Salesforce, Inc. (CRM) Earnings Call Transcript & Summary

Good morning, everyone. Welcome to today's session, Unlocking Unstructured Financial Data for Deeper Customer Insights with Agentforce. Thank you all so much for joining. My name is Natalie, and I'm on the Corporate Marketing Team here at Salesforce. Before we begin, I would like to cover a few quick notes about our webinar platform. Today's webinar will be available on demand after we wrap up and will be accessible through the URL that you're on now. Please note the slides will advance automatically throughout the presentation. [Operator Instructions] We'll get to as many questions as we can at the end of our session. And with that, I'm going to hand it off to Mohith to get us started.

Thank you, Natalie. Hello, everyone. Welcome again to our webinar on unlocking unstructured financial data for deeper customer insights with Agentforce. Now before I begin, I want to make sure I remind you on our forward-looking statements. Please do not make your purchasing decisions based on anything futuristic in this webinar, always make sure that you make your purchasing decisions based on the features that are currently available in the product. With that out of our way, I want to take a moment and thank you. Thank you for being customers, our partners and members of our community. We hope to make today a valuable use of your time and definitely, your success is our priority. And we are committed to being your partner and driving customer success together in the days and years ahead. So it's one of our core values, which have always led us the way. So thank you again. With that, my name is Mohith. I'm a Principal Developer Advocate here at Salesforce. And here's our agenda for today. So today, we're going to take a look at what is unstructured data for financial services. So we'll spend some time understanding about what does it mean to have this unstructured data? What is the value out of it in this AI era? And then we'll give you some examples of how you can use this unstructured data to activate it. And of course, we'll introduce to our product, Agentforce, for financial services. There will be product demos. And then we have a wonderful panel who will discuss how to accelerate and ensure success with Agentforce for financial services. And then finally, I'll share some resources that you can use to stay connected and informed. All right. So moving on. So we have heard it multiple times over this past year that the future leans on AI. So here at Salesforce, we believe that the future is AI and humans working to drive success together. So that's why we have introduced this Agentforce platform. So now whether it is humans gleaning insights to create these personalized experiences or these agents becoming the first line of response and customer success, each and any of these experiences are built on reliable, relevant and trusted data. So you know what, this data foundation is what that brings people and AI together. And data comes in various formats, structured and unstructured. And just to give you how important this is going to become, right? I have this slide. And if you take a look at this slide, right, it shows all the AI eras here, and we are in this third wave of AI revolution where we are working with the agents or we'll be building these assistive and autonomous agents. Now as we move up this AI wave, data will become a critical ingredient for success, as you can see from the slides. And you know what, we all laugh if you have an assistant and that makes a mistake, let's say, or hallucinates. But in enterprise setting, right, you are going to rely on assistant for critical metrics, like, for example, it could be next quarter growth of your sales quotes or sales quota. Or it might be something that is very important to your customer. And so your brand value is at reputation here, right? So definitely, getting it right is important more than ever. And it's not just about the structured data. As you can see in the slide, 80% of the enterprise data is unstructured. Think of your PDF, knowledge article, notes that you have written. So for a human agent, right, they have access to multiple screens. And they can look at these screens, they can browse like multiple sites and answer. But for agent to be effective, you want to have this unified data for the agent, so that it can access all of that data, it can read through it, it's accurate. And one of the challenges with this whole structured and unstructured data for all of the enterprises is all of this data is trapped and it's complex, and it's difficult to activate without tools. So these data are in silos today. And if you have built any agent or you're working with agent technology, you know how critical it is for these agents to understand the complete context. And that's unlocked again in -- kind of locked again with your unstructured data. It might be in your sales e-mails or it might be in notes and attachments or knowledge articles. So you definitely need system, which you can trust on. And definitely, if you do not have systems like this, you will see that the AI outcomes can be unreliable. So how do we unlock this trapped data? How do we unify this data? So to do that, right, we have Agentforce for Financial Services. So Agentforce for Financial Services, right, is a comprehensive set of tools designed to create and customize these agents specialized in financial services business process. So by integrating this data, AI and human, these agents can complete tasks