Slackbot Empowers Team-Wide AI | Slack Keynote, Dreamforce 2026

There are doors for you and doors for your friends. Some doors go around in circles. Some go 200 miles per hour. Doors can go up, down, underground, or high above it all.

Woo!

And then there's this door. Slack is the front door to Salesforce. And behind it? Everything. All your people, platforms, apps, and AI, all working together. When customer data meets conversational context, you make progress and Slackbot makes magic. When people can build with agents and open channels, Solutions get shared. Questions get answered. Ideas go farther and farther and farther. That's why the biggest, brightest, fastest, most trusted brands work in Slack. Salesforce gets stronger.

Apps get smarter.

Business gets better. All behind one door. All that's left to do is walk through it.

Please welcome Executive Vice President and General Manager, Slack, Rob Seaman.

Good morning, Dreamforce. I have been working on Slack for close to 6 years now, and this is the most excited I have been to talk to you all about it. We are seeing unprecedented adoption of Slack by the Salesforce customer base. We're seeing unprecedented behavior change of our users in Slack and how they're adopting AI and deploying agents into channel. And we're seeing unprecedented innovation from the ecosystem in building their products into and on top of Slack. And it's because of that and because of all you that Slack is now the fastest growing product in all of Salesforce. Thank you. You and your agents are sending over a billion messages a day. That's more than are going out on Twitter on a daily basis. We've acquired over 100,000 new customers this year alone. 1.5 million people connect to the Slack MCP server through the agent of their choice on a weekly basis, and 1.5 million of you are using Slackbot every month to make yourself more productive. Now, you heard Marc, and you've heard Salesforce throughout the course of this event talk about the concept of AI Force. And the fundamental premise behind AI Force is that every single user should have choice about where they use their technology, the interface surface that they choose. And we're going to spend some time talking to you about why we think Slack is one of the best interfaces for you to do that. And we're going to do it through the lens of the, some of the fastest growing companies that are out there today. If you look at the AI ecosystem, if you look at the most innovative companies in the world, the fastest growing companies in the world, Anthropic, Meta, xAI, OpenAI, Vercel, companies like this, they literally run their company wall to wall on Slack, but they're also building their products into Slack. You can see Kat here from Anthropic, we'll have Boris up in a second. Kat say that it's the core operating system of Anthropic, because the work happens where the context already lives. Now, they get disproportionate value out of Slack and their investment in AI for a few reasons. First, you may think that they use different AI, but the cool thing here is they are not using different AI. What they're doing is they're using AI in a fundamentally different fashion, and we're gonna talk to you about how they do that and how Slack fits into that picture. So, most of you probably have an AI stack that looks something like this. You've got an investment in a foundation model or a relationship with a foundation model provider. You might have a data lake on top of that. You might even use Data360. You're building some MCP servers on top of the apps that your IT team already has. If you're in a regulated industry, you might have, you know, a guardrails framework or a trust framework built on top of those things. And then you have some applications, agents, and interfaces that you're deploying out to your employees. Now, what we're seeing from the labs and what most companies are missing are 2 fundamental pieces, what we call the bookends of the AI stack. The first of those is a context layer. And that context layer is where all the rich conversation from your company happens. This is your, if you listen to Alex Karp or Chamath, they call this a company's alpha, their differentiation. These are all the conversations your employees are having about the most important projects at your company, the topics, the teams, the work items, the products you might build and why, the ones you walk away from and why, acquisitions you might make, et cetera. That conversational context, when imbued into your AI tools, gives you a fundamental advantage over your competition. But one of the things these companies also see is that most AI tools are what we call single player, meaning it's a person sitting in a terminal with an AI tool or a person sitting in a SaaS interface in a chat experience with the AI. And Toby from Shopify, I think, said this very intelligently, that the downside risk of that usage pattern is that the AI learns and that person learns, but the person sitting next to them or the person like them in that role sitting somewhere else geographically Or the company as a whole doesn't learn from that experience. So what we're seeing from these frontier labs and these AI companies is they're picking their AI up and they're putting it down into a multiplayer interface. And that multiplayer interface is Slack. And we think when you have these bookends to your AI stack combined with your AI stack, that is the new work operating system for the AI era, and that is Slack. And it's the only place where AI becomes multiplayer. And what's interesting, when you have this, you start to see a flywheel of value emerge. So when people use AI together in channel, the overall AI fluency of your company will go up. You'll see what other people are doing with AI. The exhaust of every one of those interactions becomes shared context that collects and becomes a knowledge artifact for individuals within your company, but also becomes context fuel, if you will, for other AI agents that you use. And that loop happens over and over and over again, and intelligence compounds for your people and your AI, your company becomes smarter, and your company moves faster. Now, we're seeing this happen across every single industry in the world. It's just not, it's not just the technology industry. We're seeing companies in retail like reMarkable, we're seeing companies in fintech like Robinhood, in the tech space like Intuit, or in auto like Rivian. Companies across the board are moving more and more of their AI workload and experiences into Slack for the reasons that we just outlined. And I want to bring up one of the most innovative companies in the world to talk about that. I'd love to bring up— first of all, before I do this, who here has used Claude Code before? Okay, many of you. So you can have a little bit of a fanboy or fangirl moment here. I'm going to welcome up Boris Cherny, who is the creator of Claude Code from Anthropic. Boris. Thank you for joining us, Boris.

