Inside Salesforce’s Marketing Machine: CMO Ariel Kelman on AI, Data, and the Future of Marketing

0:00I think the biggest missed opportunity is people who forget about the benefits of education and they go too much to persuasion. I think there's a lot of products that they don't get stuck on convincing people why they should try them, but on educating them in really smart ways, using smart marketing tactics on how to be successful. So, you know, some companies get this great with really strong PLG mentalities, but there's a lot of marketing teams I see where they don't look at education as really part of their job. I'm Eric Fulweiler and this is Scratch, bringing you marketing lessons from the leading brands and brains, rewriting the rulebook from scratch for the world of today.

0:52Hey everyone, my guest today is Ariel Kelman. He is the president and CMO at Salesforce, obviously a business that you've heard of that is now very focused on realizing the future of the agentic enterprise. So we get into a lot of that good stuff, AI agents, AI, the three-layer model that y 'all will hear about from us pretty soon. But so many learnings and some great perspective from Ariel from the seat that he sits in within the industry that is really driving a lot of the disruption and the change that we're going into. And I really love how he talks about it. And, you know, we were kind of going back to, well, Salesforce is an original challenger brand.

1:31You think about 1999, 2000, you know, web 1.0 and how much of a challenger and a successful challenger they were to the traditional tech industry. And he talks about thinking about technology and AI as just helping you get rid of the dumb things that you have to do in your job, helping them go away. I thought that that was really interesting. He also talks about how to drive innovation within teams in a way that I hadn't heard about before. This idea of pushing teams to relax constraints as a way of driving innovation, which I'm definitely going to be thinking about. And then we talk about, of course, if you know Salesforce, you've probably heard of Dreamforce.

2:07And, you know, the massive conference that they do now 16 years in, but also the broader investment that they've made into community and how that's really been a flywheel to how they grow. And he gives some advice for anybody in a business that is looking to get buy-in and budget to do more community and kind of brand level marketing. So without further ado, please enjoy my conversation with Ariel Kelman of Salesforce. Ariel, how are you? I'm doing well, Eric. How's it going today? Things are good here in London. I know you are very busy over in San Francisco, so I really appreciate you making the time.

2:45Yeah, of course. So, Ariel, before we get into the conversation that we prepared for today, there's an icebreaker question that we ask every guest, which is, what is one challenger brand that you're particularly passionate about right now and why?

3:04Let's see. One challenger brand. So a brand that is really trying to shake things up and do things differently. Yeah. Although actually, I think it's interesting, especially speaking to, you know, the CMO of Salesforce, how would you define challenger brand or what does it bring to mind for you? that well i look i mean i think um i think companies go through various stages between being the incumbent and being the challenger you know you can't and you gotta i always think you gotta market a little bit differently in some places where you know the obvious ones the competitor situation do you talk about your competitors um when markets where you're the dominant player i think it doesn't make sense but even if you're a big company if there's a new space that you're going into where you have a you know single digit market share that you're very unhappy with and there's a big player in there you're going to treat that differently because you're the challenger in that market um you know i think back to when i was at amazon and we were trying to get into the cdn business against akamai like we talked about that differently than when um we were talking about um you know cloud infrastructure and uh the other competitors in that space where you know we had a lead so i think you know unless you're staying in the same market the whole time.

4:18I think you constantly have to look at yourself as a challenger in all of your new markets. I really like that. So it's less about, it's actually, you know, something that we say all the time. It's not about the size or scale of the company, uh, or even the stage of it. It's more about kind of the situation and the situation. You know, you think about it from a brand awareness perspective that there, you know, there are specific people who have, you know, ideas in their head for different things they're going to buy of who do they think of and who do they prefer. And, you know, people are, you know, not simple focus group subjects, you know, they have very complex views and often, you know, their views of what companies are good at, you know, are kind of diverse and multidimensional depending on what you're asking them about.

5:05Yeah. So maybe we can stay on that for a second. So, so, you know, taking the Amazon AWS example or something from Salesforce, if you'd like, you know, what is the biggest difference? You mentioned that Akamai example where it was a different kind of approach that you used in that specific instance. But how do you think about if you need to shift gears, what the biggest changes are, if and when you do? I mean, I think a lot of it is, you know, just going back to basics and thinking from first principles of what are you trying to accomplish? You think about like a lot of our focus right now is on a product called Agent Force.

