Leading the Future of Agentic Enterprises | MuleSoft Keynote, Dreamforce 2025

Please welcome President NextGen Platform Salesforce, Eric Eichenslider.

Welcome. Welcome to the Mulesoft keynote and day three of Dreamforce. How's everybody doing? Amazing. I hope you all had a good time at Benson Boone or Metallica. I'm not going to ask you who you favored. Coming from a home of girls, let's just say we weren't all on the same page. So today we're going to Talk about how MuleSoft is empowering organizations like yours to lead in the agent era. But before I get started, I want to say a big thank you. Thank you to all of our Muleys, all of our flownatics, all of our Salesforce customers and partners, and of course, all of our employees. Every day you collaborate with us, you inspire us, and you help shape our strategy and roadmap. Now, on day one, you heard in the main keynote, the era of the agentic enterprise is here. And it's about humans and agents adapting, collaborating and working together to drive outcomes and innovation. Agentic innovation is coming from, from every direction. Hyperscalers are building it, SaaS platforms are building it. Your teams are rolling out their own. And we want to embrace all this creativity. We want to harness it, we want to accelerate it, but we want to do it responsibly. We don't want to slow it down. And the best part is, it is at the center of this transformation. We do not have to reinvent the wheel. We can rely on tried and true principles, just like we have through every major transformation. Visibility, governance, extensibility. That's the foundation that it can trust to deliver outcomes at pace and scale across any agent on any platform. We're doing this with thousands of customers. In today's keynote, we're going to talk about how you can too. First, we're going to show you how you can discover, govern and reuse agents across ecosystems to unlock trapped value. Next, we're going to show you how you can orchestrate and gain visibility across any interactions to reimagine your multi agent use cases. And finally, we're going to show you how not only does mulesoft make your agents more effective, but how we supercharge your teams by embedding agents in your developer workflows. So by the end of today's keynote, you'll have seen how all these capabilities come together to supercharge your agentic journey. In this fast paced market, it's hard to know who to trust. Every vendor claims to have the playbook, and so who you depend on matters most. Right now, what better Way to give you a sense of our vision than to hear directly from our customers. And in today's keynote, you're gonna hear from these innovators that are putting mulesoft at the center of their AI strategy. Let's get started. MSC Cruises, one of the largest cruise operators in the world. And I don't know about you, but when I think about cruises, I think about effortless experiences. And this is exactly what MSC is delivering. With mulesoft as the foundation, they've deployed a virtual concierge powered by 50 agents that anticipate guest needs before they even ask. Dining preferences done. Offshore excursions already booked. That's what it looks like when agents team with world class hospitality. So let's hear directly from MSC Cruise's Chief Information Officer, Daniela Buonoto. Daniela, thank you so much for joining us here from from Geneva. Welcome Eric.

It's a pleasure to be here and share the story with you. Mssc, as Erica anticipated, is the third largest cruise operator in the world. And we are building beautiful ships. We are ambitious, but at the same time we want to give an opportunity to our guests to benefit more from the experience. You know, when you embark on these beautiful ships, you have a challenge how to get oriented in this vastity. It's not only bigger ship, it's also the number of services we provide. There are gyms, spa services, food and beverages, specialty restaurants. There is a plethora of different experiences. So how we can help the guest to effectively benefit from all this experience. And this is the challenge we wanted to start fixing with AI. So how we made it so we started thinking that we need to provide not only information, we need to give each customer information exactly the information he's seeking for. And what does it mean, this personalization to measure things. So Markus is playing very well that we need to know the context. The context is what are the information available about that customer? What is living on, what are the services that you bought? What are the likes and dislikes of the customer? Knowing the context, giving us a lot of leverage to give the precise answer. But there is a second point. We need to understand the intent. So intent and context are the key drivers for having an effective reply with a personalized answer. Using artificial intelligence, we created a very articulated architecture that is made by 50 agents. You see here an example of these agents. So we have the welcome agents, promotion agents, because we are able to understand what we are seeking for in that precise moment. And the advisory that we give you is not the same that we give to any other person. So we have another challenge is the number of languages that are spoken on board the ship. You need to understand MSSE is an international cruise line. Each time we have a cruise, we have an average of 40 different languages. Now any cruise company can only support 6, 7, 8 languages. So how you tackle this challenge, so you can think how AI is open opportunities. And now we are supporting more than 70 languages. Where Mulesoft is giving us a huge advantage is about nurturing information inside these 50 agents. So the complexity disappears when you have a platform like mulesoft, because in that case, we are able to convey any kind of information in the right way, with the right protocol and with the flexibility we need. So this concierge is not only adaptable and extensible, it's also scalable. And this is also thanks to the ability of nurturing additional information.