Thanks for having me. This is a beautiful stage also.

It is. It's really well laid out. Thank you for everybody who joined us in here. So you use, I've seen how you use Anthropic. You have our old office, by the way, which is a beautiful space. We're glad to see you're taking care of it. Well, we've seen how you use Anthropic. You use it in a pretty fundamentally different way than most companies use Slack. Can you talk to us a bit about, you know, some of the behaviors you see within Slack at Anthropic?

Yeah, totally. For us, Slack is just at the center of all the work that we do. So every employee is using Slack. It's not just engineers, not just the researchers, it's also the marketers, the salespeople, the business people. Everyone is on Slack. And the cool thing is Claude is on Slack also. And so, you know, we're the company building Claude, and so we want to use Claude to build Claude. And a lot of this work happens on Slack. And to make this work, it's exactly what you were just talking about. You need Claude to have access to all the conversations, which means we try to talk in public as much as we possibly can. Less DMs, more public channels. And a lot of the coding, a lot of the building happens in Slack itself.

It's fascinating. So I've seen some statistics about it, but maybe we can talk about your own personal experience. Of the code that you're writing, are you actually writing in Slack versus another location?

Oh, 90%.

Wow, that's fascinating. Maybe we can pull up, I was talking to Boris the other day about how we might do this, and we were talking about the power of building in the open, and this is a change you made to the composer within Claude that's pretty fascinating. So maybe you could walk us through what we're seeing here.

Yeah, so I think this is the actual video of this conversation. This is actually it. So I posted about it.

I'm sorry if we surprised you. I thought you knew this was going to be here. This is the actual conversation.

It's more fun to show the real thing.

Yes.

Yeah, so this is an actual conversation. So I was confused about something in our products. We have these 2 icons. They look similar. I keep forgetting which one to click. So I posted about it in our product feedback channel. And Claude just jumped in. So this is the coolest thing, right? It's very cool. I didn't have to tag it. I didn't have to ask it to participate. It has the judgment to know to jump in. And this is an experience we've actually been testing for a while. And we found out with Fable, it finally starts to work. This is something that just doesn't work as well with Sonnet, other models, because they just don't have that same judgment. And so here, I'm interacting with it like I would a coworker. I'm not giving it a bunch of specific instructions. I'm saying, OK, here's the problem. Claude looked at the data, it figured out the problem. It queried BigQuery, it queried Datadog. It's connected to all our systems through MCP, so it knows what to do. And I asked it to mock it up. That's all I said. That's the only instruction. And Claude mocked it up. It has access to the code, it has access to the data, it knows what it's talking about, right? Like, this is the exact same thing a designer would've made.

What's cool is I think we'll actually see a designer pop in here, right?

That's it.

Yeah.

So this, this is a designer. This is another person on the team jumping in. And this is the coolest thing, right? Like we're not stuck in a terminal. We're not stuck in an IDE. The conversation is happening in the open. It feels like a much more human experience.

It is. And the cool thing is about this being in public is these are not necessarily designers that work on that particular area. This could be anybody with initiative that happens to stumble upon this work in flight that can contribute to the betterment of your product and your customer experience, right?

You know, we just had a team dinner like 2 days ago, and the person sitting next to me was Dan, who's one of the marketers on our team. He's actually the marketer for QuadTag.

Okay.

And I asked, hey Dan, what's the last piece of code that you shipped? He's not an engineer. And he was like, you know, like I'm shipping a bit of code. I actually don't remember what the last thing was. And so we looked up his pull requests, we looked up what he's been shipping, and he's been using Quad to build features for Tag. And this is everyone on our team. It's not just engineers. It's the designers. It's the product managers. It's the marketers. Everyone is building now.

It's immensely powerful. It's so cool to see, like you mentioned, so many people outside of engineering being pulled forward into the development process and making the products better. You talk a lot about building for where the model is going to be. And I think we can see it here. Like, you don't have a single instruction in here that's more than 3 words, which is fascinating. But you worked on this for a long time. And I assume it wasn't this good when you started.

It was not. This, like I said, this is an experience. It started to work with Fable. Fable is an expensive model. It is the frontier of what we build. And because it's at the frontier, it has the ability to code. Other models don't. It is safer than other models. So it's not going to get prompt injected. It's not going to have other safety issues that other models might. And it just has this judgment and maybe like wisdom that models don't. So really, I just have to give it a few words and it'll figure out exactly what I mean almost every time.

That's amazing. So you built this experience in Slack in the models that predated Fable and were using it. It just wasn't as good. And then, but you didn't necessarily change the experience. You waited until the next model came and it magically got better.

I mean, that was quad code also.

We built it.

It took 6 months for the models to catch up.

Wow, that's fascinating. Well, I'd love to, let me ask you, one of our engineers who is probably one of our most Claude-pilled engineers wanted me to ask you this. If you had a magical Claude and you could add any feature you wanted into Slack, what would you add?

