Dreamforce 2019: How Salesforce is using AI to improve conversion rates and boost sales

Converting clicks to sales is a constant struggle, and Salesforce is using artificial intelligence to provide good marketing insights and ultimately improve revenue.

Dreamforce 2019: How Salesforce is using AI to improve conversion rates and boost sales

TechRepublic's Editor-in-Chief Bill Detwiler spoke with Chris Jacob, product marketing leader with Salesforce at Dreamforce 2019. The following is an edited transcript of the interview.

Chris Jacob: There are four core components to our marketing solution, and we believe for any marketing solution, knowing your customer, personalizing what the customer sees through artificial intelligence, engaging on every channel, and analyzing, making sure you know you're successful in optimizing. In each of those four areas, we have one or two key announcements. A big announcement is our product called Customer 360 Audiences. Part of the overall Customer 360 Truth announcement we made in our opening keynote. Customer 360 Audiences is basically giving marketers the power to bring together the known and unknown worlds into a unified profile for the marketer that they will then be able to personalize and activate audience segments that are relevant to their business.

SEE: Hiring kit: Salesforce Developer (TechRepublic Premium)

I might want to activate to a certain group that's willing to respond to a certain Black Friday offer, for example. It's compelling to a certain SKU or whatnot. So that's Customer 360 Audiences, giving marketers insight through this known and unknown unified profile and being able to activate that data in segments.

For the personalized angle, we're amplifying our investment in Einstein and AI, and we have three great announcements that we're going to showcase in the marketing keynote: Einstein Messaging Insights, Einstein Copy Insights, and Einstein Content Selection. Briefly, Einstein messaging insights gives marketers the ability, when they log into their marketing cloud console in the morning, insights of things that have happened and things I may need to take action on. For example, I've launched a campaign the previous day, Messaging Insights tells me, "Oh, the open rate was down 30% or down 30% versus the equivalent campaign last quarter or last year. And it was down because the subject line was wrong or inaccurate."

That's important because suddenly you have all this data coming together, artificial intelligence, not just telling you something is wrong, but why it's wrong. What's the next thing Einstein can then go and do? It can then tell you what to do to fix it. That's what Einstein Copy Insights is. Okay fine, the subject line was the problem, what should I do about that? Einstein Copy Insights tells me what exactly I should change it to, the words and phrases I should use to improve the open rate. It actually starts using artificial intelligence, not just to alert you with what's wrong but give you the corrective action to fix it. And then Einstein Content Selection, probably my personal favorite of the three announcements for personalization, the ability to not just have Einstein and AI recommending the content that appears in an email, so the email you received versus the email I receive, but also changing that content in real time based on engagement.

Say we were both part of a campaign, we both received an email, and I happen to open that email on day one because I had nothing to do. You're a busy man, Bill, you don't open that until day three. Based on all those people that opened the email on day one or previously, we may engage or not engage with those offers that are in that email. When you finally get the time to open that same email and day three, that offer, that hero image might have changed in real time because all of us engaged or didn't engage, didn't respond to that offer. So there's no point showing you something that's not relevant to you and didn't engage the previous audience, so personalizing in real time.

Bill Detwiler: There's a little bit of real-time AB testing that can happen?

SEE: The ethical challenges of AI: A leader's guide (free PDF) (TechRepublic)

Chris Jacob: Yes, exactly right. Still, based on you and who you are, you don't want an irrelevant offer, but also taking into account all these people who have engaged with that content previously, it's very, very powerful.

That's personalized announcement. With Engage, our big announcement is interactive email. I think all of us know in marketing that when you're doing a conversion rate funnel, things always drop off. I get an email, I click to the landing page, there's a drop off in conversion rate. Filling in the form, there's a a drop off in conversion rate. Buying the product, going to the shopping cart page, etc., there's always a drop off at each step of the way. Interactive email brings all that power into the email itself. Meaning, that I don't have to leave the email to take an action. I can embed reviews, I can embed forms, I can embed a SKU of product selection, all within the email itself. Meaning that I don't ever have to leave the email to go to a landing page or anywhere else to complete the action or for the brand to get that information. So we think it will be very, very compelling. I personally think it's probably the biggest thing to happen to email as a channel since the smartphone itself in terms of changing the form factor of emails. We think will be very compelling, and hopefully it'll improve conversion rates and everything else for people, and people will love it.

And then analyze, the important part of knowing whether you're successful or not and taking action, we launched Datorama Marketplace. Datorama Marketplace is basically the ability to build applications on top of Datorama, to make it suitable for your business. For example, you can build skills that are based on things like voice or things that may be suitable to industry. In the marketing keynote, we're going to share the ability to use Amazon Alexa with Datorama, to query the dashboard using voice to then get insights back, i.e., "One of my top-performing channels," and then a dashboard will come up and show me the top performing channels for a given campaign. It gives marketers another interface to interact, not just with their hands and fingers.

Bill Detwiler: That's really been the theme here at this year's Dreamforce, is voice and the announcements around that. Talk a little bit more about how it makes it a more natural interaction. Maybe then, like you said, using your hands, being able to come in in the morning and get that information you need to make a decision without having to sequel, type a lot of queries, understand how to interact, waiting for someone else who does have the data to get back to you and tell you, "Oh, this subject line is better than this," or "This is why this campaign is performing badly," or, "This is my top-performing campaign over the last several months or over the last several weeks." As someone who actually does this for my business, as someone that's working with our audience to try to figure out, "Hey look, is this story a good story? Is this email a good email we sent out? Is this tweet a good tweet?" Being able to sit down and say, "Alexa, tell me about this." That's pretty powerful.

