How Salesforce Drives 2–3x Pipeline with Data Cloud and AI | Dreamforce 2025

I want to share what it means when we say we're customer zero at Salesforce. Being customer zero means that we are our own first customer. Testing, launching and using our products before they ever go to market. We get to be at the forefront of amazing innovation happening across the organization. We are the first to try it, to stretch it, and sometimes even break it. We do this so when it gets to you, the product is vetted and ready to scale. But what does it mean to be customer zero? Ah, sorry. I'm going to go to the next slide. But let's start with a thank you. Thank you for being here today, for being part of this incredible energy at Dreamforce. Take a look around you and look at this amazing innovation, this energy, this community. And this is where the magic happens. But let me ask you something. How did you get here? Beyond the flights and the lines? How did you know what sessions to attend? How did you make sure you were at the right place, meeting the right people, making the right connections? Because for us, the marketers, the sellers, the leader, that's the challenge. How do we make sure that every single, single one of you are the right people, are in the right seats, ready to drive the right next level of success? And to us, that means maximizing impacts and revenue and doing it all without increasing our resources. Not so long ago, our process was completely manual. We were buried in spreadsheets, lists, flying all around from different teams. No single source of truth and really no clear ownership. But now with data360 and distribution through our Salesforce platform, what used to take us months now takes us days. It's streamlined, it's organized and most importantly, it's a rallying call for our entire marketing and sales organization. Every team knows their role. New customer, that's brands queue. Let's warm them up. It's time to get them into the Salesforce feel and ecosystem in an active deal. Well, sales, take it away. Dreamforce is the moment to capture your customers attention and get them in the door. Instead of working with silos, we're bringing it all together in one smooth motion. All driven by Data360 we used to be limited in bandwidth. We had to optimize and prioritize ruthlessly. And now we've automated all of the heavy lifting and freed up time. And what do we do with that new free time? Well, we double down on personalization. We make sure that you are getting the right message to get you here. But also inspire you to think about what you're going to do while you are here. So we're not overfilling the funnel anymore. We're just filling it with the right people. And you may ask, does this pay off? And the answer is a resounding, resounding yes. Last year using the strategic audience approach, we saw our Dreamforce marketing driven pipe jump over 22% year over year. Our campaign marketing driven pipe nearly doubled Salesforce's overall pipe gen for the same period. So now I showed you the why we do it. I'm going to pass it to Ali to talk you through the how we do it through data 360.