Ooh, are you gonna build it if I ask for it?

Kevin, Kevin, are you in here? Kevin said he would build it, so yes.

Okay, okay.

We may not ship it, but we'll build it.

You know, I think the one thing I would ask, Slack is an amazing product. I don't think I would add features because it has every feature I need. The thing I would ask is remove features, get things out of the way, because unshipping is often harder than building.

It's very, very wise. We're working on that, and I can't wait to get your feedback on it as we ship it. Thank you, Boris, for joining us. Thank you for everything that you've done for our partnership. It's a pleasure.

Thank you very much.

I was fanboying. So for the rest of the presentation, to lead us through it, I want to welcome up the CMO of Slack, Ryan Gavin.

All right.

Hey, brother.

All right. I know it's Thursday. I know Dreamfest is done, but let's bring some energy, yeah? We got some energy.

Come on.

Let's go through the front door. We're going to get into this. We're going to show you some really, really cool stuff. First, how do you get 10x the investment from your Salesforce investment? We're going to take you through it. It's a vision we call Slackforce. We're excited to show it. Next, how are we going to unlock radical employee productivity? Not just employee productivity, but radical employee productivity with Slackbot. And then finally, as you just saw, we are in the age of builders. And we need to turn everyone into our companies to builders. We're going to take you through that. But let's start with Slackforce. Now, if you attended the main keynote, you heard Patrick Stokes talk about this little anecdote from 6 months ago from our lovely co-founder here who's paying direct attention to me. I can see the focus, the attention to this. Why should you log into Salesforce ever again? Maybe you never will. Now, Patrick kind of told you the story, but I'm going to tell you the behind-the-scenes story. On the Slack team, we were sitting around saying, we can do this. We actually can make this happen. But we knew we had to kind of get to work. So we started inventing. It started with Slack CRM. Now, Slack CRM, if you don't know, is native CRM capabilities built directly inside of Slack. Nothing additional purchase. If you have Biz+, you have it. AI-first, conversational-first CRM built right inside of Slack. And here's the best part. Because it's built quietly on Salesforce, It's ready to scale as you scale your small business. This is truly the front door to small businesses for Slack and Salesforce. Then we needed to take it up to all customers, and so we built a whole raft of MCP servers— Sales Cloud, Service Cloud, Agentforce, Data 360, Tableau. And then we did something really cool. We made Slackbot an MCP client. So now Slackbot can do everything you can do in Salesforce. All you have to do is just ask. But it's kind of just a little bit better because it's right in the flow of work, right there working inside of Salesforce. And I'll tell you, there's something magical that's happening, and it's kind of overwhelming at the same time. There's a period we're going to get to where there's going to be more agents inside of many companies than there are employees. How are our people going to navigate all those agents?

Which skills to use?

Which agent to pull at the right time? Well, Slackbot can speak agent. Slackbot knows the skills and capabilities of the agents deployed inside of your organization, and it can pull in the right agent at the right time. So as an employee, you just know, I just work with Slackbot. And this is why companies like Box say, 75 to 80% of our work can happen in Slack, in the work operating system. And this is how we're working today. Just 6 months after Parker said those words, 90% of our sellers use Salesforce from Slackbot without ever leaving the flow of work. And it's how companies are getting 10x the value from their Salesforce investment, because something really magical starts to happen when you can democratize access to the system of context, to that system of truth that is your customer. When you can make it easy to get to through just a prompt. People like me who've been in technology my whole life, I rarely saw Salesforce. I never logged into Salesforce. Today, I access Salesforce hundreds of times a week. My team accesses Salesforce hundreds of times a week because it's right there through a conversation. Now, there was one final piece that we knew to really deliver this front door, and this thing we call Slack for Surfaces, and this is transformative. If you can imagine it, Slackbot can build it. This is dynamic UI, on-demand composable UI that's multiplayer, And I'll tell you what, there's never been a better time for me to stop talking, because this is one you actually want to see. So, let's see a demo. So, let me bring up our Chief Product Officer, Jamie Delange, to show you Slack for Surfaces.

