Agentforce Marketing: How to Drive Growth with AI Agents

Hello everyone. Welcome. We're going to be talking today about how Salesforce marketing drives growth with Agentforce. Thank you for taking time with us. I hope everyone's had a great Dreamforce so far and are looking forward to another terrific day. Let's go ahead and get started with some introductions. My name is Michael. I run digital marketing here at Salesforce. The technology side, we're all members of the digital team here. I kind of focus on things like all the technology for our events like Dreamforce. So it's been a pretty cool week as well as our websites and I'll be talking a little bit about the websites but also we have Jen with us.

Hi, my name is Jen Atkins. I manage our data and marketing automation internally at Salesforce. So we manage the team and I manage our data 360 instance which supports and powers marketing, sales, customer success and Agentforce. So I'm really excited to talk to you guys today about our data journey and how we've built out data360 internally. I'm going to hand it over to Amar.

Thank you. Hi, this is Amar Ajla. I've been with Salesforce 12 plus years. I lead the sales and partner technology in the same organization and my role here is to really move us forward on the agentix sales journey that we are on and would love to share here how marketing and sales, the lines are really blurring and how they are basically coming together. So love to share that and yeah, engage here with you all.

Jen's going to start out by talking about our journey with data because long before there was agent force we were on a mission to really be able to create that 360 degree view of our customers and Jen will take you through that. Amar and I are going to talk a little bit about what's happened now with the advent of generative AI, the advent of agent force, what we're doing to build intelligent agents in order to drive business development, in order to bring in pipeline for marketing, in order to nurture that pipeline. And then very important from a marketing perspective is our messaging. How we can consistently from marketing all through sales, all the ways in which we go to market, we can drive consistency around what our messaging is around our products and services and how we want to show up with our customers. To start out with, let's let Jen talk about the customer360.

