Agentforce for Financial Services | Dreamforce 2025

We at Salesforce and myself personally, I tend to be an AI optimist. And what does that mean? It means that AI has this incredible opportunity to truly transform how we work, make it better, make it more efficient, how we serve our customers. The really important thing to remember here is humans still play a very critical role. And in fact, what agentic AI is going to do is let humans do what they do best. Humans will build relationships while agents are going to do the busy work. Humans will build trust while agents will do the work of assuring compliance. We're a highly regulated industry in financial services. And then humans are going to be able to bring more empathy to their interactions and their relationships with their customers, while agents are going to be able to bring scale. And you're going to hear that really come to life in our Absa story. So last year when we announced AgentForce at Dreamforce, we challenged our financial services product team and we said what if we could help all of our customers quickly deploy agentforce? But not just any Agent Force. Agent Agent Force that has financial services context. So agents that already have that built in financial services expertise. And that's exactly what we did. We launched AgentForce for financial services back in our summer release so a few months ago. And you know, at Salesforce, at Dreamforce, we love to take the opportunity to change names up a little bit. So the thing that you used to know as financial services cloud, we're bringing that together with AgentForce and we now have AgentForce Financial Services and we've had 29 releases on what we used to call financial services cloud since 2016. And what does that mean? There is a lot of that context. Financial services data, financial services, business processes, financial services. Now agent templates are all built in. Another thing that's really special about it, it's open and interoperable. We know that Salesforce is not your only application, although we think it's certainly a great one. But it's not the only one you work with and our agents can work across other applications. It's an open, interoperable ecosystem. And last but not least, it is embedded in the flow of work. You do not have to stitch it together to bring it to your teams, to bring it to your customers. It. It's natively embedded within our beautiful Salesforce platform and all ready to go. So with that, I am gonna welcome first Pato Matalong to the stage. Welcome. Thank you.

Thank you.

Welcome, welcome.

Thank you very much.

Can you start by just give the audience a little bit about who you are, what you do at APSA and what was it like to launch your first AgentForce agent?

No, thank you. And thanks again for the opportunity to be here. It's my second time at Dreamforce and it gets better every year. I run technology for the private of personal and private wealth business and I'm also joined by my colleagues from business banking over here, head of data science, colleagues in the room. Our journey with Salesforce has been goes back four years or so. We've done a whole lot of different things. But now today want to talk to you about two use cases. And it's not just about the use cases, it's about the opportunities that it provides in terms of scale. The one use case is in business banking where it's really about documentation to our frontline colleagues and some of the frustrations that it removes in dealing with customers and servicing customers. The second one is really an email to case which is fully automated, no human beings involved at all and the opportunities that that also provides. So that's what we want to share today in the room.

Fantastic. So with that we're going to get to the demo and here it is. Here he is, the best mohawk in banking. Ben Altom. Take it away.

Well, you all usually get to see a demo that I've built in this situation, but I'm really excited to be able to be here with Toddo today where we're going to go in and actually look at their implementation. We're going to look into their org, see what they've built and see where we're providing this power to the bankers. Now what I really love about what we've put together here and framing up this story is ABSA is really looking and viewing their retail bankers as superheroes. And what I think is really interesting about superheroes is they don't want to be superheroes. If you follow any of their stories, something's going to happen, A force comes in where that superhero has to step in and save the world. They don't know how to do it, but there's a guide that steps in that helps them along the way. And what Absa has done to put together their implementation of AgentForce is they are using that agent as the guide to help their heroes. So as we look at this today and start to envision how this experience is happening, our hero here at ABSA is going to be using these agents for pre meeting, for post meeting wrap up and to be able to see where that agent steps in to help the customer as well. So with that, Tado, go ahead and let's Talk about what you guys are doing starting with those pre meeting notes right there.