Hello, hello everybody. I am so excited to show you how Slack really becomes the front door to Salesforce. So, last year at Dreamforce, we announced Salesforce Channels. So, here you are in a channel, right here in Slack, and not only do you have access to all of the Salesforce information associated with this account, but you and the team are actively collaborating. You're strategizing in here. You're building documents, canvases, executive briefings, and all of that is right here in the channel. You can see every update happening to the record as it happens. This is great. This is amazing. It's already revolutionized how you work, right? But now you have something new, too. Up here in the tab bar, we have a surface, a Slack for Surface. So instead of normally what people do, where they pull in a bunch of documents from different sources to try to get a clear picture of the account, the team here just went to Slackbot and created a central health system for Presidio Tech. So we've all seen a lot of really wonderful AI-derived UI like this. This is the whole point of AI for us. But there are a few things that make Slack for Surfaces special. First, they're connected out of the box to all of your enterprise data. That means anything that's connected to Slack, whether it's Salesforce, it's Outlook, it's your other systems of record, all of that is connected here too. So you don't have to do any special connecting to make it work. Second, it's right here inside of Slack. which means it's in the flow of work. It also means that the team is going to be checking it. It's live, it's updating, and it's secure. This isn't on somebody's GitHub repo that everybody's going to forget to check the next day. This is something they can actually action on. And third, because we built it at Slack, we built it for the way that teams actually work. That means that this surface is collaborative by default. So, if I see something going wrong with the account, I can click right into the record, And not only that, but I can tag somebody else from my team in to help get this account on track right away. It's pretty incredible, right? So, that's amazing for the sales team, and Slack truly is the front door to Salesforce, but I work in product, and I want to know what's going on with my product launch. So, as the product lead, I would normally read through this entire channel, try to understand what's happening with feedback ahead of a launch. But I don't have time for that, so I'm just going to get Slackbot on the job. I'm going to ask Slackbot to pull together a surface that highlights all of the feedback, groups it by topic, and lets me know what we need to fix ahead of launch. Just like that, Slackbot is going to take that simple prompt, look at all of the feedback in the channel. It's also going to understand what we're actually doing because it has access to our systems of record, things like Jira, or Linear. It also has access to all of the product documentation, and it's going to pull together a beautiful surface that tells my team exactly where we need to focus ahead of launch day. So those bugs are going to get squashed before we hit production. Okay, but Ryan probably needs a surface too, right? The marketing team has things that they need to do. Maybe they want to build a beautiful interactive website instead of a slide deck ahead of a customer launch. Sure, Slackbot can do that. Look at this, it's gorgeous, it has our narrative, it has our branding, right out of the box, all with simple language, no development time needed. But you know, this is all really productive, this is great, we're getting stuff done, we're knocking out tasks, but the teams that actually win together, they know how to have fun together, right? So Slackbot can help us with that too. Slackbot actually built this interactive game where we can open up deal doors to close out the quarter and crush, all right? Crush. So basically anything your mind can dream of, surfaces can do. So my challenge to every single one of you is to think of at least 3 surfaces that you can build with your team. You could do it today, you could do it when you get back to the office next week, but start playing around, have some fun, and see what you can make. All right, Ryan, back to you.

Jamie. I predict the most popular words typed into Slack in the next little bit is going to be Slackbot, build me a Surface. Those 5 little words are going to be pretty powerful. All right, let's talk about how we bring AI to every employee. What is a company's most important asset?

People.

Their people. Someone listened to rehearsal. 30 to 60% of what companies spend their money on is their people. So what can happen when we can take our most important asset and make it 2x more productive or 3x more productive or maybe like 10x more productive? What do you need for that to actually happen? Because to be honest, it's not happening for a lot of companies today. Well, you need a consumer-grade AI that knows me, it knows my context, it knows my business, it knows my data, it knows my systems. Oh, and by the way, it needs to be budget-friendly. And what do I mean by budget-friendly? I mean one fixed cost, unlimited usage. And that is what Slackbot launched in January. It's the fastest growing feature in Salesforce history. I'm looking at our CRO right here who's been crushing it in that team. 1.5 million monthly active users. But what's really impressive is the number at the bottom, Nubank. 1 million company-wide hours saved per year through Slackbot. Our most engaged users inside of Salesforce are saying they save 20 hours a week. Is there any leader in the room that wouldn't want to give their people back 6 months of time? It's not like they're going on vacation. Let's be super, super clear. But they're not doing the work of work. They're not doing the look for the thing, go build me X, Y, and Z, get ready for this. They're doing the work that drives value to the company, the top-line revenue, driving value for customers, and lo and behold, actually doing work that's more fulfilling. And that's what's happening with Slackbot, and we are innovating unbelievably fast. We have a channel inside the Slack team called Slackbot Unbaked, where you can watch the team build in the open. You heard Boris talking about building in the open. You see the iterations happen. We had 50 features since we've launched. We have 50 more coming, and I'm sure the product team's gonna be like, Brian, it's way more than that. But I wanna tell you about 3 that are really, really cool that we're gonna show you here today. The first, you've all seen Slackbot take unbelievable context from your business and turn it into, like, a beautiful canvas or a document. How about it turns into a beautiful PowerPoint presentation from Microsoft, or beautiful Google Slides? Better yet, how about it actually turn it into a video? Slackbot can take all that unstructured data, that structured data, and turn it into these rich artifacts, so you can just focus on adding value, not the work of work. It's incredible. I can't wait for you to see it. This one's my favorite, not gonna lie. You talk to Slackbot, and Slackbot talks back. Not in a sassy way. Steve Fisher's here, he would know about talking sassy sometimes. Not in a sassy way. Slackbot talks back like your favorite colleague. In fact, it's even a little better than your favorite colleague, because it always does what you ask. Slackbot 2-way voice is amazing. And then finally, I'm here to dispel the rumors, it's true, Slackbot is coming to a channel near you. We've all been there, right? You're sitting in a channel, someone asks a question, You see 15 eyeball emojis and no one knows the answer? Well, guess what? @mention Slackbot, Slackbot knows the answer. The same rich intelligence you've had as a personal assistant now comes onto the team. And why is that important? Because when one person learns, they get smarter. But as Rob talked about with that example from Shopify, when a team learns, everyone gets smarter. And when a team learns, the whole company improves. And that's that compounding advantage. So to show you this in action, I'm going to bring up the incomparable Amy Bauer, our product lead from Slackbot. Amy, take it away.