within your organization. So we offer like a lot of out-of-box agents like you see on the screen, and they are easy to configure. They can efficiently handle all these simpler tasks. The goal is to boost productivity and efficiency for sales, service, your marketing team. To be able to easily transition a workflow from this to a human agent is also something that we have built in. So we have all that guardrails in place so that whenever things become complex for agents, they can easily sort of transition it to a human agent. So -- but the core of all this is data, right? And data is the trusted foundation. And Data Cloud is that hyperscale engine for that data, which is native to Salesforce. And we have created this so that it can not only unify your data by bringing structured data, but also all of the unstructured data that is trapped there in your notes, attachments, in your knowledge articles, in your website. So it can bring in all that data, right, and provide that Customer 360 context to the agent and has deep integrations with our Salesforce platform. So it's just not like any other data lake that you would have seen. This platform differs because it's deeply integrated to Salesforce. And also you can -- we have built low-code tools, and we provide a lot of locale tools to make it easier for all of you to be able to activate this data, whether it's for agents or Customer 360 applications or for building any of the agentic workflows. So -- but how do we unlock this unstructured data and derive meaningful insights from them? So the first thing is, we provide a lot of connectors so that you can like connect to the systems that have the knowledge about your customer and extract all the knowledge from that unstructured data. So you can like bring together, unpack insights from diverse knowledge types like PDFs, documents and more just with a few clicks because we make it very easy to configure all of this. And then you can activate this data. So it's not only about like being able to have access to this data. We make it easier with hybrid search and vector indexes and semantic searching so that you can search through this data. So it's almost like having your own search engine for your unstructured data. And then we make it very easy with our low-code tools, so you can automate this with this unstructured data or use it for analytics. So we have both low-code and pro-code tools. And then finally, how do you bring customer context to every AI experience? So for that, we provide best-in-class RAG experience. RAG means retrieval augment generation. And this is one of the technologies that we know how complex it is if you are doing it yourself. So we provide all the low-code and pro-code tools so that it's just clicks for you to set this up and get it -- get going. So for all of you technology folks out there, how do we do this? How do we have industry-leading RAG capabilities? So we do this because we have vector databases, hybrid search and we provide retrievers. So for your unstructured data, what you can do is you can unpack this unstructured data by indexing it, and we do this using low-code tools that we provide. You can categorize this data, store this and then action this across agents, analytics through automation or for Customer 360 apps or even for search functionality. So we have all the tools in place. So you're not building everything from the scratch. It's just focusing on the business value and configure these tools to get to the business value quickly. So with that in place, right, so the next thing I wanted to -- you to understand is this unstructured or unlocking of the unstructured data can elevate your businesses, right, across different teams and different use cases. So take a screenshot of this slide because this shows like how useful this unstructured data can be across different teams and also across various tools like for AI agents, of course, but you can use this for automation, analytics and search experience as well. And just to deep dive into a couple of these, right? So I have an example slide here where you can see that if a sales rep were to prepare for a meeting with a client and if the sales rep did not have this tool Agentforce with Data Cloud, you will see that it will require agent to really dig through all of these pages of documents. For example, earnings report, they'll have to dig into all these and then sort of have to like synthesize this and prepare presentations, it might take a lot of time, like 2-plus hours, as you can see. But if you had access to Agentforce with Data Cloud, you see how easy it is. You're just asking the right questions to the agent and the agent is able to read through this document, summarize it for you and give you the exact details that you need for the meeting. Same thing, imagine you're a service rep, right? Or you're building solutions for service rep. If there was no Agentforce or Agentforce with Data Cloud, you see that to resolve customer cases, they might have to read a lot of knowledge articles, go through different Slack channels and analyze that data and then get to the customer solution. It might easily take hours. But with agents -- by providing agents all the access to the knowledge, you're just asking the right question and the agents are here to help you with your answers. Now with that, we can talk about all these through the slides, but we have Jose, who has an excellent demo to show. So I want to hand off to Jose so we can see all this in action in the live demo. Jose, over to you.