What a story. I love how you're leveraging the architecture. So I know you're ambitious. I know you want to keep moving it forward. What's next and how are you thinking about how mulesoft fits into that?

So with mulesoft, we want to enlarge this experience and we want to augment the abilities of our crew, of our employees to serve the guests. So it's not about, in any hour, replacing them. It's about giving them more power in managing the guest. Think that the guests have now, not anymore, the body language. They can be served in a very fast way, in a digital way. And the crew can also even support better what is the intent of their customers. At the same time, we have launched the biggest terminal in the world. You heard about this Miami terminal. It's a half a billion investment, the most beautiful terminal in the world. And we are using biometric technology to make a very easy experience to get on board. And when you embark, this is a very critical moment to achieve. So giving to you this experience requires the reliability and the velocity of a reliable platform. And it's Mulesoft again. Mulesoft is integrating Miami Terminal in a very flexible, reliable way with 99.9% of availability.

Thank you. Danielle. It's so great to have you here. So great to have you.

So much.

Mulesof keynote. Thank you for making the trip. I find Daniela's vision so inspiring. And the best part is he's just getting started now. We want to dig deeper into the vision. We want to show you more of where we're taking mulesoft. So please help me in welcoming to the stage SVP and GM of mulesoft, Andrew Comstock.

Thank you. Eric, and thank you to the crowd, Muleys. And I think a few future Muleys out there. I'm the GM for Mulesoft now. We're the third day of Dreamforce. So what I'm going to say next is frankly unsurprising. AI and agents are the next great frontier in technology. But despite this being relatively new, the moment feels oddly familiar. Who else was here at the early days in emergence of cloud technologies? At those first days, we saw new technologies, new APIs, new tools drive massive business opportunities. Despite the relative imperfection of those early technologies, we drove massive innovation. Companies had to digitally transform just to stay ahead of competition. And as we think through in the AI era of today, frankly, I just feel deja vu. But in the best possible way, because I know with the parallel with digital transformation that the people in this room are set up to deliver an incredible transformation now. And even better, they're gonna do it best in class. Because with mulesoft, we have helped thousands upon thousands of customers do digitally transform using innovation on top of a composable foundation. With our knowledge and collective experience, we're driving new innovations to bring together and accelerate your journey to the agentic enterprise. And this is why I am so thrilled to announce the general availability of the mulesoft agent fabric. The mulesoft Agent fabric is the solution to empower your transformation to become an agentic enterprise. And it does this by bringing together agents from built on any platform in a way that enables them to be connected, governed and enterprise ready. And we do that because we know the challenges of today. We have teams innovating incredibly rapidly building agents across the enterprise, but we hear that they're building them in disconnected ways. We also know that many of the agents built on third party applications have inconsistent guardrails, increasing our risk. And as we set up for the next generation and we start thinking about the exponential results that our teams are asking us to deliver, solving for these two challenges of discovery and governance are going to become critical unlocks to make that happen. Which is why as the part of the agent fabric, we always start with the Agent Registry. The Agent Registry allows us to bring together agents and AI tools built across the enterprise and on any platform into a single central repository. This allows developers to see what other teams are doing. It has a central place to manage and govern access and usage. And perhaps even cooler, agents can use the registry too, to discover new AI agents and MCP tools to complete the jobs that they need to do. This is a first step in unlocking that exponential growth. Because Once you can find what is happening in the enterprise, you can start to reuse it, making us more efficient and productive. But once we know about the agents that we have, we need to safeguard them. And that's where we bring together and extend the mulesoft that you know today into the agentic era. Just like you can with APIs, you can now manage and safeguard any agent to agent or agent to system interaction. Regardless of where the agents live or or what systems they're connecting to. You can apply policies to safeguard your agents and agent interactions, improving, protecting your data and ensuring your robust performance. But you might have heard me say that this was generally available already. So let's see this in action and to help, let's meet Horizon Expeditions, a world class cruise line company. Horizon Expeditions wants to go agent first with a new personalized experience in their new mobile app, the Horizon Assistant. But to do it, they're running into some of those challenges that I was describing before. So what we want to see now is exactly how they can solve those challenges with mulesoft. And just like I said, you have to start with the registry. Horizon Expeditions will bring together in one place all of their agents and AI tools into one single repo. This allows them, as you can see, all of their agents, MCP servers and LLMs are now available in one place, protecting and enabling them to understand what's happening in their enterprise. We can drill into one like that Spa Experiences Agent to see what this agent is doing without having to go look at the underlying implementation. Now our Spa Experiences agent is handling bookings and cancellations for this agent for spa Experiences. It's really well named and this is giving us access to it. But we now we know about it, we want to make sure it's protected and if you think about it, if it's handling bookings and cancellations for customers, it's really possible that customers might share just a little bit too much information with it. But for compliance, we want to make sure that that agent never stores or handles any customer payment information. To do that we're going to go to policies. Now you can see one policy is already enabled, but we want to add additional policies to protect the agent and underlying services there. So here we can browse all of the pre built policies. These are the safeguards to protect these agents and systems underneath them. Now you'll see that PII detector. So let's apply that PII detector and from here we can now specify the type of PII data. Now we don't want credit card information and Just like that, with a few clicks and clicking Apply, our agents and our systems are now protected, ensuring we have confidence in the actions that they're doing. So in just a few minutes here, I have shown us how we can discover agents across the entire enterprise and drive reuse. I've shown how we can protect and safeguard those agents. But even more, I've done this through using the MuleSoft tools you already know and love because we've extended them into the agent era. But Eric gave you four pillars earlier. And to take us a little bit deeper, please welcome on stage Ashley Hathaway to tell us a little bit more. Ashley.

