited. Great place for Coca Cola to do promotion. So this looks all pretty good. So I say thanks and I can confirm my order and that's it. In a few clicks, I could have this order completely automated and don't have to think about it too much. And I can rely on Einstein and Salesforce AI to make the right kinds of decisions for my business. Now I'm done with the order now. Now I want to see what's going on. I get, of course, right away an email and can now track my shipment. Just like in the consumer space, we're making this super easy for larger retailers. Now, I see in my email here that the order is estimated to be delivered only on March 13, but my coca Cola cooler is already running low. I actually want this order to be much faster. So instead of having to call, I can click here on a little bot signal. And this is the first version of a bot. I actually anticipate a lot of exciting innovation and incorporating some of our breakthrough research and deep learning and natural language processing to eventually get into this. Here we have a simple options. In this bot. We can look at shipping, billing, and other kinds of issues. This is a clear shipping use case. So I just type in shipping it already knows the last order that I've made. So, yes, I confirmed this was our order from just today, but I want this sooner. Now, this triggers different shipping options. We can go to Express Delivery and select that one because we want it as soon as possible in our store. And now we can actually use the bot also to incorporate an upsell opportunity. Say, well, if you just add four more cases of Coke Zero, you get free shipping. Of course, who doesn't like free shipping? So we add those Coke Zeros and we're done. Now this was another example of how we simplify. There's a lot of complexity behind all these use cases. You have to have the right GPU infrastructure for computer vision. You have to have the right compute. You need to integrate different models and deal with a variety of different data set issues and complexities and labeling. We simplified all of that for you with Einstein AI. You can just use AI in your business. Even if you're a small mom and pop shop, a large company, huge business, they can all have very simple AI usage. Now Coca Cola has one of the largest marketing budgets in the world, maybe the largest. And they might also ask, how could Salesforce help us to use artificial intelligence to make our marketing a lot smarter and improve it? And to answer that question, I'd love to talk to you about Crux, an amazing acquisition Salesforce made last year that is basically a data management platform that can capture very securely a whole host of different kinds of endpoints. In fact, they collect over 4 billion endpoints every month. What are endpoints? Those are cookies on the web, app usage, promotional website usage, and so on. Lots of different kinds of endpoints, including, of course, CRM. Now, they can use artificial intelligence as well as simple rules to help you understand and build targeted audiences, and then have very personalized and timely advertisement through a whole host of different partners. So what does that mean for Coca Cola? Let me walk you through a quick demo of what Crocs would look like for Coca Cola. So here we see the Crux screen. You can see that Crux has helped us identify 12 million people here from first party Coca Cola. Data that's owned and operated, securely stored, encrypted trust is our number one value. So we have here 12 million people who are using 285 million different devices. Now we can zoom in and say, all right, we're in this March madness, March month. How can we actually use this data to create the right kind of targeted audience? Now, we could manually do this, but what's amazing is we can also use artificial intelligence to group different users based on a whole host of these different patterns. Maybe they've clicked on some advertisement, maybe they've watched a video there, they looked at some promotion. They. They visited a specific website around March Madness and so on. And we can now group these and immediately target one particular group of basketball fans that would be very excited to see this advertisement. Now we choose one of these groups and can now look at how many are there. There are 38 million basketball fans that we could market to across 93 million devices. Lots of web users, lots of app usage, and so on. But more importantly, and this is again, really amazing, we can now choose a bunch of partners and immediately crux also simplifies this. There are over 200 different partners, and we can now target this advertisement and send it at the right time over Pinterest or Google, Amazon, Walmart, lots of different options to bring this advertisement targeted to your right audience. And so the final result of that, we'll see on different websites. For instance, here you see on NBA or the Food Network, the right banner ad, maybe a video ad at the right time delivered to that person. And all of that complexity, again, identifying different patterns, going through millions and millions of marketing hypotheses to find these clusters and target audience. All simplified thanks to Einstein AI. And with that, back to you, Marc.

Well, thank you, Richard. And we're very excited about the huge progress with Einstein since we announced it at Dreamforce, We've now delivered it to all of our Salesforce customers around the world, and they have access to this incredible artificial intelligence capability right inside the Salesforce platform. I think the demonstrations that you made, specifically showing us Einstein vision and the ability to be able to get that object identification and do inventory management just with a simple cell phone camera was pretty incredible. And then to use the DMP to deliver the one to one marketing is great. So congratulations, Richard, on your research and execution. And now I'd like to introduce you to Barry is the CIO of Coca Cola. So welcome Barry.

