3 Service Cloud Secrets You’ll Wish You Knew Sooner | Dreamforce 2025

I want to begin by just clarifying something because there's a lot of confusion, right? Three Agent Force Service Secrets Service Cloud Agent Force Service. Einstein for Service Agent Force for service Agent Force Service Agent is the thing you would talk to the autonomous agent you'd talk to if you were having a chat conversation with an autonomous agent. So Agent Force Service is service cloud or Service Cloud as AgentForce for Service these days. But just I know there's some confusion around that I want to clarify. But the idea here is we're going to talk about three things. You're going to hear a lot about Agent Force. You're going to hear from Avi about how they're deploying AgentForce. There are other things though that you should be doing to set up Agent Force for success and there are other things you should be doing with AgentForce service to succeed. And we're going to talk about those also. We will cover the roadmap at the end. So there is a forward looking statement. Salesforce is a publicly traded company. Please make all purchasing decisions based upon products that are currently available. And thank you, thank you for being here. Thank you for spending some time with us today. I'm Drew shieldvector. I'm a service cloud solution engineer. ABHI will introduce himself when it comes to time to pass the mic to him. But very briefly about myself. I am a Salesforce solution engineer. I also have worked in implementations for contact centers, predominantly in regulated industries. So I like to think I bring a little bit of a perspective that is honest and realistic about how Salesforce products actually perform in the real world. This is our agenda today. We're going to cover the challenge before us. We're going to cover those three secrets that you've been waiting so long to hear about. And we're going to talk about Zoda and what they've done to be successful. And we'll close with the roadmap and some calls to action. So talk about AI exploding with complexity. This is a second career for me. I previously had an earlier career shooting, marching and digging. And when I joined the software world, predictive AI was the leading edge of what we could do, right? Look at a set of data, decide what the next most likely step will be. Learning all the new things that come into AI is a full time job. So for a variety of people who are doing another full time job running a contact center or they are running a sales operation, trying to learn about all the things AI can do is quite a demand. And people honestly are having a hard time keeping up with all the things that they could be doing with AI, I think all of us as consumers, when we think about service, I always think about myself as a consumer. And I'd like you all to think about yourselves as consumers. Your expectations have grown tremendously. It's no longer just the Amazon experience, it's the everything experience. Everything is easier and you expect every interaction you have to be easier. And for industries where services sometimes laggard in technology, it's really obvious that people are not keeping up with all the tools that could be available to them. So the challenge is growing more and more for people to deliver good service. And what we've realized that Salesforce isn't really just a technology problem, it's also a human problem. I was recently in a big contact center for a company I won't name, but it's a health care provider or healthcare payer, rather an insurance company. And a woman called in looking for an in network gynecologist. The contact center agent had in front of him a tool where he could search all of the information that he had at his fingertips. It had a text inpace input. He could ask a question, it would give him whatever answer from all the data they had available. It took him 20 minutes to solve that call. Do you want to know why he didn't know what a gynecologist was? So when we talk about equipping AI, you have to also think about equipping it like you would equip a person. You have to give it the information it needs to make the right determinations, not just having fancy technology. You have to equip it with the things you expect it to need in order to do its job. And that's really where AgentForce for Service comes in again. Agent Force for Service previously Einstein for Service is the suite of tools that we've added to help service providers do the things they need to do. So Salesforce is the best in class service platform. We are the market leader by far and all the things that we are doing. Rather than reinventing the wheel and adding a new tool that someone may forget to use or or may not want to use because it's slow compared to their original process, all we've done is we've added AI into the processes that people are already using successfully. And by doing this, it helps AI and humans to work together to be more effective and more efficient. Without reinventing the wheel, we've done something good already. Let's add AI where it makes sense, when it makes sense in context. So Agent Force for Service really comprises all of those things. And there are a few things called out here at the bottom, but I don't want to linger here. I want to jump into those things that I specifically want to call out. The first one is knowledge, and knowledge is huge. Knowledge doesn't just mean Salesforce knowledge. It means using all of the knowledge you have at your fingertips and empowering Salesforce with that knowledge. So when we talk about Salesforce knowledge specifically, there are a lot of tools that you'll hear in other sessions about here. But the key thing is you have to get the information into the the platform so AI can benefit from it. So that means in some cases leveraging data cloud to bring in other disparate sources of knowledge into Salesforce so it can be used and brought into context. You can do this with Agent Force service agents, you can do this inside the console. But if you don't give it the knowledge, if you don't tell your agent force service agent what a gynecologist is, it won't know, right? So ultimately, if you are thinking about embarking upon AI, you also have to get your house in order in terms of where your knowledge is and, and how your knowledge is connected into Salesforce. Next, we talk about Omnichannel for a long time. I think people most often immediately thought about routing in the context of other systems connected to Salesforce. The external routing option was often flagged. However, Salesforce routing and in particular Omnichannel and Enhanced Omnichannel have made tremendous gains. So you have the ability to route from any channel to the right resource. And, and then those resources have the Omnichannel inbox that allows them to manage all those pieces of work. And all those pieces of work are now organized not by tabs, but by status. Or they can be. So you can say, hey, I sent an email to my customer. I'm waiting for them to respond. That shouldn't count against my capacity. Let's pause this work. It's still here in my inbox. I can deal with it later, but it's not preventing me from receiving other pieces of work. And then the ability to blend human and autonomous agents together and see the their status across all your whole enterprise at once allows you to really be efficient about the way you're using all of the agents inside your enterprise. So I want to show a little bit about how Omni channel works, because I think this is something I spend a lot of time talking to customers about, but incoming work across any channel and I want to be clear about That, I mean, any channel you can connect to Salesforce, someone may have some channel that you know, can you connect to this one? There are a variety of channels you can connect to Salesforce.