Thanks, Andrew. All right, let's talk about how we get agents to work together. Because think about how we all work. None of us operate in isolation, right? It turns out agents are no different and for one very important technical reason. What we've been learning is that when you start to overload an agent, when you give it more tools, well, its performance starts to degrade, it starts to hallucinate. This is precisely why we know that the future it's not one super agent. No. Organizations are already starting to build task specific agents. But this is where the opportunity lies. Because these task specific agents alone, well, they can't fulfill all these requests. But when you start to orchestrate many together and you can still observe every transaction, well then you can start to solve more complex multi step business processes with just a single prompt. So to help you take advantage of this opportunity, we moved fast and we just launched the Agent Broker, an intelligent orchestrator as part of the agent fabric. And you might be saying, Ashley, what's an Agent Broker? Well, I'm so glad you asked. An Agent Broker operates within a given domain like HR accounting or has access to a set of task specific agents and MCP servers. And then using an LLM of your choice, it serves as an intelligent router and it's figuring out the right agents and the right tools to call in the right order to solve those complex business problems. But I know that's still a lot of words, so I do think this is going to be a little bit more clear if we talk about it through a demo. All right, let's go back to the Horizon Assistant, but this time we want to talk about how it was actually built. We're here in anypoint Code Builder and this is where Mophie, the senior developer at Horizon, is able to open up the project for the Agent Broker. This is the Agent Broker that is powering the Horizon Assistant. Now we know an Agent Broker is only as powerful as the agents and the MCP servers that it has access to. Right. Well, on the left hand side, you're going to see those agents. We've got a dining agent, We've got a payment agent, and on the right hand side, we've got the MCP servers. There's a housekeeping request tool. All right. She wants to add more, though. Mophie wants to take advantage and expand its capabilities. Well, luckily, she has access to the agentic development experience, and this is directly integrated with with that agent registry that you just saw. She's already asked, hey, do I have any agents related to the spa? And it came back with two options, including that spa experiences agent that you just saw earlier. This is the one that she wants. So she's going to click, she's going to say, I want to add this to the project. Now, notice, what did I do? What didn't I do? I did not go off and just build another agent. Instead, Mophie was able to reuse an agent and she was able to dynamically reference it from the registry as a result of adding just this one agent. Now, the horizon assistant can route customer requests across dining, across housekeeping, and now across the spa. Well, this has been deployed, so let's see it in action. Right. I'm going to take the assistant for a spin myself. It's the first day on the cruise, and I knew you could tell because I've been getting my tan on already. Right. It's also my friend's birthday, so I'm going to reach out to the assistant and I'm going to say, hey, help us plan a day. I want to make it really special. And I'm going to send that request off. And we notice the broker's already hard at work at the bottom here. It's figuring out the tasks, it's breaking them down. It's figuring out which agents it wants to coordinate across. It's reached out to the spa experiences agent that has the catalog of the spa offerings. It's passed it off, and it's working with the scheduling agent because it