You're such a gem, Ryan. All right, so we just saw Slackbot building dynamic surfaces, bringing together trusted data from across your connected systems, but I want to build upon that now and show you how Slackbot can adapt to the way that you're working wherever you are, whether you are on the go or you're working together with your team. So, it's Dreamforce week, right? And I'm sure, like many of you, I have been running around this place going from meeting to meeting, but I've also unfortunately not had a lot of time to prep in between those meetings. But I have had Slackbot, and I've been using that a lot to prep. So, let me show you what I've been doing. Hey, Slackbot, I'm about to go meet with Presidio Technologies. Can you give me the TL;DR on everything I need to know walking into that meeting?

On it. I'm pulling together the key notes now. Still fetching details.

Almost there.

I have checked Salesforce and looked through the files and messages. Just wrapping up. Here's the quick brief. Presidio Technologies is a $450 million tech company, and you're $220K into the cloud platform expansion deal. Now in negotiation with a September 25th close target. Elena Marsh, the CIO, is the exec sponsor. David Okafor, the IT Ops Director, is your champion. And Priya Nandakumar, VP of Procurement, controls sign-off and is the timing risk. Heads up, competitors are pricing 12 to 15% below you, and legal's still finalizing MSA redlines on liability cap and data residency.

Okay.

Yay!

Yeah, that's pretty cool.

Thank you. Now, what's interesting about this interaction is that Slackbot wasn't just summarizing the Salesforce record, right? That's synthesizing context from across Slack, Salesforce, and my connected systems, so that I'm actually pretty well armed walking into that meeting, which seems like there's a lot to know, because we want to close that deal at the end of the month. But I can now use that same interaction pattern after the meeting, to move that work forward or to do deeper analysis with Slackbot just while I'm on the go. So let's try that one out. All right, so guess what, bud? Meeting went great. I think we won Marcus over with the demo, but I'm going to need you to update the next steps in the deal because we're going to meet next Monday, and the pricing negotiations are probably going to get a little intense. So I think I'm going to need to loop Polly in, and I'm actually curious what you think we need to do to get ahead of the competition in those pricing negotiations.

Great news! Let me get that queued up. Still working on those deal updates and a quick plan for pricing. Almost there, just checking the deal record and team info. Still in progress. Thanks for your patience. I'm still on it, and I've just updated the record. Great news on Marcus. I've updated the deal's next steps to reflect the demo win and Monday's pricing negotiation. Polly Smith from Deal Desk is looped in. As for getting ahead, I'd lock in David and Elena to reinforce value and quantify your ROI edge so it doesn't turn into just a price fight. Also, close out the MSA redlines before Monday so pricing is the only open item on the table.

Yeah.

That's good advice. All right, here we go, Slackbot. So, this is where voice gets incredibly powerful, because if you notice, I didn't really have to organize the work before I gave it to Slackbot. I could simply talk to it very naturally, and Slackbot was able to parse through the information, what actions it needed to take, and what we needed to do next. But if you think about it, what needs to happen next is not really going to happen with just me alone. It's going to happen with my team. So, let's go back now into that account channel, and I'm going to have my awesome demo drivers, Daniel over here and Camilla, zoom in here so you can actually see that Slackbot is right here with the team in the account channel. It's here, it's researching, it's handling tasks, and it's helping them move account work forward. You heard Rob talk about this earlier, but there is something incredibly important about this work happening in the open. People can add context, they can redirect the work, but most importantly, they can learn how to use AI better by watching other people use it. So here we see the team is asking Slackbot to build a presentation. I know Ryan's super excited about it. I love doing it too. Build the presentation for that next customer touchpoint. So Slackbot is pulling together the relevant context, it's running the right skill to build that presentation, And it's leveraging MCP to pull in those ROI stories and those customer case studies. But we also see that another teammate jumps in here to say, hey, there's a newer corporate template available. But that doesn't trip up the work. Slackbot simply adjusts without missing a beat. When teams work like this in the open, it's searchable. It is discoverable by other people. It becomes context that Slackbot can now draw from in the future, and it becomes compounding intelligence for your entire organization. So across these experiences, you have seen Slackbot move from mobile to desktop, from understanding context to taking action, and from helping one person to working alongside your entire team. Ryan, back to you.

That was incredible. We're very, very lucky. We're actually joined by a very special guest here. Stripe is an unbelievable customer. It's one of these storied customers in history, and we actually have John Williams here, who's the general— head general manager of sales at Stripe, and he's going to be joined by Zach Ciliotti-Young, our CRO. These guys have learned how to maximize the value of Salesforce with Slack and Slackbot. So give us a quick little view, guys.

All right, all right. Yeah, welcome, John.

Happy to have you back at Dreamforce.

Thank you so much, Ryan.