Right.

So all of these channels coming in are then assigned to a routing flow. And this routing flow can bring in CRM data. So it's more than just, I think this piece of work looks like this, and this is where it should go. It's. This is a piece of work from someone I know, I know they own this. I know this is what their case history is like. You can incorporate all that CRM knowledge into your routing decision and, and route it to someone who is in the right queue or who has the right skills. And in doing so, or even route it to an autonomous agent. In doing so, deliver that work very quickly to the right person. When it comes to routing it to the right person, it's routed through a capacity model that's shared across all the different channels. So you understand, like, maybe I could take a phone call if I'm working on an email. And that shouldn't prevent me from taking the phone call because it's a higher priority of work. Right. I want to see all the things you're working on and I want to be able to route them with a full understanding of that. And then lastly, the ability to escalate and supervise this work. So if it goes to autonomous agent, and the autonomous agent is fumbling the conversation, I want to be able to see that and take it over. As a human supervisor, I want to see the red flag from the agent in my omni supervisor and take it over. If I am a supervisor that I can see all these things coming in. Lastly, Service Rep Assistant. This is a very powerful tool that was recently added over the last year and it's going through some great changes as well. So I'd encourage you to attend sessions about that. But essentially what Service Rep Assistant does is come up with a way to generate a plan for every piece of work you receive. So you receive a new case. And Service Rep Assistant will say, based on knowledge, based on topics you've given me. Because Service Rep Assistant is an agent, it's an autonomous agent. Here are the steps I think you should do to solve this case. So you think about next best action. Next best action is this thing has happened. Here is one action. This doesn't necessarily make that obsolete, but this is a more complete picture. So it's, hey, you have this type of work. Here are all the things you should be doing. One of those Things might be completing a next best action or you can actually complete actions inside Service Rep Assistant now as well. This is going to become dynamic Service Rep Assistant, which is what you could attend sessions about while you're here, where there even more context is going to be added into this feature. But this is a tremendous way to create consistency and expertise across your service organization. Where people may lack the tenure or the specific expertise in one topic to know exactly what to do, it helps you decentralize processes and plans for all the types of work. So we've gone through the part now of my little prologue of those three secrets and now we're coming to the most important part, which is hearing from another customer like you. And Abe is going to talk about the success that Zoda has had with autonomous agents and I'm going to turn it over to him.

Thank you very much. Thank you. I'm very happy to see that we have a full crowd today and there are no more invites. So very happy that you are here. And it's also broadcasted. Let me start and begin with the first secret. I think let's dive right into it. So my first secret is in order to have an agent force, you need to have a license. This is a secret, right? So bear it in mind.