Thank you, Mohith. How is it going everyone? Good afternoon. I will be sharing my screen momentarily here. And here we go. We've -- my screen is on. Okay. So now sticking to the same theme of servicing, some of the slides that you just saw about the time saved. I'm going to stick to the servicing theme, but maybe pivoting over to a bit of a different use case, more of a quick example of how we can leverage some of this unstructured data for perhaps making the service agent a bit more efficient in searching and looking for information and retrieving information. So let's talk about Sam here, right? So meet Sam. Sam, she's a member of the support team here at Cumulus Bank. Sam has just opened up her computer and has received a message from one of her key contacts, one of her accounts that she covers, Omega Inc. She's reached out -- the contact has reached out asking, for example, about the minimum loan amounts for equipment financing. Now Sam is somewhat new to Cumulus Bank and has not had any direct experience with financing for equipment. So typically, without Data Cloud and Agentforce, we all know what would happen, right? So we would expect Sam to go and search the company directory or share drive or internal links and docs to try to find the latest documentation related to equipment financing, right, and read through it to see if we can find the answer while the client is on the line, et cetera. But what if she doesn't know where to look, right? And instead, now we need to go ask a colleague and maybe that colleague doesn't know and we have to hand it over to someone else or a different team and yet another person, right? Or perhaps a dreaded, we'll get back to you with a response, which will totally ruin our chances of a first call resolution, right? And that's exactly not what we want. So this whole process takes time and could easily, as you know, take from minutes to several minutes to then hours or maybe next-day resolution, which is not what we want, before Sam is able to essentially get an answer for what the customer needs. So meanwhile, the customer at Omega is getting frustrated, minutes turns to hours, like we said, and all the answers from the customer's perspective is what seems to be like something pretty straightforward, a question about financing, right? Well, luckily, Sam can take an entirely different approach and leverage a new AI service agent that Cumulus Bank has implemented to help her more quickly resolve those exact issues. So let's demo that real quick. Let's open up Salesforce and see how it all works. So using the chat interface here that's embedded in Salesforce, Sam, who's our service agent, can ask the AI agent the question about equipment financing that she just received from her contact over at Omega. You can see that we are working and the agent -- and bear with me, this is all happening live. So after receiving the question, the agent is quickly going to go through the process of looking at safeguards, guardrails, instructions. And after a few seconds here, we can see that she gets the answer that she needed, right? So let's scroll and see, right? It says, yes, we offer equipment financing and the minimum loan size is $25,000. So for more details, we can refer to equipment financing documentation. We have a link for that. And then we also get some intelligent recommendations, some other questions that we could either answer or that may come up or that we could ask of the agent to get some additional information, right? So this summary is exactly what Sam needed to quickly respond to the customer, right, which is great. We were able to get that turned around without having to do all the manual work of going through all these different sources to find the information, right? So how does this actually work though? What's happening behind the scenes? I want to take you back to the agent builder experience, so we can kind of see this step by step how the agent got to this answer. So this here is our agent builder. This is our native interface for configuring and deploying agents. Let me ask the questions here. So now agents are built using essentially 3 simple parameters, topics, which are essentially what the actual questions or conversations or categories of things that the agent will do will be about. Then we have instructions, which are essentially what we want the agent to do step by step. You can think of it as somewhat of a process flow or a set of guardrails or things that we want the agent to explicitly do or not do and then actions. So actions are essentially what is the agent enabled or empowered to do on our behalf and they can execute these things in the context of the question. So let's take a look at topics, right? So the topics that have been assigned to the agent in this particular case, there's a specific topic called product questions. And if we read the instructions that have been written for this topic, you'll notice that this pretty much matches the exact use case we encountered with Sam who needed help answering some service-related questions, right? We can see, use this function when users inquire about details. We have a bunch of instructions here for additional context that we can add and things that match exactly things related to answering questions. Now when we open up the topic, we see a set of actions. And when we ask our question here, as you can see, we get the exact same answer, which is great, right? That is what we want. Now if we look in the middle of the screen, we can actually see the order of operations of the instructions executing by the agent and how the agent got to that answer and what it went through to get there as well as some of the actions. So in this particular instance, we're going to see that the agent started with a user prompt. And then based on context and using some of that reasoning in the reasoning engine, selected product questions as a topic, right, because it knows that the client was asking a question about a product. So based on that, it knows that it has all these different instructions that before creating a response, it