5:40And we really believe there's a massive transformation happening with AI agents that are really augmenting people's capabilities and letting companies do things that really weren't possible before. And this is something that didn't exist for anyone a year ago. We launched Asian Force at Dreamforce last June, September, but really AI agents, the whole concept in a meaningful way is about a year old. And that's forced us to think about things in a different way. This isn't talking about the 143rd Salesforce automation feature. You know, this is about talking to people about a new technology, what's possible if you have that new technology, how it's going to change the way we work, and then why is it worth it?

6:24What are the benefits that you're going to get? So when we talk about letting people become two, three, four times more productive in certain activities, it just opens up new possibilities. But it also, it just transcends the typical incrementalism that you see in the technology industry of here's the latest release, which I think is fun. I mean, as a marketer, I think the best challenge is when you can have something disruptive, new technology that really causes people to change, oh, sorry, causes people to open their minds to looking at new solutions, whether it's new types of solutions or new vendors in a way that wasn't possible before.

7:03So we have, you know, Now, we don't make event management software at Salesforce, but one of the first agents that we did last year at Dreamforce is we built an event agent so that people coming to Dreamforce could go and answer any questions from, is there an analytic session tomorrow? Is there vegan food at the party? I happen to be asking, is there any beef at the party? but being able to have that direct connection with your customers at a massive scale with customer support resources with infinite capacity you know that now i guess we're in the event industry you know we're the event vendor industry now we have something for that so you can't take for granted your earlier perspectives and and for me entering these new markets is always pretty exciting.

7:50Yeah, it is. You know, you mentioned day one thinking, which I don't know if originated or at least is best known from Amazon. That's actually the reason that the name of this podcast is Scratch, because we think what it means to be a challenge is think and act like from scratch for the world of today. And another way that I think about it is, you know, my previous role before starting Rival, I was CEO of a fintech business, and we built a lot of kind of next generation digital first products for banks for the most part. And of course in that world, um, and then a lot of big tech, you talk a lot about technical debt building up within organizations, right?

8:27So the technology that was fit for purpose for 10 years ago or in banking world, a lot of it is from 20 or 30 years ago, obviously it's not fit for purpose for the world of today, but it's really hard to get rid of it, to rip it out, to replace it. So many of them have to kind of build something different. And I think marketing debt is the same thing in many organizations. So whether it's scratch, whether it's day one, whether it's marketing debt, whatever way into that perspective and opportunity you take, I think it's the same thing where what it's all about is thinking about the need of the audience, the category convention, and the value of the product from scratch as it is right now without any of that baggage or legacy.

9:09Yeah, I think that's important. And also, when you talk about the technical debt and thinking about things from scratch, I also think about being empathetic with the customer of what they have to deal with when people have built all of these systems uh based on doing things in a certain way um you have to think through well what do we need to deliver to help them be successful so you think about with we deal with this issue with agent force where okay if you want to apply uh generative ai and ai agents to helping consumers with cooking like okay you're just good to go because they went and trained on all the knowledge of the internet and all the recipes.

9:47But in the business context, you cannot have accurate, useful AI output if you don't have your data in order. If the LLM doesn't have access to understand the context of the situation, so you have a customer who has a customer support problem, you need the AI to understand, well, what products have they bought? Do they have any current customer service issues? Who am I talking to? What issues do they deal with? Are we trying to sell them anything at this point? Where are they geographically? Because that may affect certain compliance things like solutions even can or are not allowed to deliver them.

10:26And so if you don't have your data in order, you can't have accurate, predictive, useful, generative AI results. And And that means your agents will not be able to do things that are useful. So we have this product called Data Cloud, which is our customer data platform. And a lot of the work we've been doing with our customers for the past three years, this product's been around, is using that to really harmonize all the different data sources they have in the company so that AI can now take a customer ID and understand everything about it so it can help that customer, whether you're marketing to them, whether you're providing customer support.

11:06And so we do this internally. So I have a data cloud implementation. It reaches out to 902 different data sources we think of as data streams. So it's everything from traffic from our website to our Snowflake marketing data warehouse to third party data sources. But it is taking all this data from over 250 million raw profiles of customers and harmonizing them down to 151 million unique profiles that we sort of grouped together in about 800 segments. And so that pre-work that we did was necessary to really get the full value of AI agents. and so building that solution to pull the data together and understanding that that was required to really get the most out of ai um is really how we think about it to think of the whole way through not to just think of customers as having this you know magical utopia of turning something on uh and it immediately works yeah we have a big piece coming out around kind of our perspective on ai and the headline is basically that you know these cycles of technological disruption do repeat themselves because it's a lot of the human psychology and sociology and economics that drive what happens within it.