That's not a secret. We tell everybody that.

People tend to forget that it's required. Anyway, that's my first secret for today. Nevertheless, let me tell you a little bit about our company, our story. Zota is a payment gateway hub. We'll get into it a little bit further. But my story begins actually in my previous employer, who was a well established company, worked for a couple of decades, maybe a decade or so, and decided to go listed. They didn't have a proper CRM back then and decided obviously to add Salesforce as their CRM. I saw what a complex process it was for an established company to actually move all their data and the data sets into Salesforce itself. And I realized that if and when I will do establish my own company, I will start with Salesforce. So our story begins actually with the company that we actually established, Salesforce, before the first employee came. So we had all the protocols, licenses we've built with partners, great partners. The methodology, how to open an account, how to communicate, and so on and so forth before the first employee came. So whenever they came, there was no going around the processes. And in the end of the day, it led us to be a very lean team. So maybe I can move forward to what we do. Xota is a global payment Processing payment company. We are kind of marketplace for payments. On one side we have merchants global businesses around the world, mainly in Europe and the western countries. From the other hand we serve end users the ability to process money from emerging markets mainly in places like Southeast Asia, Africa, China, Japan and so on and so forth. Where actually credit cards are not the most common payments leading payment methods. They have QR codes, instant bank wires in China, for example, Union Pay, Alipay, WeChat. So these are things that are not being supported by the traditional acquire banks payment system, Stripe and others that you are connected with. But we need to find us as a payment company, gateway payment hub as a marketplace. We always find the local providers in the countries that can support global businesses. And basically our product would be a cashier page for end users to come and engage. Doesn't matter where they are around the world. They can come from Brazil, Thailand, Vietnam. And they will see the local process of making payments, currency, the local payment methods and so on and so forth. So we have decided to build the whole infrastructure of the company based on Salesforce as a single source of truth. We process currently today almost 10 years old a company. We process a couple of billions of dollars every year and managed to get to almost 360,000 cases every year internally from email to case communications transfers that we do. And basically we backlog every movement and action that we do as a company on the Salesforce. So that was our starting point which was really really good. Nevertheless, it's coming from up to bottom, meaning me, I'm very dedicated to this process. I believe in the 360 view. So everything should be connected to somehow to a single source of truth. For example my go to besides the LMS and learning myself and then explaining everyone else what they should do or where should we look for more information on how to do it better. Besides that, we also build the entire infrastructure on the Salesforce itself. If we don't find anything that for example some kind of need that a manager will come to me, I will first thing ask him, did you find or did you look at the App Exchange, the Salesforce AppExchange? Because there are so many things that we can use. And the first thing we go either Salesforce Lab or a really good third party vendor that already from the community that already enabled such technology. And we can solve this need internally. We also have the Salesforce Direct license which in some rare cases you mentioned there is no some platforms. You cannot really connect the rare ones. So give us the API and we can connect directly from our Salesforce environment to them. And we did it a couple of times. It looks beautifully and it empowers our employees. So company the same scale as ours, volume, transactions, number of queries a day would be more than 1,000 people. We are only 200, 250 people in the company licenses. So also the headcount is very, very lean and mean as a company. So it's managed by doing all of these investments priority. Before we started the company automations and connecting everything it really allowed us to empower each employee and to stay and learn even we grow very lean team. So that was one of our my goals. I don't see any target by growing employees only if you need them. So we do grow. And I must tell you when just before the agent started. Second secret in order to have agent you need to have data cloud. So it goes together. So data cloud would be our first stage where we started way before or maybe a year before AgentForce was launched last year in here in Dreamforce. So when we realized that AI is coming and it doesn't matter which model you want to use, it can be your own and I urge you to seek what the data cloud as a platform has to offer. You can run your own data on your own AI models on your own data based on there and actually you can customize your own AI model. You run or let's say choose all the fields that you want to be carried and the model will tell you you can come back in a few hours and I will make the AI model for you. So the data cloud by itself is a really, really powerful tool. And I think we haven't really tipped it, only the tip of the iceberg over here. Besides that I understood the first thing we need to do is to connect all