Thanks Michael. So I'm going to start off with a little context around Data360 and at Salesforce our vision is to enable our customers, our companies to connect with their customers in new ways. What we're really trying to do is allow those companies to unlock the value of their data so that they can connect with their customers, they can deliver exceptional customer experiences and they can really grow their business. And so data 360 is really at the heart of that vision. It's the trusted foundation for Customer360 and for AgentForce, it unifies and harmonizes all the data together across many sources, whether that's structured, unstructured data or real time data, in order to create that single actionable view of your customer. Data360 also enables cloud companies to engage with their customers so that they can across marketing, across sales, across a service, you know, activate anytime, anywhere to give the right message at the right moment, on the right channel. But Data360 is also really, really important in the sense that it is the. It is where we can create real time knowledge and context. Not just content, but knowledge and context to drive the power, power our agents, right? It provides AgentForce the intelligence to make the right decisions so that we can scale our business. So once you have this data foundation in place, everything's possible. So I'm going to talk to you a little bit about our journey. We started our journey at Salesforce trying to deal with our data and creating that unified foundational customer data. Almost four years ago we were experiencing very fragmented, disparate data. We were getting data in from all these different sources, whether it was our own CRM orgs or whether it was our data warehouses, our data lakes. We were inundated with data and we could not create a single view of our customer. For example, we couldn't even foundationally get our customer demand that gets generated on the website and take that demand like when a customer or prospect comes to the website, fills out a form, watches a demo, translate that to a lead and get that to a salesperson. That was taking us over 20 minutes to do. And we were also just inundating the sales team with low quality leads, completely overwhelming them. So we knew we had a very core data customer data problem and we needed to hit it head on. So we embarked on that journey and we built out our data 360 infrastructure. And this is a picture of kind of what that environment looks like from the left hand side. We are pulling data from all types of sources, whether that's data warehouses, our CRM systems, and we pull in 68 billion records, 68 billion signals, and we pull that into our Data360 instance and we're running over 1,000 data streams in order to pull that information in. Now all that data that's coming in, over 70% of it is actually zero copied in. It's using the Data360 zero copy technology. And so what that means is we're pulling in data from Snowflake, we're pulling in data from Amazon, Bigquery, et cetera. And we're actually virtually harmonizing that data in data360. We are not physically moving the data we which means that saves us time and cost. We're also using standard connectors, the CRM connectors, Marketing cloud to ingest data in and then most importantly, we're ingesting all our website data, real time into data360. Right. So all that comes in and then it comes into data360 where it's harmonized and unified and we're creating our unique customer profile. Once I have that unique customer profile, that individual, I can link it to their company information, their account information. We're also collecting all those engagement signals I talked to you about, right? Whether those engagement signals are from marketing, like you guys all attending Dreamforce, that's a signal that will come into Data Cloud, right? Sorry, that was my one miss on calling it data cloud data360 and then things like clicking on marketing campaigns or going to the website and viewing product pages or watching a demo, all that comes in and is fed in as signals into data 360 and harmonize stitched together with that one view of the customer. And then we manage consent and preferences. And all that information that I just talked about is available real time to activate across any channel. It doesn't have to be a Salesforce system, it can be a non salesforce system. And we also take that information and we create insights and predictions things to be able to predict customer behavior and intent. And so then the far right hand side is all about activation, right? It's all about activation across marketing, sales, customer success and agent force. Right. So when you think about that right hand side, to also make Agent Force hum, we also pull in over 700,000 content records across 20 different systems to make that real time knowledge base that I was talking about to support Agent Force. So the Agent Force is pulling both from the unified profile as well as that content base in order to make its decisions. So here's the command center. This is where you see the homepage of Data360, our actual production environment. And you can see here we manage over 1000 data streams. And that complexity of bringing in data objects, tables, et cetera, allows us to have so much data that we can create over 2,200 dynamic segments and those segments result in almost 900 real time activations. And those are things like paid media campaigns, email campaigns. It could be driving insights, real time insights via Slack to AES. But the real core, the most important metric here is that we are pulling together 283 million raw customer profiles from across those systems and then we're unifying them using the unification engine within data360 down to 153 million unique customer profiles. So so it's about a 46% consolidation rate. But by creating all those unique profiles is what allows us to drive personalization and to get it right in terms of a customer experience. So now I'm going to show you what the customer profile looks like in real life. Bringing it all together. We're using Matthew here as our example and Matthew on the left hand side here. This is what we call the baseball card. It's all your real time stats, right? Whether that's your title, your email, your phone, it puts all of your stats in one place. Hence the baseball card. Really that's harmonizing that from all these different sources across what we're ingesting in terms of data. And it also shows you how many different source systems really make up that profile. The stitching together over 300. Then we go to the right and you can see our consent and preferences. So we manage marketing consent within data360 and this ensures that when we're activating, we're activating consent aware segments. Right? So we're always ensuring we're delivering a personalized experience but also a compliant one. Then on the right hand side is the engagement feed. So if you look at the engagement feed it's all about a transactional history of marketing and also website behavior. It shows you email clicks, email opens, it shows you web pages that people visit all real time. And then at the bottom is the trailhead activity. So this is really about the customer's learning journey within the salesforce ecosystems. What badges they have, what certifications they have. So all this can be activated and used real time. This is just a summary view that we give to our different teams. So we take the unified profile and we marry that with all that engagement data I talked about. And basically then we identify the lead. When we have the lead we then do real time scoring of that lead. And what I'm showing you here is actually all the lead routing rules within Data 360. What we do is then we created within Data 360, a dynamic lead assignment engine here that allows our sales team to be more agile so they can go in there and they can change the rules in terms of what they want to do to route the leads, whether that's to a human, whether that's to an agent or whether that's to further nurturing. In this example here on the dynamic lead engine, it's pulling all this data from data 360, right? All the data is from coming in from data 360 to do this. It's for AMIA and the AMIA sales team wants to look at their sec, the commercial sector leads from the education industry and they really want to basically up the scoring of those leads. They only want the best of the best. So they basically say I only want four those that are scored four or above to go to my sales team. The rest are going to go to agents. And so this really enables them to quickly change and modify based upon their business needs. And it shows you the power of how they can assign the leads to either humans or to agents to really scale their business efficiently. So I'm going to hand over to Michael Gonnella who's going to talk to you about our personalized agentic experiences we have on salesforce.com, which are also enabled by Data360. Thanks Michael.