needs to fit that in amongst our other various activities. That's requiring a little bit of back and forth. And it's also reaching out to the payment agent because I need to know, hey, are there any additional costs involved? I love the schedule. And what I love even more is all that orchestration just happened under the covers. All right, I love it. I'm going to give it my credit card number. It's my treat. But remember that PII policy that Andrew applied earlier around credit card numbers. Well, it's coming back and it's saying, hey, I can't take credit card numbers in the chat, but there's one on file if you want to use it. So that's perfect. Let's go ahead and confirm the schedule. Now, this whole experience was powered by an agent broker. That orchestration happened under the covers. Meanwhile, I just had a seamless experience. But developers who are building this, they're going to need visibility into what's actually happening. So let's talk about another critical part of enabling your identic enterprise, and that's the ability to actually observe your agents. Because without this capability, agent interactions, they become a bit of a black box. We use the ability to trace their workflows, to debug failures, to monitor performance. And these problems, they only compound as you add more, you add more and you add more agents. But with Agent Visualizer, this is the fourth capability of the Agent fabric, we turn what could have been this black box instead into a source of clarity. Here you're able to observe the performance, detect failures, and build trust in your agent network. So let's see how. Okay, we've been talking a lot about all those agents that Horizon's been building. Well, here in Agent Visualizer is where they can see all of them centralized. They see the agents, they see the interactions, they see critical metrics about each one. So we're going to zoom in on the Horizon Assistant. We see the tools, we see the agents that it's talking with. Notice it does not matter where these agents have been built or where they're running. There's agents on Bedrock, on Vertex, of course, on AgentForce, and all those edges, that's based on real time customer interactions with the app and the traffic that was generated. If we click into the assistant, we're going to see critical details. We see its description, we see where it's deployed. But if we scroll, we can see if it's in compliance with our governance standards. There's that PII policy from earlier. If we scroll further, well, we can see key metrics like, have there been any failures? And if so, we can drill in and click into the logs. Now, this powerful tool and all the metadata that's behind it. This is the culmination of the past four demos that you just saw using mulesoft Agent fabric. Horizon Expeditions centralized all of their agents into that one registry. They could discover, they could reuse them and then they could orchestrate them into this Horizon Assistant. They made sure that every Agent interaction happened with the necessary guardrails in place and then it was automatically centralized them back in the agent visualizer. Pretty cool, huh? But all right, yes. But this also isn't just talk. Customers are getting started with this today. In one example is Varco. Varco's been leading the industry in their imaging technology. This goes all the way from live events to surgical suites. And now they're transforming how they operate. They've been moving from a hardware first business to an AI driven enterprise. And with mulesoft, they've built out an orchestration layer where all other agents in their IT estate are working and that is built on their API layer. But I know you don't just want to hear more about this from me. So let's welcome the expert himself to the stage. Please join me in welcoming senior integration architect, Joris Stependales.