No, absolutely. But I love your haircut by the way. So I really like us to paint a picture. APSA is a bank that comes from four banks that were actually put together. So you can imagine technology capabilities come from those four banks. And in the eyes of a banker there's a whole lot of frustrations is actually putting data together to be able to have some sort of insights before we engage customers. That takes a lot of time and today we want to show you somewhere an example of how we actually been able to take that frustration out of the customer. In this particular case it's really about meeting preparation which really starts off with a new meeting with the suggested agenda for that particular meeting and the basic stuff about the dates and what to but really the most important part is really about this particular piece. In here we are able to have a lot more insight about the company that is going to be engaged. And you also see the individuals that have ownership in this particular business and we can also assess whether these owners also have any relationship with the rest of the bank. And it just creates a whole lot more opportunities for us. Now Ben will run quickly to just to to save the actual pre meeting as you see in front of you now consider that has now been saved. Now it's important now to bring in a relationship assistant into the picture. And what we're starting to see here is that we have the opportunity to have some sort of interaction in preparation of that particular meeting and what some of the products that we are offering and some of the product that the this particular customer do not hold with us. In that way you are able to have a far much more meaningful conversation with the customer. This is just an example. There's a whole lot more insights that we have an opportunity to create to be able to have that meaningful relationship. Now the likes of Gavin has really worked hard with us to really ensure that we can actually start consolidating data curating it so that we have a lot more insight for conversation with our customers. This is a far much better experience from what we used to have. There's zero time spent in actually doing this because that is all been prepared for the customer.

Toddo I spent 15 years in retail banking prior to coming to Salesforce. I would have killed to have some of this stuff specifically that pre meeting. The amount of time and effort that it would took to prepare specifically for a small business meeting was unreal. But being able to not only have the agenda but then have those pre meeting notes, have them saved and then have an agent that I can talk to. That's a big deal. But we're talking pre meeting here. Yeah, the pain really comes post meeting because listen, I'm sure a lot of you in this audience have said the statement, if it's not in Salesforce, it doesn't exist. That's a big deal. And that post meeting wrap up, the bankers don't like to do it. Now what I thought was awesome was what you guys did. Looking at post meeting and simplifying that for the banker. You want to take us through this?

Our adoption rate of Salesforce utilization in the bank, as she will tell, is over about over 80%. So therefore if it's not in Salesforce.

Yeah, exactly.

It doesn't exist. Now you have an opportunity here to experience a journey of a post meeting whereby there's particular nodes that has actually been captured. It so happens that the banker captures these nodes. They are scripted on a piece of paper. We have 11 languages in South Africa. You can use any of those and we can still be able to help you. And here is an opportunity to upload those notes. And as we Ben runs quickly into the analysis, you will see that the details of the description of the notes have now been translated into English. Not only that, they suggested actions that one can also consider for further engagements with the customer going forward.

Now, I don't want to steal thunder from my theater session I'm doing tomorrow, but one of the things that I love here is the learning loop you create with that. Because it's one thing to be able to just have the notes and it's saved for the account team. But when you have that and it's stored like this now that data is there for other parts of the bank to be able to utilize. Are you all using that data in any way to create that learning loop? How's that being leveraged right now?

Tato, as I indicated, our head of data science is in the room now. Think about thousands and thousands of information coming through. What Gavin is building for us is opportunity to personalize sales for frontline teams. And there's some of the new opportunities that we're starting to see and that we are able to actually leverage this capability from all the interactions that we have with our customers.

That's fantastic. This is one of my favorite features that we've seen, to be real honest. Now let's switch gears a little bit and talk service because as we talked about sales for a while, we know that that's obviously what we're trying to do, but the biggest time suck comes from service and those bankers, either in a branch or over in a call center, having to be able to adapt and have to change based on that request that comes in. Now when we look at that, I know you guys have spent a lot of time looking at case deflection. How are y' all answering the call to free up that time for the banker on that front?

In this particular use case, what we would like to share with you, it's a problem statement whereby there's about 35,000 or so emails that come through, about 39,000 calls that actually come through for service. And when the customer calls a bank, you know that there's some issue and therefore they're not just phoning for to check in. There is some issue and usually these customers are in distress. In this particular use case, we're talking about a transaction dispute whereby there's a debit order that has gone through and that the customer is disputing that. Now you're seeing an email coming through to the bank. There's some sort of an interaction with a customer on email about this particular case. And eventually this within a minute, this particular case, the debit order has been reversed and the case is closed. Now that used to take a whole lot of hours to be able to get to that outcome. Now and can you imagine whereby some of these calls are actually dropped and actually not going through? Right. And we just extending this frustration with the customer, within seconds, that particular case is resolved.