Thanks, Tim. So we're going to do a little demo on the website. I'm going to do a little setup, we'll do a little demo. Everything you see is running on salesforce.com so I'm sure you're all going to be inspired after you see this to get on salesforce.com, talk to the agent and see how it can help drive marketing pipeline. I know that's going to happen. Here's our homepage on salesforce.com and you'll notice we've put the agent up in marquee space right where you'd expect from an eye level perspective to be interacting with it. We're gonna go ahead and do again a live demo. Once we got started with AgentForce we realized this was a huge opportunity for us. We have 1,000 pages of product content on the site, more than 1,000 pages. And that really is the definitive messaging by Salesforce about our products. In tone, in language, in concept of what we want people to understand. This unstructured data, our content is actually our knowledge for our agents. And I'll explain how that works because what we did is we ingested our content management system in order to be able to inform all of our agents how to talk about our products. 84 million visitor sessions a year. So that's an opportunity in order to be able to engage with AgentForce, in order to be able to not only answer people's questions but drive sales. So here's the backstory on what I'm going to show from a knowledge source perspective. We've ingested our content management system. It's now a knowledge management system for our agents. We've ingested structured data around customer references and stories around product attributes like pricing and very importantly digital offers. Because at the end of the day what we want the agent to do is to provide offers to people to bring them down a sales pipeline and convert into customers the agent itself. Taking all of this from Data360 has largely two major topics. That's basically answering questions and providing offers as well as transfer to a human. Because whenever someone wants to talk to a salesperson, we want to make sure we can make that happen as expeditiously as possible. We'll go ahead now and step into a demo of all this. So here's our homepage and as I mentioned, we've gone ahead and made the agent extremely prominent as well as up in the header where we normally had site search. We now have the agent available also to take questions from people. We add pre canned prompts to help people understand what are the kind of questions you might ask and some of the most common things people are asking. In this scenario. Let's say we have a prospect who's come to the site who's interested in Agent Force and AI and wants to know does Salesforce have an AI product for a small business? The agent immersively takes over the screen and is now basically looking at any information we have to answer. And you'll notice that it basically already is taking the qualification that's occurred. It's a small business agentic AI for small businesses. Learn more, learn more. This is the first step of providing offers to bring someone further down the pipeline of learning about our products. The prospect might ask about customer references like what customer stories do we have associated with Agent Force helping small businesses? Here's where we see not only a diverse set of customer stories based on what we've already learned about the person, but because we ingested the cms, the agent can choose to render website content. And this is a little bit of a unique innovation. It's about having the knowledge so that the agent knows how to render. A very important signal is when a customer asks for pricing. And this is where we know we want to continue this process of providing engaging offers, including can we get this person into a free trial into our product, actually using the product and on their way to being a customer. Try for free Prospect Loves that is brought to the page in which they can activate the agent, continues following their journey in the sidebar now in order to still be available, but also to help with activation. This particular prospect though wants to, before they start using the product, talk to a person. And this is where the escalation comes in. What's important to point out here is as the agent is asked to bring in a sales development rep, we're further qualifying information in order to be able to build out our lead record and how we're going to interact with them, but also all of the context of what's occurred, the conversation, what pages the individual has looked at in the session, what we know in Data360 about the individual, this entire interaction is all provided to that sales development rep so they have a chance to quickly understand, maintain context, maintain continuity and basically continue their conversation. In this particular example, Naya joins and is able to provide support for the prospect. So I'll quickly just talk about some year to date results and this is really since sort of late January when we launched this, we've had 200,000 total conversations with the agent, which is very exciting that we're seeing this level of engagement. But what is super exciting is 15% of those conversations drive marketing pipeline and a sales lead. This is without a person getting engaged just from the agent, providing information, prompting people, providing offers. It's basically creating a 15% conversion rate which is extremely high for us in terms of seeing someone coming to the site and converting into an agent. As a result, within this space of traffic we're seeing a 14% year over year lift. Now we're just starting out, we've got a lot to do to optimize this, but for us this is very promising in terms of innovating and exploring sort of this idea of there being a new role on the website to work with our prospects and customers. And lastly, I just want to point out all of this needs to be done through AB testing, experimentation, managing our metrics. We're very fortunate to be able to have Tableau next up and running where not only we can see our metrics, but as you see over on the right Agent Force is with us on the internal operations side, being able to respond to targeted questions around our data and being able to provide us insights associated with how we're managing agentforce now that we have a lead, let's go ahead and talk about how that gets nurtured and the messaging developed further. Amr, thank you.