12:27And so we kind of relate it to obviously the technology and probably the impact of it is going to be different, but like the internet in 1999 or 2000 around that time. And really kind of what we're saying is that CMOs need to prepare for the marathon, not just the sprint. And there's kind of a lot of sprinting going on right now. And within that - You can do both at the same time though. I mean, I think that's one of the things to think about is that, you know, with the new technologies, you can't just say, I'm going to tackle the two-year problem up front. You got to get quick wins. Otherwise you lose momentum in the company.

13:00So that's the good thing with AI agents. Because if you want to get an agent up on your website that can answer product questions, hand off leads to your salespeople, we did that first. That took a couple of weeks to get up and running. And we now have had 700 ,000 conversations with customers, with our AI agents on our website in less than the past year. Meanwhile, while we're also developing longer-term, more sophisticated projects that almost reinvent how we think things like Leadflow. So I like to always think about, you take the new technology and you get the quick wins in parallel with doing some of the longer things.

13:39Yeah, it's about perspective really and having both of those horizons because I would argue there are a lot of businesses that are only might be unfair, but maybe too focused kind of on the sprint. But I think the bigger, you know, there's not having the momentum internally, but I think the bigger risk is just kind of getting left behind. You know, I think one of the biggest things on CMO's mind is how do I keep up with the pace of change, especially around technology and AI. And you have to be out there. You have to be using the technology. You have to be finding ways to apply it internally. But a big part of kind of this work that we're putting out is talking about what we call the three-layer model.

14:15It's very similar to what you're talking about. But the technology itself, AI agents in this case, is just one layer, right? And underneath that is data. And on top of that is the humans, the talent that you have using the technology. And so again, to perspective, which is going to drive strategy and investment, it needs to be all three of those layers, not just the technology itself. Yeah, absolutely. And I think, you know, a lot of our customers have, you know, they've sort of said, well, if I'm going to start with AI agents, do I think about autonomous use cases or augmentation use cases? You know, the autonomous use cases when the AI agent will just talk directly to the customer, augmentation is when you think about your salesperson having, you know, the world's best infinite capacity salesperson sitting on your shoulder trying to help you.

15:03Same with customer support. So, you know, a lot of our customers start with the employee facing agents to begin with and, you know, and then start to combine them together. So you think about like a company like Finnair, which is one of our Agent Force customers. So they're Finland's largest airline and they get a lot of customer service inquiries. And what they're able to do is both reduce the amount of inquiries that come to their human reps. And then when their human reps are working the leads, sorry, the leads, the always thinking marketing, the customer service issues, they have agent force next to them, making sure they're always giving the best possible answer.

15:46So we find like, you know, you got to handle both of those aspects at the same time. People get a little bit myopic of like, how do we remove the humans? And I think that's not really the point right now. There's so much productivity gain and being able to augment the capabilities of your human employees with AI agents. Yeah. It's the specific objective. It's the job to be done at the end of the day and how technology and or, but in many cases it's and, humans can best solve that. So, Ariel, do you have a challenger brand that you want to share? There's a company called Ramp. Have you heard of them you're in fintech yes that's a good one so i met those guys a couple years ago and you know that it just uses example of you know you think about managing expenses for a lot of people and and this kind of gets to my world of thinking about like what are dumb things that we should be able to get rid of with technology well the idea that you have to pass out an expense guidelines document and tell people to read it and say eric you really need to follow the guidelines.

16:50Did you look at the wiki and tell salespeople you can only expense lunches with your customers? Like, that should be obvious, but then they don't. And then they get in trouble and people do audits and ramps like, well, why don't we use software to create rules? If this rep is a, you know, if they are a sales rep in New Jersey, well, we're going to reject, we'll tell the system, reject every credit card transaction that's outside of the state of New Jersey or, you know, route it to their manager for approval before it goes through. And, you know, software defined credit cards and they do a lot more than that, but it's just not how the credit card and expense kind of vendors thought at the time.