has to follow and make sure that within the context of the response, it sticks to what we are allowing the agent to do. Now when we see actions here, we see the different actions that the agent is able to take based on the actual question. So the next step that you see here is that it selected an action, and that action is a preconfigured prompt in prompt builder that actually retrieves the information from unstructured data that is powered through Data Cloud. Now -- we can see finally the response, and we can see we got the next response. So let's actually take a bit of a dive into the prompt, so we can see how that works, right? So here, we can see the prompt and the prompt has been built for the service agent that Sam has leveraged. Now with the prompt, you'll see some specific sections of text here that stand out in blue and purple. The color text essentially represents the retrievers that have been added to the prompt to augment it with our bank's valuable business data and some of the inputs that we are passing over to the prompt, essentially our questions, right? So this retriever at the top that's labeled Einstein Search. That retriever is actually allowing the AI agent to search within all of the bank's unstructured knowledge articles, product documents, et cetera, anything that we give it for a search index and it has the ability to essentially search through all of this unstructured data, and that's all thanks to Data Cloud, right? So let's see how this is all set up inside of Data Cloud. Sorry about that. A bit of an issue there. Okay. So the first thing we noticed here in Data Cloud, and we see that we have an unstructured data lake object, so essentially, the bank has created by ingesting all their knowledge articles, product brochures documentation that existed as PDF files. We use either Einstein data libraries or a blob storage like S3 or Google Cloud Storage, something like MuleSoft Connect as well to be able to bring that in, right? So once the documents are ingested, there are essentially 3 data streams that Data Cloud automatically creates, which provide either -- which provide 3 different sets of documents, right, directory table essentially creates and index the chunk contact -- contents of the actual documents and then the actual vector embeddings, which are essentially our numerical representations of our data that our model can understand semantically. So we have data that was ingested and transformed at the Data Cloud. Now we need to make it searchable, right? So luckily -- and we got here, luckily, Data Cloud also automatically creates a search index based on that object, right? So when creating a search index, essentially, we create and we can define the parameters like the type of search that we want, whether it's hybrid search, semantic search, or keyword search. We can select our chunking strategy and our vectorization, our embedding model. There are a couple with more coming soon down the line. And then once that is created, we can head over to Einstein Studio, which is our third part of the equation, which is our retriever. This is also automatically created by Data Cloud, which essentially creates this retriever. And you can think of a retriever as essentially a helper so that we can search for this information using declarative formats, right clicks, not code, so we don't have to create complex Apex code, et cetera, to be able to look for this information. We can simply use that pointer in our prompt builder to be able to look for the information we need and then retrieve that data without some complicated implementation or development, right? So that retriever is created against the search index that we just looked at, and it's this very retriever that we used earlier to enhance the prompt and the prompt builder. Now another very important piece of information, especially in the regulated industries that we operate in is governance and security, right? So through Data Cloud's upcoming governance and security features, they're pretty much critical for Data Cloud and Agentforce because they ensure that data remains protected, compliant, accessible only through authorized users and systems, right? We enhance governance, security and safety, data sharing for Agentforce solutions by implementing things like automatic tagging and classification, taxonomies, organization of data, creating access policies and ensuring essentially that only consistently high-quality interactions are part of the Agentforce agent responses without a massive governance lift. So now with Data Cloud all set up as a trusted foundation for Agentforce, the bank can help employees like Sam more quickly resolve customer service issues that rely on AI agents, right, and find those answers that are very deep inside unstructured data or documents or links, right? In fact, actually, another question just came from Omega asking whether additionally, if we offer equipment financing specifically for construction equipment. So let me ask that question to the agent as a follow-up. Let's see how this agent responds. As you recall, it's going to go through the procedure of looking at the search index, looking at all of our documentation, following those instructions and guardrails that we implement so as to ensure there are no hallucinations. And then we're going to get a response that should be grounded in the context that we gave it, right? Now we're going to get, it's finishing up. I'm going to get a response. And bingo, yes, we offer specialized financing solutions for construction equipment, and I can get more details. Now it's going to populate some recommendations and additional questions. And just like that, I was able to get the answer that I needed as a follow-up. So it looks like Cumulus does offer financing for construction equipment, right? This is how AI and data can essentially come together and drive a better customer relationship, right? This is essentially what AI was meant to be, a reliable and efficient way to get us to get work done faster and better than ever before so. And now back to you, Mohith. Thank you.