Well, John, Stripe has been at the forefront of both payment and financial infrastructure for the internet for well over a decade, and enterprise sales has been key to that growth. So, tell us a little bit more about yourself, your org, and the transformation that your entire org has been on.

Yeah, thank you so much. I appreciate the opportunity to be here today, and hello, Dreamforce. At Stripe, we are— we're trying to change the way folks go about doing financial automation, and with that, our sellers are engaging day by day all the time. I run our enterprise business for our existing accounts, and the major problem with that is we're trying to move as fast as we possibly can through what we call our Stripe Way and customer engagement framework. And in that process, Zach, there's challenges. There's challenges to move fast. There's challenges to get to the customer first. It's a competitive landscape in fintech. It's competitive and it moves very quickly. And with that, we want to make sure we're bringing the full power of Stripe to the customer, to the user. We have a methodology, users-first operating principle, and in that, we're trying to, again, gather that information, bring a relevant point of view, and drive that with the customer, both qualitative and quantitatively in terms of business outcomes. Not easy. So in that process, we use Slack to really bring that high-level customer engagement framework to life.

Amazing. And one of the things that you've done better than anyone else, because I've got a chance to see it, is bringing Salesforce into Slack and into the flow of work. You have over 1,300 channels, you have 3,900 sellers, and all of your cross-collaboration teams working together in Slack. Tell us a little bit more about the business outcomes that you've been able to achieve.

Thank you, Zach. Yeah, it's a key unlock for us, right? We implemented channels recently, and it's been massive, massive. production component for us in terms of productivity, but also getting there faster. Not a lot of things keep me up at night, but if we're not at the customer first, that does, right?

Yeah.

Because that can lead to poor buying decisions, it can lead to technical debt, all the things that we don't want to happen for our user and customer base. So with this unlocking of channels, we're able to manufacture and get to the customer first and foremost, but with relevant data. So I'm going out to Boston next week, I'm meeting with a customer, we're having an EBC, and in that process, Yesterday, I just ran, give me an understanding, everything about this EVC. I'd like the full account plan for this customer and user, and let's see what happens. Boom, this thing just dropped everything for me. I've already been briefed on it. We have a brief next week. Not even necessary for it because we can async all the time through the power of Slack.

Amazing. And look, that's one of the things that a lot of companies struggle with is AI adoption because people, one, don't know how to prompt and they don't know how to bring all the tools into the flow of work. You've been exceptional at it. Your company sends over— Stripe sends over 62,000 messages to Slackbot every week.

Right.

What are you doing to drive that AI fluency? And what's one final thing that you would tell the audience they should be thinking about for the future of work?

Well, look, AI is here to stay. It's productivity. We still need the relationships through people. So we want all the work to transpire through our products with AI, but also through the tools that we use in go-to-market. But everybody uses Slack at our customer, from engineering, from coding on down the line, as Boris was mentioning earlier, we do that heavily as well. But this is a major, major piece for us. So we want that personal touch to happen more, as I was talking about customer engagement and framework on that front, right? So the more we're using Slack, the more AI that's transpiring, the more that's going to happen. That's going to be good for all of us.

Amazing. Well, we really appreciate your time here, Don, today. And look, Stripe's been on an amazing journey. We can't wait to see what's next. And you heard it here first, folks. Everyone in your company should have Slack. It is the best place for all of your work, all of your people, all of your agents to come together. So can't wait to see what's next for you.

Appreciate it, Zach. Thank you.

All right, let's give him a big hand. Okay, let's get to that 3rd chapter. I want to go all the way back to that conversation with Rob and Boris. How do we have humans and agents building together in the open? And we thought there was a really important innovation that was going to be required to really take that to the next level. It's called Slack Code. We released it a couple weeks ago. Take a look at this film. Slack is already where we do so much of our work. Code channels are taking this to the next level.

We've been using it to build our integration for GitHub, which is so cool and so meta. It allows you to bring Claude directly into a Slack channel as a member of your team.

Traditionally, software development has been an individual sport.

Slack Code changes that, and that is a reflection of how people now work. No matter if you're a developer, a designer, or just someone with an idea, Slack is the new evolution of multiplayer vibe coding.

With multiplayer, we're able to develop such creative coding projects.

With the added layer of AI and the tools we already know and love, it's an unbeatable combination.

Amazing. Amazing. What do you guys think? Slack Code is crazy. It's going to take single player into multiplayer. We actually launched it just a few short weeks ago with partners from Anthropic, obviously, OpenAI, GitHub, Vercel, Cognition. It's incredible because, as Rob mentioned, these companies aren't just running on Slack, they're building their products on Slack. And why? Because that's where their customers want to work. We have this incredibly rich ecosystem of 2,600 applications that's getting bigger every single day. Now, I don't actually want to talk anymore about Slack Code. I want you guys to see Slack Code. So for her very first Dreamforce keynote, please welcome the head product lead for Slack Code, Katie Steigman. Katie, get up.