Hi everybody. Actually, it's a pleasure to be on stage with you today.

Welcome. Thanks for coming all the way from Belgium. It's his first dream for us too.

Thank you. Well, I'm actually honored to share the Barco story here today. Very proud of what we've been doing. So let's get started.

Thanks for being here. We've been talking about innovation behind MuleSoft Agent Fabric. You're the ones making it real. We see your architecture here. Can you walk us through how you're bringing the agentic future to Barka?

Sure.

It's been an interesting journey so far. We started exploring the rack like everybody else over the past years. We embraced the agentic shift this year and we started actually building explosives, flooring and so on with our first agents live. So it's been quite a transition going from apps and systems to agents. So we had been challenged quite drastically, I would say. So we started building and our first agents went live. Actually.

I love it. So we've seen things like A2A and MCP on here. Can you talk us through that and how it's fitting in?

Of course. So what you can see on the screen, or was on the screen was the architecture behind our Barcogenius. So Barcogenius is an AI knowledge assistant. It's a collection of headless agents which are the backbone of our digital workforce. So the digital workforce is actually the embedded agents that you have in Salesforce, for example, agentforce, Copilot from Microsoft, but also Jewel from SAP. So these headless agents, being the backbone of this digital workforce, interact through A to a. So agent to agent protocol amongst each other. We've also been building different tools. The tools we offer to the agents through mcp. So the model context protocol. An important note on the tools is that these are actually our composable building blocks which we've been investing in over the past years in the MuleSoft platform. So our APIs as part of our API led architecture.

I love that you bring up the APIs and that reuse because we know speed is a really critical business imperative right now. I feel it, I'm sure you feel it, but it does bring up the question of how you go fast without breaking things.

The pressure is indeed on. We need to deliver experience, experiences and AI capabilities fast. I think everybody can acknowledge that we already have been investing in governance on API side using mulesoft API governance. So we're quite strong on that. It also means that with what's ahead of us, we need to start also managing the API life cycle end to end. So it's very important to make the shift towards the governance of these agents as well. And that means that with the new capabilities of agent fabric, we can actually expand our governance of these APIs also towards the agent governance all in the same platform. So it's really easy and useful for us doing it centrally like the discovery with the registry, but also making sure that the APIs are compliant and secure really helps us in achieving this in a fast paced way without actually taking any risks. So we need to scale fast but safely. So the compliance is really important.

I love it. Thank you so much for joining us.

You're welcome.

Thank you.

All right, we have been talking about how you build the agentic enterprise, but we haven't forgotten about our favorite people. We have not forgotten about all of you. So to take us a little bit deeper and to to go back to our roots, talk about our developer productivity tools, please join me in welcoming mulesoft CMO Kurta Mukherjee.

Thanks, Ashley. All right, developers, buckle up. Today we are not just coding, we are wipe coding. So I want you to think less lines of syntax and more lines of magic as we turn your ideas into your agentic superpowers. We want you to have an AI partner in code, a co worker, and when you need it, your favorite debugging buddy by your side. Now, we've all heard people say, good vibes, bad vibes. I am sure, fellow parents in the audience of teenagers, you have heard this in every context. I know you get it. Today I am going to take that energy and bring it into our code. We are so thrilled to announce mulesoft Vibes. Vibe coding is not just about coding. It's about making that developer experience simple, simple and intuitive. With an AI partner right here in anypoint Code Builder. It learns from your projects, it understands your patterns and it produces enterprise grade code every time for your developers. It can help you design and generate APIs, it can create tests for you, and it can also deploy and manage your applications. You heard Patrick say on day one Mark Keynote from ideas to anything instantly. And that's not all. You can also create custom connectors using natural language prompt right here in Mulesoft Wipes. Now we know your journey doesn't end there. You have to keep those applications up and running. And so with our new MuleSoft diagnostic agents, we make troubleshooting easy for everyone. You are able to quickly pinpoint the issues as well as provide recommendations, essentially reducing that time it takes to keep your applications up and running. Now, if you are like me and I've spoken a lot about these new cool features, I am sure you want to see it. And who better to show us than someone who lives in our product day in and day out. Please join me in welcoming Mofi Oluwalana, our Director of Product Management.