I love it. Now continuing here, I know you all did a lot behind the scenes on how you actually built this foundational piece. So when we look at this, I know we're looking now at some of the agents that you guys were able to build. There's a lot. It's not just looking at those transaction disputes. You want to take us through here a little bit and show us how you built some of this stuff?

Chihuahua runs also the middle office of the bank. It involves collections fraud. In this particular case, service jobs. These are different functions that we run within what we call retail banking. And across all of these, we actually worked with colleagues in those environments to create a whole lot of use cases and actually start building agents that support their particular functions. And this has been a collaborative work between the colleagues and the technology teams to actually build these agents. Just throws opportunities that are there to extend the capability going forward.

And you guys used Agent Builder to do a lot of this, right?

Yeah.

So when that information came in, you're using the Standard instructions, standard actions, they're calling those flows. And we're looking here at how you all handled some of this. Tell us a little bit about the agent that you built and some of those actions that were leveraged on the screen.

You're actually seeing on the left hand side just the classification and the description of that and how that is translated in the middle into a particular flows within there. And you will see that also in the next grid.

So those emails that went back and forth there, it's using the flow behind the scene to power that agent. So when it hears what comes in that email, what's going to happen is this is orchestrating that process, Right. So it's listening for that response.

Absolutely.

And then as in return it's going to fire back exactly what's needed to up step in that conversation to get to your resolution, right?

Absolutely, absolutely. But this is not the crux of it. Ben.

Yeah.

Now I think it's important that we talk about the impact because this is where it really, really matters. And you will see on the next slide we're starting to show the impact that this actually brings to our frontline. And as we said, the idea here is productivity gains. It's about efficiencies and is working with colleagues so that they can actually do a lot more tasks and activities that actually add a lot more value. And you'll see on the right hand side, we're starting to see that the actual handling time reduced to one minute. You are seeing an increase in colleague capacity and also create a whole lot of capacity to be able to create leads and sales opportunities for us. That's what matters. And I think using technology and capability on Agent 4 to be able to create this opportunity for colleagues, that's really what we're trying to achieve.

Now this quote is what really stuck out to me right here. When we start putting dollar amounts to this value, can you tell us a little bit about what hitting that 30% of cases did as far as the impact on the bottom line? Yeah.

We also built some of these capabilities on our mobile app so the customer can self service. But in this particular case, these are customers that choose to engage with us through our contact center. And this is now you're starting to see the real impact and commercial outcome for the business in terms of the results coming through. But this also present opportunities to do a whole lot more. And this is part of the next phase of our journey to actually scale these use cases into other functions that we've spoken about earlier on.

That's fantastic. So in what we've seen today here, you guys have done an amazing job empowering these bankers. And if we think about the superheroes and the guides that have, Luke Skywalker had Yoda, Marty McFly had Doc Brown, Moana had the ocean and her grandma, they always have a guide. And here at Dreamforce this year, our retail bankers have Agent Force. And you guys have done a phenomenal job being able to paint that picture on how this comes together. So as we wrap up here, what I'm going to ask you all to do is go ahead and get your phones out and take a picture of this screen first. Your QR code will give you the guide for everything that you need for the day. But in particular, to build off what we talked about today, make sure that you guys are in the keynote tomorrow, and make sure you stop by the BWAM theater session, because we're going to get deeper into data. We're going to show more of what the agents that we looked at today. Actually, the demo, I'm doing a theater session tomorrow, I ended up building after seeing this awesome stuff that Tado built. I showed them just before this. So come hang out with us tomorrow, see more of this. And if you want some coffee on us, scan this QR code and scan it. And as we wrap up here, if you have any questions, anything that we can talk to you all about, we're here to chat and we'll see you around Dreamforce this week. Thank you all so much. Toddo.

Thank you, bud. Thank. You.