All right. So that was a great demo, Jose, and thank you for showing all of the capabilities that we have. And with that, we have an exciting panel discussion today. We have Julie here along with some panelists. Julie, would you want to take over?

Yes. Thank you so much, Mohith. My name is Julie Jutte. I'm a Product Marketing Manager here at Salesforce. I'm on the industry's financial services marketing team. I'm very excited to be joined here today by my colleagues, Rob and Stephen, from Customer Success. They both work with financial services customers every day here at Salesforce, helping them get the most out of our products. So I'm happy to have them here to share their point of view on data and AI and financial services. I'm going to let them introduce themselves. Rob, let's start with you.

Thanks, Julie. Great to be here with everyone. I'm a Customer Success Manager here at Salesforce, and I've been one for 9 years. I work primarily with financial services customers and particularly those in the insurance vertical. And for those of you that are not familiar with the role, Customer Success Managers essentially work with our customers who have Signature Success Plans to help them achieve their critical business goals. So very specifically, we work proactively with our customers to understand what they're trying to achieve with Salesforce, and then we collaboratively create success plans that provide them with recommendations on how to leverage the platform and the Salesforce resources available to them to drive adoption, to optimize their design and architecture and to address issues quickly and transparently. Stephen, over to you.

Thanks, Rob. My name is Stephen Bennett, I'm also a Customer Success Manager here at Salesforce, I've been working in the ecosystem for over 8 years, and I do work in the Financial Services vertical, specifically with customers in the banking, wealth and fintech sectors. I'm aligned with Signature Success as well and strategically focused on helping organizations digitally transform and build trusted relationships with their customers, and I'm really excited to be a part of the conversation today.

All right. Let's get started right off the bat with a question about Agentforce. I know it's top of mind for everyone here at Salesforce and for our customers. So Rob, starting with you, do you mind sharing about any of your customers who started using Agentforce and specifically the use cases you've seen them have success with.

Sure, Julie. Happy to. First and foremost, I have to say that my customers are very excited about the possibility of leveraging Agentforce to transform their businesses. And so I think early on, I'm seeing 3 types of use cases. So use case 1 is where an Agentforce agent is actually customer facing and is able to provide customers with self-service for tasks where the agent can perform the task from end to end. So a couple of very specific examples. One of my customers is planning to roll out an agent shortly that will take care of insurance policyholders who need their ID -- user IDs unlocked or passwords reset on the customer portal. The same customer is also looking to create an agent to run in-force illustrations for life insurance policyholders in the 80% of cases where the illustration is fairly simple and can be done by an agent wholly. The other 20% will be redirected to a human to help. On case 2, also customer-facing, involves situations where agents are more providing information back to customers. I think for use case 2 and 3, very similar to the use cases that Jose was able to demo for you. But basically, in use case 2, a customer will -- maybe needs status on a loan request or a policy disbursement. And so the agent will read through all of the knowledge articles, case notes, PDFs with standard operating procedures and come back with the relevant answer. And then use case 3 is not customer facing, it's actually internal facing, and this is where agents -- the agent is providing a customer service rep with the information they need to answer a customer who might be on the phone. So for one of my customers, this is going to greatly reduce the need to rely on specialized CSRs, who are on call to answer the questions that the first-line CSR might have. So in both use cases 1 -- 2 and 3, unstructured data will be critical to grounding the Agentforce agents. So very similar to what Jose demoed, it will reference knowledge articles, case notes, et cetera. And in all 3 use cases, the business value to my customers is going to be seen in the form of deflected calls, higher first call resolution, lower average handle times. And importantly, those humans who would have been doing the task of the Agentforce agents are going to be freed up for higher-value tasks.

Yes. So it's all about optimizing your workforce, whether it's your digital workforce or your human workforce. I love that. Okay. Stephen, what about you, what are you hearing from your customers about Agentforce?