Yeah, this is the best part, turning all this awesomeness into actual money. First I want to kind of really highlight in this new agentic enterprise, we are seeing the old paradigm really shifting. Instead of marketing and sales being looked at two distinct pillars, we really see them merging into a collaborative growth engine which is all about enhancing the buyer experience, bringing that alignment between marketing and sales really together. And this is especially happening because of the agents. And I would love to kind of now share a few examples internally of how we are doing it. So the first, let's start talking. Let's talk about nurturing generating leads, right? We had a big problem here at Salesforce. I'm sure some of you have this too. Every four leads sent by marketing to sales, only one was getting engaged with for good reason. I mean Jen talked about all the scoring attributes, kind of the rules that we are applying, but we just had that much capacity to go after, right. And that led us to the rest of them just getting ignored. So that is 75% of marketing source leads basically never getting a follow up. That's a massive opportunity lost. And just to give you a sense of size, this amounted to about 20 million over the last 10 years for us. Right. And we set out to fix this with what we call the lead nurturing use case lead nurturing agent. So agent does what you would expect, right? Basically able to personalize email outreach. It's respecting the opt outs. It's, it's able to schedule meetings and really get that back to basically a human touch when needed. Almost working like an extension of our sales dev team. And just in the last three months this agent has internally reached out to 80,000 leads for us. Those were otherwise gonna just get ignored. So imagine that. Right. And that has resulted in 8 million of pipe for us which again it's all about really driving that growth. That is growth being delivered. One important thing to really understand about this agent, it's not like a simple email drip that's happening. This is two way conversation. Jen showed the amazing unified profile, the data360 it's really getting powered with the magic of that to gain the context being grounded in the CRM data already. And, and with all the marketing information it has, it is able to go very targeted, personalized touch with the buyer and that also keeping the brand top of mind. And this is a great example we wanted to pull up from our real world Production. There's a prospect here you can see. Did not find the webinar to his satisfaction. But instead of the agent just doing a very direct standard follow up, it empathized a little bit. He understood his frustration reflected back and basically went back to him with the exact information the buyer was looking for. So this is really turning a bad experience into a sales conversation. Now that we have a lead, we really need to progress the deal. It's basically about coaching, right? So the next use case is about taking that lead to more of progressing the deal. And that's where we are trying to really get into the message integrity of coaching. And this is what does the coach agent does for us. It's really getting into making sure that you have on your tips all the sellers that we have. 15,000 of those.

Right.