Thanks Lisa. Cool. So let's take a look under the hood in terms of how we actually build some of our segments that are helping us optimize our audience targeting for Dreamforce. So the secret ingredient really is our deeply unified platform. This consists of our Salesforce platform data360, the customer360 which we call our Truth Profile and Agent Force. And this rich 360 view really is the game changer for us. It's allowing us to make much deeper decisions on how we target our audience and be a lot more deliberate and intentional in how we communicate with our customers. Interestingly, last year audiences built using our unified platform data using this rich targeting method generated 3x more pipeline than audiences from more traditional sources, more traditional targeting like job title, job function based segments. So if we dive a little deeper into what our deeply unified platform gives us, we have all the usual stuff in there. Account data, contact data, opportunities, campaigns, et cetera. All the usual stuff you'd expect inside customer360. But we also have some highlights. These are some of the things that are really helping us take that extra mile in terms of how we target our audience. So we have access to all of our historical event attendance. So if you've been to Dreamforce in the past or this Dreamforce, any sessions you've attended, we can use all of that to build a picture of what your first party interest is. So rather than make an assumption based on your job title, which doesn't really work anymore, job titles aren't quite as predictable or as standard as they used to be. We're now able to target based on what you've told us you're interested in, what you've shown that you're interested in. Similarly, across our digital properties, we have our engagement metrics so we can have a look at what type of content you consume on Salesforce, what you've been interacting with on emails, what you respond to, what forms and websites you've visited, and also More recently, we started to look at your trailhead engagement. So here where you've shown an interest in learning about a particular product or technology, that again, is an extremely strong signal for us on an area of interest. So we can use that to again target our audiences more effectively. We also have buyer Personas, so these are more best practice types of approach for us. So this is where we've standardized and used a number of different methods to come to more of an agreement on what does our buyer actually look like. So rather than every marketing team or every sales team redefined for every campaign what a sales buyer might look like, we've gone ahead and taken research from our business partners. We've had a look at historical opportunities. We've used some machine learning here too to come up with some standardized baseline definitions of what these buyers look like. That means we always starting from a consistent, repeatable baseline that we can then just refine, but we're not having to reinvent the wheel every time we build a campaign. And then lastly, what really ties this all together for us is our machine learning models. This is all well and good, but actually what it represents is billions of data points. Tens of billions of data points in fact. And actually it's quite difficult to segment on that much data. So we rely quite heavily on our machine learning models that bring that all together and actually deliver us a really nice, easy to use score. We're quite fortunate actually that we can make decisions across all of these data points just by using one or two different models in our segmentation strategy. So what does this look like for Dreamforce? So for this year, we took an approach of building out this kind of pyramid view of our audiences. You know, I've been here for a little while now. We've done a few of these Dreamforce audiences and the conversation used to be ali, I need an audience for Dreamforce. And it's not singular, it's plural. We're talking about a lot of people here. You can't really bucket everyone into one group. We really need to start thinking about our audience a little bit more, a little bit more granular. So the way we tend to target, and we did target this year for Dreamforce is we come up with this prioritized view on at the top of the pyramid. We have our priority one, high value customers here. Elements like product telemetry, how much of the product they're consuming and using versus perhaps their opportunity value, what does the current engagement look like. And then we'll use also some next best product predictions to make sure that their experience at Dreamforce is tailored towards that next best product for them. It's not a great customer experience if you're already a customer of Data360 and then we target you with more Data360 content. So we really want to make sure that they have a good experience there and for them to be able to see the vision of what's next for them in terms of their, their engagement with us. As we move down, we see kind of growth potential customers. So this is where perhaps we don't quite have the same level of footprint, but we see that there's potential for growth there. There's some indicators and we're seeing some signals that maybe there's higher propensity to buy. So that's where we're again using some different machine learning models such as propensity to buy. We'll also then use product telemetry, but maybe also have a look at customer lifetime value and see where they are in that journey with us to try and build out some predictions on where we see the growth. Then we move more into our prospect space. This is where it starts to get a little bit more tricky. And the customer side, I tend to say it's a little bit easier to target because we know a lot about them. As you move into the prospect space, that's where we start to know less and less about you. So in our high propensity prospects, in fact, we do know a little bit about these folks. They've probably come to an event before. They're probably on our website, we have their email address and they're in journeys. So we understand a little bit more about their engagement and we'll use that to, as I mentioned earlier, retarget them. Then my favorite part, priority four, that's the hardest part. And bit of a glutton for punishment of course. But in priority four, these folks, we don't know anything about them. They're effectively white space. These folks haven't really engaged with us in the past. It's a big pontification. There's millions and millions of rows of data there. So how do we start to prioritize here? It's a bit more of a hybrid approach where we are able to score and model based on maybe much fewer data points. We'll use that. We'll also leverage some third party insights. So we'll look at kind of web engagement across broader digital assets, not just our own website. And then lastly we may actually lean on some of those more traditional job title function and attribute based segments and to target this audience. So let's take a look at this in action. So here we have our data360 production instance as it happens and we're looking at our truth profile, our customer360 and let's build that high priority one segment in real time. So whilst this is building itself, I'll just quickly talk a little bit about our methodology and we can watch this in the background come together. So we use a canonical segmentation approach. This is where we have a collection or a library I should say, of pre built macro segments. So what we've done here is we've gone and we've looked at many use cases across the business and we designed pre built quite large global typically segments around those particular use cases. And the reason we do it this way is much like I was talking about buyer Personas, it allows us to standardize our approach to segmentation where again we're not redefining all of the attributes and rules each time we build a segment, we're using these pre built library of segments. So the requirement then changes from what different fields we want to use down to what elements do I want my, what features do I want my audience to have? And which is much more accessible to more people in the business. You don't necessarily need to be a data nerd like me and kind of have a whole encyclopedia of our data assets in your head. You can scroll down the library and say okay, yes, I would like to have high value open opportunities because that's the type of audience I'm targeting here. So we effectively build segments from segments and we stack different segments together to build out the audience. Whilst this is still coming together, I'll share another example of where we're using this canonical methodology in data 360 to drive efficiency across the business. We have typically every year we'll have global top priority campaigns. These are CMO driven and again every we build audiences for our global marketers and when we get requests to build them, it's different every time for the same campaign. So if we use service for sales as one of the examples, every region ou will have a different definition of that. So it means there's a lot of inconsistency in our approach and sometimes we often cannibalize our audiences because we've put everything into the first one and not really thought about the hierarchy. Right. These are, these are kind of prioritized in a certain order and if you're not starting at number one, you'll often kind of use up all that audience before you had a chance to execute on some of the other ones. So the way we've gone about this is we're working with Lisa actually quite closely and some of our product marketing teams. We've defined the baseline acceptance criteria for what qualifies you to be in one of these segments. And we built those as canonical libraries. And really the beauty of that now is we can pull in a canonical library, just customize it and localize it. We may add a little bit of machine learning in as well to help us optimize the audience. But the baseline starting point is the same for everyone. So we now have global consistency in terms of how we go to market for a particular campaign. But we still have that local flavor, we're still able to localize it and make it relevant for that particular industry, country, whatever it might be. So I think we're almost there with our segment build. We can see here we're just adding in the last few components. So we're looking at here now our buyer Persona. In this case, we're going to include decision makers and advocates. So let's bring those across. And again, these segments that we're bringing in all have the rules baked into them so that the business rules have all been agreed. We spend a lot of time working with again, our marketers and stakeholders on what defines an advocate versus a decision maker. We've reached consensus on it and then it's locked into the segment. As the business might change, we may change those definitions. We can update the baseline segment and any segment built using that will also then adopt that new criteria. So it also allows us to drive change at scale quite quickly. Cool. So that's the segment done. Let's do a quick run through of what we did. We have a standard data, standard segment that really is more of a compliance type of thing. It gets rid of any junk stuff that shouldn't be in there. So we apply that to every segment we build. Then we have our high value opportunities because that's the type of segment we're looking for. In this case, we're only looking for customers of sales or service. Then we've sprinkled in some machine learning. So where our next best product recommendation indicates agent force or marketing cloud is the next cloud. And then lastly, as I said, we were looking at decision makers and advocates. So with that we can hit recount and let's see what that does to our universe. So we're starting at 17 million and in a few seconds we should see, as if by magic, 3776 people. It's the first time I've done this, honestly. So now that we've got that, this is what we were expecting. This is quite a highly targeted audience. It's the top of the pyramid. It should be small, but we're ready to go. So now that that audience is built, all we need to do is set up our activation targets. Are we sending it to a paid media platform? Is this going into an email campaign? It's very easy for us to now deploy that audience across multiple channels or single channels, or reuse that audience many times over as we see fit. So let me hand it over to Lisa to talk a little bit about how we now drive personalization on these audiences.