Yes, absolutely, Julie. So I have been also collaborating with customers who are exploring Agentforce use cases to augment their service and sales teams and the conversations are very exciting, exploring what potential capabilities are in store. So to highlight 1 particular use case my customer is exploring is a sales use case to help them qualify and warm leads, right? So today, both sales and account managers at my customer are tasked with a number of responsibilities, including managing relationships and also seeking out new business in addition to selling, right? And Agentforce will act as a digital assistance for those sales executives and account managers handling lead qualification responsibilities, outreach, schedule meetings, preparing for those meetings and some of the manual follow-up tasks that are involved in that process. And this will allow their teams to work smarter and focus on building and nurturing their books of business while being handed off quality prospects by Agentforce. So a very powerful use case. And accessing unstructured data is key to powering this particular use case. Many organizations like the ones that have been mentioned today store product and service documentation in PDFs, SharePoint repositories, Confluence pages and by leveraging data cloud, they can bring in this unstructured data and use it to ground Agentforce to deliver precise, accurate and highly personalized customer interactions that are needed for that type of use case. This approach will drive measurable ROI by increasing their lead conversion rates and also uncovering upsell and cross-sell opportunities within their business.

That's great to hear. I'm so excited to hear that customers are already getting into Agentforce and seeing great results already. So speaking of unstructured data, we heard earlier about how important that is so you can get a more complete view on your customer data. So Stephen, would you mind sharing any more perspective you have on how Data Cloud has really helped any of your customers or been an important tool for them?

I'd love to. So in the presentation, we saw a key statistic that 80% of data is hidden in unstructured content, images, PDFs, case notes, presentations, PowerPoint presentations and knowledge, just to name a few. And my customer understands that there's a massive opportunity to leverage this data to power analytics and AI capabilities, deepening their understanding of their customer and how they interact with and are supported by their business. And to double click back into my previous example, my customer, this customer in particular, has a wealth of customer-facing product and service documentation that's spread across their organization in multiple Salesforce orgs, systems outside of Salesforce and unstructured data sources. And Data Cloud is a pivotal tool that will unify and activate all of that data to augment both human and AI-led capabilities. So in the sales use case I mentioned, Data Cloud will allow autonomous agents to connect the customer product and service data that exists within their orgs and other -- in order to deliver those highly personalized journeys to qualify leads and hand them off to their sales and account managers. In service use cases that are also being explored, Data Cloud will bring in that same product and service documentation to power AI insights and provide those prescriptive recommendations and solutions to both their Salesforce users and directly to customers.

All right. Rob, anything you would add to this?

I think Stephen did a great job of summing it up. So I'll just kind of emphasize some of the similarities with my customers, which is Data Cloud is critical for unifying the data, both structured and unstructured data, from both sources outside of Salesforce as well as even within different Salesforce orgs across an enterprise. So for one of my customers, the sales organization and the service organization are in separate Salesforce orgs, being able to pull that data together to inform autonomous agents is really critical and Data Cloud does that perfectly.

Yes. Okay. It makes sense. You've got to make the data actionable, right? And it is not helpful if you just have a bunch of data coming in and you can't make it actionable for your agents or your humans. So I know in this topic, data strategy can be very overwhelming for customers. So I'm curious if you can share some of the main obstacles you see in financial services that your customers are facing when it comes to managing their data. And how have you seen Data Cloud make a difference for them. Rob, let's start with you for that.

Yes, absolutely. Sort of following on my last answer, for most enterprise customers, I would say, the data is spread all across the organization. So as we've said, pulling it together is critical to enabling agents to take action. And the data that's required for them to take action and even the actions themselves are not always in the sales teams' job of seeing the data up and then providing a mechanism for the action to take place outside of Salesforce. At the same time, and this is really important for some of my customers, many enterprise customers already have data lakes and data warehouses where they are similarly trying to pull together data from multiple sources. They don't want to replicate that data into 1 more data lake. So one of the most important things we tell our customers is Data Cloud is not a data lake. Data Cloud is there to provide a mechanism to action the data that you may have pulled together in your data lakes and data warehouses. And we do have the magic of zero copy to ensure that if you have data lake assets that live in Snowflake or AWS, that through the Data Cloud connectors, you can read that data where it sits without having to move it, without having to worry about synchronization and latency. And so this zero-copy feature really resonates with the enterprise architecture groups for my customers.

Stephen, would you add anything here?