17:36And it's just so obvious that that's what I always love is when you come up with something that, you know, really, you know, you'd say changes people's life in a consumer way, but in a business way, you know, upends how people think of, like, what's possible and what this type of solution should do for me and this type of solution. Like in these companies, it's, oh, well, I have a corporate credit card vendor over here, and then I have, like, Concur or whatever my expense report system is, and they're doing different things. so they're there you know that that's a company i think they're they're just they're crushing it and they continue to do well because they're they're not bidding into the existing categories it is interesting from a um so you know again i work worked in fintech before and we've done a lot of work in fintech and financial services we actually work with one of raps competitors um but how you describe what they do has given me a completely different perception on them because i would argue that actually a lot of their advertising and of course marketing is not just advertising but a lot of their advertising i'm not sure communicates their point of difference if you ask me what's good about ramp i'd say they seem innovative and you know they're the classic make um you know business financial operations easier but um i'm gonna go check out more i actually think that they're on our list i think they got nominated for we're putting out this rival 50 list, this annual list of like the top 50 challenger brands that we've come across.

19:05And I think they were definitely on there. So it'd be interesting to see how they do in the actual ranking at the end based on their marketing or their advertising against the principles that we have for challenger brands. Hey everyone, just another quick interruption. If you're listening to this podcast and you're not already part of our AMP community on WhatsApp, would love for you to get involved. So AMP is a community that we started at the beginning of the business really just to kind of bring together some of the interesting people that we were talking to, the marketers that we were coming across, many of whom have been on this show as well.

19:37And so now AMP is about 300 senior marketers from around the world from various categories, different stages, challengers, incumbents. And it's really a place where people can share questions, can share resources, can make new connections, can use folks as sounding boards. And for those that are in it, you know, it is a very vibrant community. It's really taken on a life of its own, which is amazing. So if you're listening to this and you want to get involved, please go to our website, wearrival.com. There you can find the application form or reach out to me or someone at Rival directly. Thanks so much.

20:13Obviously everybody listening has heard of Salesforce, but maybe people don't know the full kind of universe of where the business and the brand is today. So And we'll get into the conversation that we had prepped for today. Sure. So if you hadn't heard of Salesforce, a lot of people will ask, what is Salesforce? So let me give you an overview. So Salesforce was founded in 1999. So we're about a little over 25 years old. And what we're really focused on right now is helping customers build what we think of as an agentic enterprise. So it's taking AI agent technology that can allow humans and AI to work together to solve problems all around the world of customers.

21:10How you sell, how you market, how you service your customers. And really giving you a data foundation to pull in all the data that you have about your customer, organize that, and then to harness that data through interfaces for all of your human employees so they can really have the most effective interactions with your customers. and to have incredibly effective AI agent power interactions with your customers through all the different engagement channels, whether it's AI agents for customer service, integrations into digital media, email, SMS, WhatsApp. Think about Salesforce being at the center of all of your engagement with customers, both through your employees and through AI.

22:00Love it. And I was excited for this conversation. And obviously Salesforce is a big name, but I think more importantly, it's one of the, um, one of the OG challenger brands that we always reference. And, um, I feel like that's, well, maybe it's a question actually, it feels from the outside, like it's still very much part of the DNA and the culture of the company, that kind of challenger mindset. And I know that we talked about it a little bit earlier, but, um, how much does that manifest for you as CMO, kind of the original Salesforce as the challenger within the old tech industry? Yeah, I mean, it is in the DNA of our company, especially when you have a founder who's still the CEO and the company is larger.

22:40We're constantly looking at things. I would say it's a little bit of a sense of entitlement where we just feel this strong belief that everyone is entitled to have all the dumb things they have to deal with go away. So you think about back to when we started the company and Mark said, why can't Salesforce automation be as easy as buying a book on Amazon? that you think about what people had to deal with these year-long, super expensive implementations to do things like Siebel. So you could just do something as simple as entering a forecast and to have any change in a field or record. Oh, I want to store a different type of field here.

23:31Oh, well, that's going to be a two-month-long IT project. It's like, no, people shouldn't have to deal with that. We're going to have a completely different way to do it where you can just go into a web interface and say, add field, save, and then boom, it's all out there. And that mentality of looking at problems in a way that we think that things shouldn't have to be as hard as they are today, you know, as part of that, if you kind of look at, and a lot of it is you have to be willing to disrupt your own business, to look at what's not acceptable with the current way of working. you know this is something i think apple is you know always been great out of being not afraid to cannibalize its own products so we look at our our own business processes with a challenger mindset of if you look at lead follow-up so you know how like this is in the b2b world you get um you generate a bunch of leads for your sales people you score them you route the high ones to your sales people and then all the rest uh you put into some kind of email nurture campaign but really a lot of them are just in the garbage.