Thank you so much, Ryan. Hi, everybody. I'm Katie Steigman, and I'm so excited to show you Slack Code today, because my team built and shipped Slack Code with Slack Code. So, you just saw Slackbot working in channel like a teammate, but we know the modern AI stack runs on many agents, whether that's Anthropic's Claude, Cognition's Devin, or Vercel. I'm going to show you what it looks like to work with an agent like Devin with Slack Code. Now, if you're not familiar with Devin, Devin is an autonomous AI coding agent that can do a bunch of cool stuff like building applications and writing code. Who here has worked with coding agents before? Some yes, some no. I know I have a bunch in the past couple months. Well, whether you have or not, I'm going to show you how to do that in Slack right now. So, let's say we're working on a launch and we want to update our website. I can tag in Devin without having to leave Slack to help me do that. So I'm going to ask it to update the website with the brand team's latest designs. And when I do that, Devin just doesn't respond in thread like it normally would. It does something new. It spins up a code channel, and it prepopulates that channel with the right context and the right teammates so we can get straight to work. Now, what the heck is a code channel? A code channel is a dedicated space for the whole team and the agent to work together on the build. In the code channel, the agent can do a bunch of cool stuff. It can build HTML artifacts, it can help us plan, it can preview the build, and it can do all of that without needing any specialized tools because everything we need is right in Slack. The way we work at Slack with our coding agents, they're set up with GitHub. So as a non-technical person, I can just tag in the coding agent and start working right away. It's really cool. It's really seamless. So we're working on our website. Requirements change, as they often do. The really neat thing here is that anyone in this channel can talk to Devin. So Akshay is going to chime in and direct the agent here. And this is the absolutely magical thing to me. Slack Code democratizes building. In this channel, maybe 2 engineers understand what's happening under the hood, but the other 6 teammates like me and Akshay, we don't know how to read and write code, and we don't have to.

We can just look at this preview.

We can give the agent feedback in plain language, and we can participate in the build. Coding is accessible to everyone now. How incredible is that? I know. I'm excited. But we're not going to get carried away, right? We still want to check in with engineering to make sure we haven't done anything too wild. So I'm going to tag in my engineering lead, Chris. Chris is going to take a look at the code. He'll open a code diff tab, and this is unique to code channels. You can actually see the code and review it right here in the channel. Crystal makes a few suggestions on how we can improve the code. Devin will apply those changes, and then we can see that preview updated right in the Slack channel. So, we're looking at a real-time HTML artifact in Slack of our website. So, once we have that all updated, let's take a look at our preview. What do we think? Should we ship this thing?

Do we think it looks good?

Should we ship it? Okay, let's ship it. So, what I'll do is I'll ask Devin to create a PR. Devin will create that PR, it'll document all of our changes into Canvas, and it'll push a summary back to the channel of everything we just did. Then, it will cleanly archive the code channel. Our task is done, our Slack has stayed tidy and focused, and because this happened in Slack, all of our conversations, our decisions, all that code we just wrote is 100% searchable by our entire organization. What was once individual learning is now team learning. So we just did all that. We just updated the website in a couple minutes. We got to a PR faster than ever. Pretty incredible. So I showed you this power with Devin today, but you can imagine this with Anthropic's Claude, with OpenAI's— oops. We have Anthropic's Claude and OpenAI. And also your custom agents. With Slack Code, anyone and everyone can become a builder, working together with agents in an incredible multiplayer experience. Back to you, Ryan.

What do you guys think? All right. I didn't tell you guys earlier, but amongst you, we are chock-full with special guests. They might be to your left, to your right, and actually one of them is right here. I actually want to bring up someone who's going to help bring all this together. Ali-Kay Miller, is maybe not just the most followed voices in AI. She's maybe the most importantly the trusted voice in AI. She advises some of the world's largest companies on their AI transformation. She was named by Time as one of the most influential people in AI. She's the CEO of Open Machine. Allie, come on up here, please.

Thank you. This is— hello.

Hello.

Do a full 360 for every answer.

Okay.

Twirl. Yes, we twirl when we get on stage. So, Allie, it feels like every 3 or 4 weeks you send me a text, and you've, like, unlocked some new magical way of running Open Machine with agents and Slack, and you're like, Ryan, check this out, and I'm always, like, blown away. So, can we start with a little bit of the art of the possible? Like, maybe how do you run Open Machine today?

Yeah, so we are, we are a lean and mean machine, and I made this transition a couple months ago where I told myself, If all I'm doing is repeating every single thing that we did in the human world, I have failed my team. And so we are trying to rethink our org structure with agents and what is happening proactively. Can we bring something up on the screen?

Yeah, go ahead and hit that clicker.

That's a little—

There you go.

OK. This is going to be an insane use case, but sure, why not? So every single 30 minutes, we have Claude going through and looking at all of our meeting transcripts. The goal is to codify every single thing we are doing. Every single thing should be legible to an agent. So this is a meeting that I had with a teammate, and we basically have like ghost agents that we're talking to, where into the ether we say, for the AI that's listening, can you make that Google Sheet happen? And then every 30 minutes, work gets done.

Wow.

So this is going to be a weird other hack. I highly recommend this one. When we are talking to these ghost agents, come up with some sort of word, a trigger word. We use pineapple. Do whatever makes you happy. And we are going to take all of that, Claude Code, which we integrated into Slack, is going to auto kick off all of these tasks, bring in summaries, bring in messages that I have to follow up. This is like a list of AI tips that I told my ghost agent to pull together. I had also told it to build that Google Sheet. It just makes it, it shares it with the right person.