Rob did a fantastic job covering some of the key challenges in how Data Cloud can help overcome those challenges. I'll just add or plus 1 to customers that have to navigate data in disparate systems. As we know, enterprise customers have numerous systems that stores that structured and unstructured data, which can lead to a fragmented view of their customer, right? So they're looking to bring that all together. And also, they have a wealth of data that they're looking to utilize. But are sometimes unable to effectively provide those personalized experiences and utilize their data the right way. And then the last challenge is navigating and maintaining data security and privacy, right, and meeting compliance requirements. Working in a highly regulated industry, financial services customers need to leverage that. So Rob did a great job explaining how to overcome some of the disparate system gaps and also utilization of Data Cloud. And I'll just add that Data Cloud provides robust tools and features to enhance data security and compliance.

Yes. That leads right into my last question. I want to talk more about compliance and regulations. It's a very important topic in financial services, as we all know. So back to you, Stephen. Can you share more about compliance and regulations? How it plays a part and how you see your customers approaching data and AI?

Yes. So our customers, like I said, work in a highly regulated industry where compliance and security is not only top of mind, but a nonnegotiable consideration for our customers' businesses. And I've had the privilege of working with -- closely with security, infosec officers at a few global organizations. And the consistent priority for them is adhering to regulations like GDPR, CCPA and industry standards that govern how data is handled, stored and processed. And for these customers, every conversation about AI and data begins and ends with security and compliance. And specifically, within the conversation of unstructured data and bringing those sources of data from outside Salesforce into the platform, Data Cloud enhances the ability to establish those governance, security and control access to that data with built-in tools like policy builder, which was referenced during the demo. Data Clouds' advanced security features, which includes enterprise-level encryption for data at rest and in transit, combined with the Einstein Trust Layer, provide additional safeguards. And with Agentforce my customers specifically really look forward or they really consider the control around setting guardrails and using their existing business logic to drive actions as being mission-critical. And they also need to have transparency of Agentforce's behaviors, which comes with the offering as well. They have those capabilities. So this comprehensive approach enables our customers to confidently test and deploy Agentforce while ensuring compliance and mitigating risk in their organizations.

That's awesome. Rob, what would you add here from your customer perspective?

Sure. I think to Stephen's point, the Einstein Trust Layer is a huge contributor to getting our customers comfortable with Data Cloud. The fact that enterprise data can be anonymized and fed to large language models with a guarantee that it will not be stored and will be instantly forgotten is a game changer. It's something that differentiates us from some of our competitors. There are still open questions. So as we move to autonomous agents, we're going to have to find a strategy to effectively supervise them, to comply with regulations much in the same way that our financial services firms currently supervise their human representatives. I think the one thing that's encouraging, though, is that compliance and security officers at my customers at least are taking a much more collaborative approach than we're used to finding these answers with us. Maybe it's because they understand the transformational potential of Agentforce or maybe it's because they're being pushed by their executive leadership. But either way, the end result is I'm very optimistic that we'll get to these answers.

Yes. Bringing compliance in early and often, I think, is the key for anyone looking to have success and being able to take advantage of all the opportunities of Agentforce. So I'm glad to hear that they are doing just that. All right, that wraps up our panel. Thank you so much, Stephen and Rob for your time and sharing your point of view from working with our financial services customers. So Mohith back over to you.

Thank you, Julie. And this is a slide for all of you, take a screenshot. Here you'll learn more about our success plans and how you can work with great customer success leaders like Stephen and Rob to get the most of your Salesforce products. So please take a look at that first QR code. Next up, we have a great guide here, built specifically for financial services customers about data maturity, so you can truly prepare your organization for this AI era. And lastly, we have a great new financial service demo for Agentforce transaction dispute's use case for all of you to take a look at. With that, I want to make sure that we have -- we take some Q&A from all of you. So use that Q&A tab, get in all your questions and get it answered from our experts here.