Awesome. Thank you. Thank you so much Karuna. So yes, I usually live in product as the Product Manager, but today I am the developer. That means that I'm the one under pressure to move fast, build APIs and make sure that the code actually works. What's my task today? To build an API for our Spa Experience Agent so that it can manage and recommend spa treatment. Normally, a project like this could take weeks, even months. The good news? I have Mulesoft Vibes natively in anypoint Code Builder. My first step is to see if there's an existing API that I can reuse. In the past, this meant bouncing between portals, catalogs and internal documentation. Every developer, and I mean every developer knows just how frustrating to that experience can be. With mulesoft Vibes, I don't leave my workflow. I can ask the agent right within anypoint Code Builder. So it turns out that Horizon doesn't have an API for this use case. So I'm going to ask the agent to create one to create an API that we can use for managing and recommending spa treatments with Horizons reservation System mindbody here's where the magic begins. The agent is able to read Horizon's API documentation, understand my requirements, and begin to intelligently generate the application logic for me. But it's not just pulling in my application content. It's also adding in any connections, any dependencies, and any configurations that I need so that I can fast track deploying this application within seconds. My API is ready. All right, before I deploy this application, I need tests. And let's be honest here, tests are not of anyone's favorite part of the development experience. With mulesoftvibes that changes. I'm able to ask the agent to create tests and it does so from for me, removing the burden of what would otherwise be time consuming and tedious. All right, with my test complete, my application is ready to go live. I asked the agent to deploy and I am all done. Almost as we know, building fast is just half the story. Making sure that your applications are running smoothly are just as important it's almost Halloween and Horizon is launching a special promotion for hot stone massages. Spooky. I know it's a smash hit, but the surge in reservations gives the system a fright. Response times are taking longer than they should. When this happens, I'm immediately notified through an automated alert in Slack. That alert is going to direct me directly to Diagnostic Agent, which not only pinpoints the problem, but also recommends a solution to enable auto scaling on my application. With the solution handy, I'm able to navigate back to the agents in anypoint Code Builder and ask it to redeploy my application with auto scaling enabled. Just like that, without ever opening up a single log or dashboard. We've resolved this issue together. MuleSoft vibes and diagnostic Agent give me a complete agent powered experience from building and deploying to monitoring and fixing issues. With these capabilities, I don't need to leave my workflow. Instead I can focus on real outcomes. This is building better, faster and smarter with mulesoft. Back to you Karna.

Thank you Morphy. Now we need to keep this momentum going not just for our developers, but for our architects, our admins, and everyone in the audience that wants to innovate. And you are extremely excited that with our new AI connectors you can bring AI to everything that you already have. With our Inference connector supporting more than 30 plus models and our Vector connector supporting 15 plus Vector stores, you can bring AI right there. And I'm really excited to announce that we are partnering with Nvidia to announce support for their new NIM Microservice that accelerates. That accelerates their model deployments as well as support for the new Nvidia Nemetron Open model family. With this dedicated connector, you can build enterprise grade AI applications everywhere. Now developers, let's talk about one more thing that is so important in our jobs and that is monitoring. With our new Integration Intelligence The Mulesoft Telemetry meets Tableau Next and Data Cloud. You are easily able to pinpoint those issues. My favorite part, we no longer have to go bug hunting. That's how easy it is. We've had a busy year. This year we have brought the power of mulesoft to our line of business teams and our admins with mulesoft for flow, making it easier for you to connect and integrate all your data and applications. But none of this really matters without the builders, the innovators, without all of you bringing it to life. In our audience today, we have a very special Muley. A long time Mulesoft ambassador, a community leader, a role model for thousands and thousands of developers. He's an agent blazor legend and agent fabric user. He has shaped the community and shaped the roadmap for Mulesoft. Please join me in welcoming Sravan Lingam to the stage. Stravan, thank you so much for being here with us today and I know this is the favorite time of year for you.

Of course. Yeah.

So you've been a part of the mulesoft community for so long. It has helped shape your career in technology. I would love for you to share with the audience what advice do you have as somebody getting part of this community community and starting their journey in technology.