24:33And for us over the past 10 years, we haven't followed up at all on 20 million leads, either because they're too low scored for the reps to pay attention to them or they're in some kind of nurture thing. But when you think of a world with AI agents where you can have unlimited capacity of your best SDR, think of the low-end sales people who do follow up on leads, if you can have unlimited capacity, well, you can change that whole business process. You can say it's unacceptable that you don't follow up on every single lead. And so we built an SDR agent that we have internally we've been using here.

25:17And so we're having about 14 ,000 leads every week that would have been in the garbage heap are getting followed up on. and we're really turning some of them into opportunities. And that's just like found money. But we could have easily said, no, no, no. Like this existing process, we've been doing it for a long time. You know, we score leads, it works great. But it's to constantly look both internally and externally and say, what are things that are going on that are just dumb? They're just unacceptable. We need to be able to think about it differently. And is there any new technology that can, you know, relax constraints?

25:56In my mind, that's like one of the biggest things of, if I sort of switch a little bit to managing a large team, that you need to, of how you get your teams to innovate is to push them to think about what's required to have them relax constraints of things they said weren't possible before. In this case, it's that, well, we can only follow up on the leads that score four and five on a one to five scale. Why? Oh, we don't have capacity. What could change that? And so I always try and apply that in different places of asking people to write down, like, what are all the things you can't do that in a perfect world you would like to do that are slowing you down or making you, you know, not as successful as you could be in this project.

26:38It is the most misunderstood role within the C-suite. Absolutely everything about the industry changed. C-mails are really f***ed. Out of all the Fortune 1000 businesses, there's only 41 of them that have a CMO on their board. I think you're going to see more functions that were traditionally in the selling realm moving toward marketing. In today's world, and especially as we look to the future, the role of the CMO is at a crossroads. They'll be taken in by cheap parlor tricks. We need to reinvent and reimagine marketing totally. Deliver on your demand gen goal or lose your job. If you can do that and make your CEO look good at the same time, actually, that's a really powerful thing for a CMO.

27:21It's not enough to just push that on the juniors in your department. What does the future of the CMO role actually look like? So, Arielle, I want to make sure that we talk about this concept of customer zero that you use at Salesforce. So, what is customer zero? Why does it matter? And how do you apply it internally? I mean, this goes in the software industry. There's very few really good breakthrough products where the companies had not done this concept. I think originally people talk about dogfooding, but it's like, look, use your own damn product. You're going to learn a lot from that. You should be the first product.

28:03If you think of a new ranual list starting with zero, you should be customer zero. and you know so i'll take an example of our for our agent force product we have you go to our our help website help.salesforce.com um there is an agent that can answer customer problems and they were after i think the first uh the first agent we did was our dream force event agent um but the first i'd say sophisticated uh agent force deployment that we really focused on was this help site. It's like, you know, customer service, the obvious, you know, most obvious use case for AI agents. You have a, what are, LLMs are really good at synthesizing massive amounts of information.

28:48You can ask it a question, how do I do this? How do I fix this? I mean, I can take chat GPT, take a picture of a tomato plant and say, what's wrong with this plant, right? So it should be great, but you have to pull together all of this data, integrate your processes. There's a lot of, you know, a lot of complexity there. So what we did is we just rolled it out and we basically created a team between engineering our it organization and our customer service group and drove one of the most intense projects in the history of our company to iterate and iterate and iterate and to have as you know any new technology you start to use it this works this doesn't work to have that instant feedback loop to your engineering team to basically be revving, you know, with every two week sprint on new code that's coming out based on this customer getting.

29:38And of course, you know, the internal customer doesn't get to say, I'm not going to try this new feature you came out with. You know, you're like, yes, we're going to try it because we're trying to iterate, make this better together. So they sacrifice as a customer a little bit of being the guinea pig, but it's just invaluable for making a great product. I mean, it's literally 10 times faster to iterate because you're removing the friction of having another company to work with. It's hard. I mean, some companies struggle with this. If I, you know, date myself, like there was this story of when Microsoft was building Windows NT, they made the developers write their code on machines running Windows NT.