So moving on, let's -- so we have a question here. And I'm going to take it, and I think Rob already answered this, but we'll repeat it. How is Data Cloud different than other data platforms, data lakes and warehouses? So to give you an answer, right? Data Cloud is not a data lake. It's not a data lake. It's a hyper-scaled data processing engine that is designed to get data from multiple systems, we zero-copy from a lot of these systems. So we have zero-copy connectors. We have connectors to bring all these data, harmonize it, create a 360 customer profile we call or that is, in a nutshell, it's data unification. So we do that data unification and then we have tools on top of it so that you can take -- you can capture insights from it, action it, especially action it in agent use case that we were showing here in demo. So I hope that is clear. It's a processing engine. It's completely processing engine designed for actioning your data. It's not just that a lake house. With that, we have another question and Jose, this is for you. Do you need Data Cloud to use unstructured data in Agentforce?

Yes. Great question. And overall, I would say -- obviously, I may be a bit biased, of course, but I would say the answer is a resounding, yes, right? I mean, for -- Data Cloud is, I think, a key piece of technology that enables a ton of functionality with Agentforce, right? You're making an investment in Agentforce towards AI maturity, and getting there and unlocking a whole bunch of capabilities, so why not add that rocket fuel, right? I call Data Cloud essentially the rocket fuel that powers Agentforce and unlocks a ton of capabilities. We're talking about things like what we saw, right, the unstructured data, being able to bring in things like text documents, PDFs, voice and audio transcriptions live into where your employees work in the context of their work, right, being able to read data from zero-copy sources without having to duplicate that or ingest it in, right, being able to tap into external models, right? If you want to bring your own LLM model, right, something like Hugging Face or something like -- something through Amazon Bedrock or Google Gemini, et cetera, want to tap into the power of all those models, right? Data Cloud is the bridge or the platform that enables those capabilities. And not to mention, in regulated industries, as we all know, audit and logging and feedback is going to be a key capability that's going to be necessary. And especially Data Cloud provides that, right? The reporting on what is being sent to those LLM gateways. What are those responses? Is there PII data involved? What is the feedback if you have someone who's prompt engineering, right? Is -- are the prompts that are being generated through Agentforce and all the responses that we're getting effective or do they need work, right? That's where Data Cloud comes in. So I would say Data Cloud is absolutely necessary for this.

Awesome. Awesome. And this is for, I think, everyone here, please chime in, and I'll chime in as well. So there is a question from Heath Gordon. Is MuleSoft required to connect Data Cloud to our data lake. Jose, I'll start with you.

Sure. Yes. So we have a great story of MuleSoft and obviously, great technology there. What I would say is we have a lot of out-of-the-box connectors to a lot of warehouses, right, and a lot of other technologies, right? We have zero-copy, and we have a lot of those things, some of those warehouses and some connectors. Where MuleSoft really comes into play is integrations with platforms where, for example, we don't have an out-of-the-box connector. It's a great integration platform to be able to bring in data from anywhere, right, where we don't already have some sort of form of integration, right? It also works great for things like Kafka, right, being able to connect in some sort of broker subscription service and stream data in. It has -- it already has out-of-the-box connectors to Data Cloud. So it's a great complementary platform for whenever we don't have certain capabilities of connections. I would say, look at what's out of the box first. And if there's nothing there, MuleSoft is definitely a great addition to that combo.

Yes. Anything you would like to add, Rob or Stephen.

No, I think Jose covered it well.

Okay. Awesome. All right. So any other questions, please feel free to use the chat as we have experts. One of the -- and this is not a question, but a use case that I just came across for financial services. So I thought I'll share here. So any credit card purchase or anything like that, even the marketing e-mails or whatnot, right, these days, they come still in mail. And sometimes even if they are digital, they are a bunch of unstructured data and a bunch of documents to read. So I had like a bunch of these credit card requirements like my -- I knew my requirement, but just to find it out, I had to personally read all at least, I would say, at least 10 pages of documents, to find out exactly what I was needing. So I was like, okay this is a very good use case for agents. If you can like have all of these offer, brochure, materials just if companies can provide these as agents, right? It would have been just simple for me to just go to that agent, just ask that, hey, I'm looking for this, and the agent just recommends me with the right results. So it was an interesting use case that I came across and I thought I'll share. I think we are almost time now, and we are running out of the questions. So I would like to take a moment here to thank everyone, especially for all of you tuning into this webinar. And then of course, our panelists here, Jose, Julie, Rob and Stephen, for bringing some amazing content and great discussions. So I hope that you all have a great holiday season here and hope to see you on the next one. Thank you, everyone, for joining.