Chris Jacob: It is powerful. I think all of us, you and I are having a voice discussion now, it's so natural for this to be the way we communicate. I think none of us who work in this industry are surprised at how fast voice has grown. So many of us have Amazon Echos and Google Homes. We use Siri, Apple's Home Pod, and all the other devices that voice is enabled on. In the consumer world, we have adopted this en masse. And this spans all generations, right? I think what you saw yesterday with the announcements, it's only natural for a lot of those capabilities, for voice to make their way into the business world, right? Things often start in the consumer world and they end up in the business world. Why shouldn't a sales leader in the organization or account representative be able to query, go, "Am I likely to make my quarter? What does Einstein tell me about my ability to make my number? What recommended action should I take?" Instead of just looking at that in a typical interface, can I use voice to go and do that?

And then the second part to it is, because we know we can't build everything, giving developers the ability to build their own skills that are relevant to their own business and making it interoperable. Because voice is not just about having a closed ecosystem. If you do use Alexa, and if you do you Siri, having the Einstein and Salesforce capabilities with voice being able to work with those devices as well. That's why I'm so excited by it all because the possibilities are endless.

I don't know where it's going to go, you probably know better than me, Bill. But being able to give business people the power of voice in their interface, whether it's querying the database in a CRM use case, querying their database in a marketing use case, I think it's going to be super powerful. The service interaction, having things transcribed in real time, AI service and recommendations, that service agent to give a bit of experience to the people calling in to the call center. It has so many possibilities, and I'm just excited to see where it's going. Hopefully people were excited by what they saw yesterday, I certainly was, and we'll see where that goes.

Bill Detwiler: Talk to me a little bit about the kind of feedback that you're getting from people that led you down the path to integrate voice, to make some of these changes. From your customers, what are they looking for in marketing, the tools they used to market and to reach out to their customers?

Chris Jacob: I think it comes back to first principles in many ways, right? At the end of the day, our goals are very much the same, right? I want to increase revenue, increase conversion rates, increase open rates at the channel level, lifetime value, net promoter scores, customer satisfaction. All those things are still important. And so, when we think about what we're doing, we're still enabling those end goals. Now, if voice is a frontier that enables that end goal, I make someone more likely to take advantage of an offer in a marketing campaign and it makes a marketer more efficient in running their business, and she's just more powerful in being able to do what she does on a daily basis in the office. Then that's somewhere where voice should be introduced, right? Just like introducing touch a generation ago, well a technology generation ago.

Voice is just another mechanism, another medium, another frontier to enable someone to do something, whether for the consumer with the brand or whether with the marketer running the campaign, or the sales person running their business. And so, that's the way that we've looked at it for everybody. Can we make that professional better at their job, and does this improve the customer experience for their brand? To their end customer, business buyer, or end consumer? And so that's where we've gone with it. As I said, we're clearly in the early days of all of this, and it's interesting to see where it will go. But we're hopeful that because we've made it an open ecosystem, developers can go and build their own custom solutions with this, to see what's suitable to their business, and ultimately what their end customers are telling them that they want and need. And their internal employees are telling them what they want and need.

Bill Detwiler: Along with voice, another thing that's always important in terms of making actionable decisions on all this data and integrating what you're doing with AI, talk to me about visualizations. Talk to me about just being able to paint a picture, visually, so that people can see, "OK, well Einstein has told me that I need to do this in an email." Salesforce, one of the big acquisitions that it's recently made, Tableau of course, and it's being integrated across the platform to provide these kinds of visualizations and providing that information that Marketing Cloud can deliver to people in an individual way. How's that work?

Chris Jacob: Yes, visualizations, I think, is interesting because you've got two types of visualizations. You've got the visualizations that give you insights, i.e., the charts, the graphs, the trends, all those sorts of things. And voice can be a mechanism that gets you there. Artificial intelligence can surface insights that are far more difficult for anyone to do manually, even a team of data scientists to do manually. You can just do things more efficiently and at scale. Particularly with all the touch points that exist. So that's one form of visualizations. 

The other one is giving professionals visual cues about what to go and do. So think like next best offers: If I am that person at the other end of that service call, and this is a problem service call, and Einstein's telling me, "You know what, you should talk to them about this particular offer. This might not only help resolve the situation but actually get them to buy something else from you." That's really powerful, right? Because suddenly that service agent is not just a cost center but now a profit center for the business. And all you've done is used AI to surface an offer that that person can see in their console that they can communicate back. Now, to do that at scale, you need the service person to be, not just empowered with that technology, but the service person to be connected to the marketing side of the business, because the offers are usually being generated out of marketing. That's where it comes back to that idea of a unified profile. Everyone in the business has to have access that unified profile, but serviced up in a way that's relevant to them. A service person needs offers to talk to the person, the marketer needs content and an AI service things to generating insights and to power the next campaign.


Bill Detwiler, left, spoke with Chris Jacob of Salesforce about new AI concepts.

Also see