Thanks Karna. And first of all I would like to thank everyone and I am so honored to share this stage on the keynote and I love my amazing mulesoft community over there. The support, yep, I want, I'm proud to share my experience on like how my career started with Mulesoft. So it was a decade ago so I have been like with this journey around like 10 years ago and when I started my career with Mulesoft I was looking for the resources like where I need to start. Of course there were like many resources online but I wanted to know like what is the best point or best place to start. Then I got to know. There is a local Mulesoft meetup group in Hyderabad, India. Then I have attended this meetup and I got like so many connections, best advisors and from there I started taking advices, cleared my certifications and yeah, here I am today. And then I started thinking of like how to give it back to the community. So then I started as a speaker, Mulesoft speaker. That's where I started contributing to the community. And then I became a Mulesoft leader for the same chapter where I have attended the meetup in Hyderabad and then an ambassador and recently I have been inducted into hall of fame this year.

Yes.

Congratulations.

Thank you. Yeah. Then I started sharing my knowledge through YouTube channel and I have probably like traveled around like 10 different countries and I always make sure that I meet at least one mule in that country. So my small recommendation or you can take it as advice to the community. Whoever wants to start their career in Mulesoft is like, you know, mulesoft is very good at connecting systems. So advice is simple. Connect with the community, go to your local meetup groups or attend the virtual meetups. Join our Slack channel and meet the mentors Ambassador. We have some amazing community out there. Please give a shout out to the amazing community over there. And last note that I wanted to tell is like, we have seen many communities here, but Mulesoft community is not just a community, it's a family. So once again, thanks for joining.

That's amazing. Well, Travan, in recognition for all the work that you've done for the community in helping us shape the roadmap map, it is my absolute honor to give you the coveted golden hoodie. Thank you, carlin. All right, we've heard a lot about how mulesoft powers the agentic enterprise. But the one thing that is even more critical is high quality data. And so to hear how data fuels this agentic transformation, I'm going to call on stage, Andrew. And joining him is the chief product officer for Informatica, Krish Vital Devra.

Thank you, Krita. Thank you, Krish. Krish, it's great to have you on stage and talk together. You've been a long time working with Salesforce. Tell us a little bit about that journey with Informatica.

Yeah, no, Andrew, first of all, thanks for having me and Informatica here at Dreamforce. When I look around the room, I see that many of the Salesforce customers and Mulesoft customers are also Informatica customers, especially the large global brands. Right. Orville, Pfizer, New York Life, National Instruments, many, many more. So this is my first Dreamforce. And I will just say the energy is vibing and I'm very excited to be here.

Welcome to your first Dreamforce. We've been talking a lot about the agentic enterprise and how customers are customers can get theirs and companies can get there. How does Informatica play a role within that AI journey?

Yeah, I think when I talk to the customers, it's incredibly clear that getting their data in order is the number one way to get to the AI success. Right. So when I talk to all the CDOs, like, in fact, we surveyed them and 56% of them indicate that data reliability is the biggest blocker for them. So we at Informatica, with our intelligent data management cloud, CLAIR AI engine and all the purpose built data agents, we kind of help these customers transform their raw enterprise data into data that is AI ready, right. So that they can have a very strong AI data foundation. And when I talk about AI ready data, I think one of our customers, Takeda, articulated it really, really well. They said, hey, with Informatica and CLAIR AI, what we do is we convert our enterprise data into AI happy data. Data that is intelligent, contextual, trusted, compliant and governed.

Yeah, it's very clear. Informatics got incredible data management capabilities. This is Dreamforce. Is there one thing you want anyone to think about? We always talk about futures. What's one quick thing on the roadmap you'd love to tell about?

I will say, like you cannot walk down the halls and not feel that the agentic enterprise is here. And that's exactly what we do. Right. Like you talked about mulesoft, with all the work that you're doing around Agent Fabric, all our agents are available through MCP so that you can actually use something like Agent Fabric. Take our purpose built data agents and build your own custom agents. The custom agents that work on reliable data and they're gonna have their data and AI agents operate in a trusted, compliant and governed way. What's not to like about it?