Yeah, so let's go back to that pyramid. You can see that now. According to that pyramid, we will determine our distribution channel, our messaging, and how we're going to reach that audience. And that's key because we are activating the audiences. And by doing so we can start building really rich triggers against the data that we've, that we've received from Ali. As we move up and down the pyramid, we can actually expand our objectives. Driving awareness, employing scalable tactics, and making one to millions really feel like one to one. So we are super excited about the way that we've approached Dreamforce this year and we are expanding this. This year we deployed around 400 different variations of emails, digital campaigns, et cetera. We are really doubling down on our personalization elements and really looking at our pyramid to try to mobilize both sales and marketing. But what's next? I'm going to pass it back to Ali.

Thanks, Lisa. So to finish us off, the last slide really is with all these hyper target audiences, we've thought about our personalization. We have all the content ready. This is where really AgentForce comes in. And we pour all of this into AgentForce because it allows us to do this at scale. We've done this for many years, but are doing it manually. One to one with lists and different campaigns is extremely complicated and actually quite difficult to get right. It's often quite error prone. Agentforce has really cleaned all of this up for us. So this is really how we've transformed our personalization journey into something very hyper targeted. So before the event, an example, you know, AgentForce is delivering for us, hyper targeted personalized agenda guidance and tailored session recommendations on site. Hopefully you've all had some interaction with AgentForce this week. Even if you haven't noticed it, it's in the app, it's everywhere actually. It's providing you walking directions. It's suggesting what to do next based on your location. Might even be booking some consultations for you. And then after the event, we'll be using it to help us send a personalized recap video on demand. You'll be able to catch this on Salesforce if you want, share it with your friends and family, and seamlessly connect you with your sales team. So this really is the ultimate evolution for us of event engagement. It's the final piece of the puzzle, turning strategically targeted audiences into guided revenue generating experiences.

And with that, we want to thank you for joining us today and being the right people in the room. We are excited and we are grateful to have you.

Thanks everyone.