Hi, Mark.

Great to be with you.

Thank you.

Well, Barry, welcome to San Francisco. And we're thrilled that you can be here. Of course, we're thrilled to continue our incredible relationship with the Coca Cola company. And you know, Barry, so much of this, what we saw today is actually your vision. We got together late last year and you said, I have this idea of a connected cooler, a smarter cooler, a cooler that can see. And you know, here we just saw this incredible. So can you just give us what is your vision for the connected cooler and Coca Cola and how artificial Intelligence can improve your business.

Okay, well, thanks, Mark. First of all, thank you very much for the partnership here. We've been able to build this demonstration together really, really quickly. And if you think about our business, we're a global business that has to act locally. And so for us to identify the biggest opportunity in an outlet is really important to us. And we have 24 million customers, we have 16 million coolers worldwide. And so being able to find better ways to optimize our inventory and to turn that into actions for our customers as fast as we can is a huge opportunity for us. And being able to then also optimise the sale within that outlet is really important for us. So if we could have the picture of success for every one of those 24 million customers and understand exactly what the right selling proposition is in every One of those 16 million coolers powered by AI is just a huge opportunity for us. And so, once again, taking global products and global ideas and executing them locally and giving our salesperson the biggest single selling opportunity that's going to drive revenue for our customers and at the point of sale is a huge opportunity for us. And up until now, until we've had this kind of capability, it's been a dream. It's been something that our best people can do. But what we can do is with the technology is really democratise that and make it available for all our salespeople and put it in the hands of our customers as well.

Well, Barry, it's a huge vision. It's obviously inspired us and it's inspired Richard. And I remember our conversations. As soon as I got out of the meeting, I called Richard and I said, gee, I think based on some of the breakthroughs that you've had with Einstein Vision, we might be able to do this incredible inventory analysis for Barry. And then to see it live and working is really cool.

It's amazing. Richard came up to us, I think it was two weeks after we met, and said, look, I've got a demonstration for you. And I thought, okay, here we go, this will be interesting. And I was Preparing for the PowerPoint presentation and Richard had the live demo built ready to go. And it's just an awesome experience to think that we can take ideas and concepts and put them into action, because once again, it doesn't make you any money, it doesn't give you a better outcome unless you actually put it into action. And we've seen that happen. So it's been really inspiring for us as well.

Well, I'll tell you, now we have Einstein Live for all of our customers around the world. And our dream for this year is that soon we'll have a billion users using Salesforce Einstein. Whether it's on our, on our email marketing cloud, our marketing cloud, whether it's in our community cloud like we just saw, whether it's in our commerce cloud or in our sales and service clouds, when we add all of those incredible user bases up and seeing the intelligence now start to roll out, soon it'll be a billion people using this AI technology. Where before business really didn't have access to AI, now that business is getting it and that it's approachable and easy to use. Where do you see this going? Where do you see it going in for your company, for other companies?

Well, it's interesting if you think about our business. We're an old business. We've been around for a long time and we plan to be around for a long time in the future as well. So we need to adapt and we need to digitize our business. We go to incredible lengths to make the experience compelling for our customers. And the customers are at the heart of our business. And from our perspective, we think, well, why wouldn't we do the same thing for our associates? And that's the big opportunity we see. We have over 700,000 associates through our franchise system and helping them become more productive, helping them have a better experience with their time with us, is all going to equate to better ways they serve their customers and also a better experience for how they work with our system. And so those AI platforms that can learn with our associates, that can tailor the experience for our associates to exactly like a one on one kind of education for them and experiences for them, I think that's a big opportunity for us. So taking that sort of maniacal customer focus we have externally and then applying that to our associates and the experience they have within our business through technology, I think is a tremendous opportunity for us and, and something we're really going to have to do if we really want to attract and retain the best people in the business.

Well, Barry, thank you very much for coming to San Francisco and I hope that we can spend more time with you today and introduce you to more customers. We're thrilled that you're here. So please thank Barry.