Wow.

And every single thing is done without me prompting. The goal is to completely stop prompting.

Wow.

I never wanna prompt again. Even if it's 4 letters, of course.

That's incredible. And did you do, like, a final cleanup in Slack at the end?

Yeah. So, there was this whole Friday wrap-up that we decided to implement, which was we want to do this big sweep. And so, it's summarizing the summaries of the summaries, just so that if all else fails, I can just read this.

That's amazing.

And it's also going to tell me that I didn't actually send that message, and that I still have to send it.

That's amazing. Isn't that incredible, in terms of a new way of working? One of the things you've taught me is this difference between AI as a teammate and now, you're starting to think about AI as a proxy for you. Can you maybe share a little bit about that?

Yeah, this is, and it's a, there's a lot of data privacy, you know, additional concerns here. So I'm gonna show, I'm gonna tell my use case. I would not yet bring this into enterprise, but we have built something called Allie Loop. We would have a Ryan Loop if you were on our team. And it was stemmed from this discovery where my team was waiting on me to answer things and I might be on a flight. on a stage, and I just can't get back to them. And it dawned on me that I don't have to answer, but my context has to answer.

Oh.

So we built the ability for my team to message Ali Loop. And so they can message, hey, did Ali email Ryan to say how excited she is for the Slack keynote? And my AI's responding on my behalf. We've set it up with different guardrails, but there's an AI avatar of every single person on our team so that inside of Slack, we are talking to I guess, ghost Allie.

That's amazing. It's kind of mind-blowing to think, like, you're going to have your own AI representation that's working on your behalf.

To be clear, the goal is to look at every single bottleneck that exists in our business and to rip it to shreds. Like, that is what I want to do. So anytime that you feel yourself being held back, anytime you find that friction, do something with AI, probably in Slack, Claude, all those things together.

Amazing. All right, so wrap us up here. A lot of the people in the room are probably inspired, but also a little intimidated. Like, how do you get from here to there? Like, how do people make this step? Like, take us through that when you're advising companies.

Yeah, and I think, you know, looking at use cases like what Anthropic was showing is very helpful. I talk to Fortune 500 companies about raising the floor and raising the ceiling. So raising the ceiling is about having that frontier unit, and they are working in probably teams of 2 to 8. They are sprinting, they have higher budgets, they have different Slack channels, and as they are testing brand new features and surfaces and Slack bots and all those things, they are taking those learnings and bringing it to everyone else. So, that breakthrough becomes everyone else's starting point, and Slack is where all of these companies are actually making that connection, where they're taking those learnings and speeding up everyone.

That's amazing. Raise the floor, raise the ceiling. Ladies and gentlemen, Ali Kaye Miller. Thank you, Ali. All right, we are nothing here at Salesforce, and certainly at Slack, without our community, and right now, it's a time to recognize our incredible community, Gigi Chan, where are you? Gigi, come on up here. In fact, Rob, I'm gonna have you join me for this part as well. Gigi's out of Toronto. She's an incredible community leader. You've been helping to transform how AI works. You've been showcasing Slack so much. Would you mind just sharing just 10 seconds, 30 seconds about why you're so excited about Slack? Do you have a mic? Did we give you a mic?

Sorry.

No, we didn't give you a mic. Here we go.

Yeah. So bringing the agents into Slack really helps unlock productivity, right?

Because Slack is already where people are having conversations.

So it makes sense to bring the work where people are.

So instead of them going into different tabs and contexts to get the work done.

Yeah, well, I'm so glad you shared that, but I got to be honest, I kind of tricked you. And Rob, come on up here, buddy, because you're here under false pretense. You are an incredible member of our community, But you're here because you're the Golden Hoodie winner for this year. We'll do a quick picture here. Congratulations. Thank you so much. Thank you so much. All right, Rob, bring us home, brother.

Thank you all so much. Gigi, I'm so impressed by what you've been able to accomplish. Now, you all have seen a tremendous amount of innovation from us in here today. We have it listed on a slide that will be up here in a second. But I want to encourage you all, one, if you're an existing customer, go check out Slack CRM, and it's an entirely new way to use Salesforce within Slack. Connect your agent to Slack through Slack's MCP server. Go build a surface. If you don't have access to this, come see us. I'll show you some of the surfaces that I've built. Get ready for 2-way voice, and if you're using a coding agent, go now and start to use Slack Code. I hope we've shown you how Slack is the start and end of the modern AI stack. And for those of you that don't have access to Slack today and you're a Salesforce customer, please upgrade to AI Force Max, which will get you access to Slack and all the amazing features that you saw today. And if you're not already and you're on Slack, you can upgrade to Enterprise Plus. Thank you. Thank you to our demo drivers. Thank you to all of our guests. Thank you to all of you for being customers. If you didn't already and you like the, the Nike Dunk Lows, look under your seat. We're giving away 25 pairs of those sneakers and there may be a random ticket under your seat. Thank you, everybody.