That's great. Thank you, Krish, so much for joining us. So we're going to talk about one last story. And when you think about innovation, you don't always think about highly regulated environments. But Moody's is looking to change that. Moody's is building the next generation of their company around a network of agents. So to tell a little bit of the story, please join me on stage, Eric. And alongside head of Digital Content Innovation at Moody's, Christina Peretti. Christina.

Welcome. Christina, welcome to Dreamforce and to the mulesoft keynote.

Hi. Thank you, Eric. And thank you everyone. It's an absolute pleasure to be here with all of you today.

Now, Moody's has been doing a lot with agents. So tell us, how does Agent Fabric fit into your strategy?

Oh, it's a key part of our strategy. So let me start maybe by telling a little bit about our agentic strategy, because we have a pretty ambitious agentic strategy that involves basically two pillars. One is getting genai data and we just heard a little bit about what that means. And the second is developing highly specialized agents. When we think about the data, and I think we've heard from a lot of our speakers today of how important the data is.

Right.

If you give your agents bad data, then the results are going to be bad outputs. And we have a vast data set. We have the biggest private company data in the world database in the world that involves financials, that invites private company descriptions, that involves information about who's managing the company, who their owners are. And we want to make sure that we provide the right data, the right context, not only for our agents, but also for third party agents. Right. So that's one part of the strategy. The second part is we are developing highly specialized agents in areas where we have a lot of subject matter expertise and a lot of data to power those agents. So things such as, of course, credit risk assessment agents, kyc, know your customer agents, sales and marketing agents, early warning monitoring agents. Right. Now, why are we doing this for our customers? Basically, we want to support them in their agentic journey. We want to make sure that those customers are operating better, faster, with higher accuracy, higher efficiency. But a really important part, it has to be accurate and it has to be compliant. And that's where mulesoft agent fabric comes into place. Right. Because as of now, we've been doing that orchestration, that compliance, that checking the behavior of the agents, that checking that they're accessing the right data ourselves. Partnering with you. Well, we have the central control plane, which where we can check and monitor the behavior of the agents, make sure they're accessing the right data, and make sure that everything is compliant, which is really, really important because the majority of our customers are financial services firms and we all know how important it is that they are compliant.

I love to hear it. So, look, you're driving the business forward. A lot is going on. That's really great. At Moody's. Where do you see this all going and how does mulesoft help you get there?

Yes. So we are in this agentic journey. This is our first priority. We recognize this is the biggest transformation, and I think you all do as well. That is going to happen in our lives. And we want to build not only an agentic Moody's, but we want to make sure that our clients become agentic versions of themselves. Right. By partnering with Mulesoft, we can basically concentrate on what we know how to do best, which is surfacing that trusted data and building highly specialized agents and work with mulesoft as the mulesoft agent fabric. I'm sorry, to do that orchestration to do that monitoring of the agents to make sure that it's Raising the hand when they're not accessing data in not a secure manner. So we basically see ourselves as being super powered by agent fabric. Really excited.

Thank you so much, Christina. Big round of applause for Christina. Thank you for having you here. Now, what an incredible story and what a day. Today we've seen how mulesoft helps you to take control of your agents, unlock their value, and accelerate your Genentix journey. Stories like Moody's show you what's possible when agents work together. And they're not alone. Today we've heard from a number of organizations that show you how MuleSoft can drive forward new agentic use cases. Today we also had the opportunity to reveal one of our biggest innovations yet. Mulesoft Agent Fabric alongside developer breakthroughs that are making your work easier. And look, as a memento of today's launch, on the way out, please be sure to grab one of these mulesoft agent fabric patches, a symbol of your support of shaping the future of agents. And we're not done yet. We're continuing to invest in the platform and develop capabilities to make your work easier. You saw a glimpse of just some of these AI connectors and integration intelligence, and there's so much more coming. So today you've heard the vision, you've heard the product, you've heard the possibilities. What's next? Well, if you haven't already, get out to the show floor and visit the Build a Bear Expo. Another great showcase story. But most importantly, we want you to get your hands on the technology. Sign up for a trial. As I wrap up here, I want to emphasize we want your feedback. Because together, we're not just imagining the future. We're we're building it one connected enterprise at a time. Thank you for joining us at Dreamforce2025.