the data that we have as a company. And I thought all the data is in Salesforce already. But I was really, I was wrong. And when we realized the first thing we did is actually to understand what we have as data sources, which data sources are important, relevant, more influential. It can be from social media, marketing, can be from a website, analytics, Slack. We have a lot of manuals and guides that are leaning into the departments, Google, Shared Drive and many, many other resources that were not actually on Salesforce. And bear in mind we used a lot heavily on Salesforce, all of the NMS knowledge articles, they were there already but still a lot of data weren't there. Grubs, water, sorry. We started to look for use cases, general company use cases. Once we had some kind of data connected into the data cloud. And for me it was very exciting because we have Salesforce and we have our core technology, payment technology, how we process the transactions, but they're never connected with each other. And the data cloud was really natural place for us to connect all the data from all the different sources including our two pillar core systems that we use. Once we did it, it gave us a lot of insights. And running modules, for example for risk related or fraud prevention in our industry is a very insightful and powerful thing. The first agent when agent first came in, I realized instead of company level wide level use cases, now we can break it. This is the time to break the knowledge or the queries into thousands or daily little prompts and use cases that each and every employee will come up with. And we don't know this and this is way more powerful than company wide use case. The first agent that we ran was actually on our website FAQ agent from a company like ours which is bank like services. The customers are B2B Enterprise Sales. Each customer we can hold a couple of millions of dollars on a daily level at each and every moment. So for us the website was actually soft launch something without risk or less risk. And also we have seven different financial licenses around the world including uk, Singapore, Hong Kong and so on and so forth. Switzerland. It's very important for us what the agents, whether it's representative or digital, will say, what kind of answers they will give. One word important for us from regulatory point of view, from AML point of view, anti money laundering, whether we enable payments or provide payments in certain countries, it really matters, especially for the legal team in compliance. So it was very important for us that the answers would be highly accurated and will serve us in the best way possible. And we could not allow ourselves, let's say, an answer that was not within the standards. What we've seen after we launched the first agent, which for us was not only a POC, but also was a tool to engage, to create engagement between all the different employees. So we have a great results of the first POC, 77% more time we reduced from B2B point of view the engagement. The first engagement time from 12 hours to a couple of seconds. And the level of answers that the agent was able to bring us was very, very high level. Okay, so when we brought all the teams and I think the second or third secret here that it's more of a human resource change rather than technological change. So you need. What we did is we brought the best people and the core team, the People who run the show the call the shots to be the agent managers because they know best what's going to happen. So we defined 50 different use cases in each and every department. What is our, let's say business value and how complex it is in terms of data. And data is the key here we have today. If you bring good data, you you get good results. You bring garbage data, you get garbage results. What you call gigo. Garbage in, garbage out. So very important how we start. We start with low complexity and the low implementation complexity with high business value or with business value. So we were looking about the problem first, how we can solve really the problem in each and every department. And I believe there is no such thing. Too many agents and the managers. Each manager, the new era manager will have couple of agents running under him before they will have employees. So there is no point to bring a new employee before you can actually show the use case of how you try to do it via engine at least and what it means. It leaves us more time for the employees so it empowers the employees. We have today internal agent, an LMS agent, an agent on our slack to summarize all the threads, we have an agent for a meeting summary before a meeting and after meeting. So we believe in like next year at Target, 50 different agents all across the company. So that's where we go.

Thank you, Abe. Every customer I've seen be successful with agents has done like Zoda has done. And it's easy to deploy in agent, but a lot of rigor should go into planning that deployment. A couple things I want to leave you with on the roadmap. Like I mentioned, Service assistant transforming into dynamic service assistant. In other words, expanding just beyond cases into messaging into more responses out of service assistant. I would definitely try to attend some service assistant theaters while you're here. So that's really the biggest thing I want to leave you with from the roadmap. And then lastly service keynote tomorrow, 1 o'. Clock. So if you want to hear about all the latest things happening in Agent 4 service directly from the the mouths of the several horses will be presenting, that's where you should be. Thanks so much for coming today. Appreciate you.