The coach is trying to really scale that messaging with the help of you can practice the pitches. You can go about using the sales coach for certifications. We have done till date 15,000 plus sales certifications using sales coach. Let me show you with a demo here how simple this is. I'm gonna highlight three things here with the sales coach. So first thing is it's embedded in your sales cloud itself. So you actually have this in the opportunity page. The seller can go and just practice the pitch here. She sees it. There's a one on one conversation coming up. Bang. You basically start practicing the pitch right there. So the second big thing is the coach is really built on the marketing content, the how you do value proposition. How do you want to really get the specific product knowledge out there? So it's evaluating based on a rubric that's coming from marketing. So that's another way of how we are merging it. And the third big thing about this is instant feedback. This is one on one feedback that you get with personalized recommendations and actions that you get right there for the seller. So areas of improvement, how did you kind of bring up the points? It gives you certain actions that you could be taking, maybe go take another trailhead for being more targeted, more personalized. And it's also taking the customer context into play here in terms of how that pitch comes through. It's as easy as that. 1, 2, 3, and then you can practice more. What we have observed is that our sellers who are using Coach are really seeing a 10% uplift in converting the opportunity. So that's how again, the growth aspect of it is coming, by using the marketing content to actually make sure that the integrity is maintained. And this is a third, one of my favorite Use cases. It's really about taking all the amazing content that marketing has created and activating it throughout the life cycle for sales. So this is our sales agent right in the flow of work for sellers throughout. So you have on their tip intel that they would want for say a competitive information or you're looking for certain content. You don't have to worry about where marketing has all the amazing first call deck stored. Bang, right there. The agent in Slack can actually respond back with specific information that they're looking for. And it has again as the Context of the data 360 because everything is consolidated. You can see the amount of assets that we have loaded up in Data360 for this. It can answer product questions and imagine the amount of changes that are happening in product for us, the innovation that we are driving all the time. This helps sellers be really ready to actually be prepared and informed in order to have that best conversation with the customer. Bringing all of this together, you can see the impact by the numbers. I spoke a bit about this before as well. You have the lead nurturing which is really discovering the hidden pipe. Right. And all the numbers you see here, Sales Coach is about message consistency, making sure that you have the ability for the sellers to do that one on one as well as mass coaching. Like I said, data360 is a great example. We did pitches basically training for all our sellers through sales coach and the sales agent which is really driving a lot of productivity for our sellers. Because one example that you'll see, I hope you all will make it to the customer. Customer zero keynote that's happening right after this. We're going to demo the point of view which is a much more exhaustive, informed point of view that is for an account and actually helps save so many hours for our sellers. Right. So that's again time saved till date as we are tracking 200,000 hours saved because the same work that you need to do to research to find that content, you would be spending easily 15, 20 hours in a week to be that more equipped for that customer conversation. So this is growth being delivered hand in hand with marketing content enabled for all the sellers. With that I'll pass it back to Michael.

Thanks Amit. Okay, so let's talk about what's next. We are obviously just starting out on our path and we talked about some of the agent force use cases we've put together. We talked about the website, agent, sales coaching, et cetera. There's a lot more across marketing, sales and service and all of the ways in which through digital channels and in person. We're connecting with prospects, customers and trailblazers as well as over on the left on our internal operations, how we're looking at implementing Agent Force and to drive operational efficiency and effectiveness. There is an unlimited number of use cases we could talk through. I'm hopeful that some of you have had a chance to experience our events agent on the Dreamforce mobile app which is built with a body of knowledge around all of our sessions and speakers, the campus and gives you a chance to interact with that. It was actually our first outward facing launch at Dreamforce last year, but obviously we're in the process of really building out our own agentiq enterprise. Some of the things we have coming up right now is we're working on an agent which looks at our website content and drives improved SEO and generative engine optimization through backlinking, crosslinking, making sure that we're getting our best optimization of our website content and also pipeline management in terms of how we're managing our forecasts. So so many different opportunities that we're getting a chance as Customer zero to innovate with and we'd love to talk with any of you about these if you're interested. So we'll just take a moment to talk a little bit about like some learnings. Not to dwell but just to say there are a few things that we noticed that we just want to touch on. Everything was changing very quickly. Our products are still growing, emerging and evolving. So there is no like big project. It's kind of like get something done, learn from it and then move on to the next. We're finding that things like our traditional ways in which we would do technology development are not as applicable in this space as ways that really more models, sort of the development of machine learning models or decision sciences practices. The right data is gold. Thanks Jen for taking us through Data360 but it's also not just data, it's the knowledge that we're creating from unstructured data as well as structured information across the enterprise. As you saw how it powered consistent messaging across many of the use cases that Amar and I talked about. Multivariate testing, AB testing, experimentation is the only way to really understand what's going on and we have to measure that impact all the way down the pipeline, not just in terms of agentic engagement or quality of responses, but how then we're seeing it impact, leads, opportunities and the quality of business development that's occurring and obviously embracing new ways of working. It's something that we're all like in the middle of figuring out what it means for each of us in terms of how we can help our teams as practitioners in this new agentic enterprise. Thank you very much for taking time this morning to be with us again. If anyone has any questions, please raise a hand. We have mic runners. Or we'll hang around a bit and we'll get a chance to talk. Thank you all.