Drive Seller Productivity for Any Revenue on Any Channel | Dreamforce 2025

You. My name is Brooke. I'm a solution engineer specializing in our high tech vertical and I am thrilled to be joined by Ashley and Kai from Anthropic. So today I'm going to be focusing on how Salesforce is enabling teams of seller productivity tools. I'll be passing it off to these wonderful ladies and they're going to talk about how Anthropic is approaching AI and how they are enabling it for their sales teams. I've been at Salesforce for eight years now and one of my favorite parts of the job is that I get to speak with customers every single day. Common thread is that sales teams are more are really under a lot of pressure right now. Sellers are facing rising targets. New tools, new technology, all being faced with higher customer expectations than ever. And despite all the innovation, over 70% of sellers time is spent on low value repetitive tasks. And when I talk with customers, they know AI can change this, but they're spending so many months figuring out the solutions, evaluating the different features and figuring out how it's all going to fit in with their existing tech stack. That's where Salesforce really stands apart. Over the last year, we've spent a lot of time really zeroing in on the high impact sales use cases that bring sellers and AI together. We're meeting sellers exactly where they are. So this is in Salesforce, in Slack, in partner portals, surfacing insights, recommendations and automation right in their flow of work. Because it's built on the Agentforce360 platform, you're not adding another tool to this tech stack. You're enhancing what your sellers are already using. Today, it's trusted, it's connected and it's a straightforward path to get AI in front of your sellers. Agent for sales unlocks growth in a variety of ways. Today we're focusing on five top use cases. Automating, prospecting, forecasting more accurately with bringing in siloed data, enriching account plans, scaling, coaching with dedicated AI and making the quoting process far more seamless. Let's start at the top of the funnel. Buyers today are more educated than ever. By the time they're ready to engage, they've already done their research and they're looking for fast answers that are relevant to their business. Now sales teams are struggling to keep up. Salesforce being one of those. I don't know if anyone here has heard the customer zero story around Salesforce SDR, but used to be that over 75% of inbound leads went untouched. That's crazy. Marc Benioff heard that number and he's talked to the leadership team saying, you have to change this. And they did with Agent for Sales. Now, when someone is getting on the website, Agent for Sales is autonomously answering questions, providing collateral like customer stories and qualifying prospects. 24 7. This is huge for our sales teams because they're getting calendars full of qualified meetings. Let's say the timing isn't necessarily right. That's fine. Our agents can continue to nurture those leads with personalized outreach that's grounded in prior conversations. By bringing agents to the top of the funnel, your sellers are spending less time chasing cold leads and more time with qualified buyers. This is ultimately shortening deal cycles. It's expanding the reach of your sales teams, and it's driving revenue with more qualified pipeline. There's something that I've heard very consistently across all of my customers is that they can't truly trust their forecast. And it's not for a lack of effort. Their sales data is fragmented. Oh, it's everywhere. Oh, that's. Is that my bad? No. Okay. Sales data is in Salesforce, but we know that's not the only place. It's in spreadsheets, it's in emails, it's in other tools. And that makes it really difficult for sales teams to make sure that their CRM is completely up to date. It's incredibly time consuming. This is also making leaders forecast their business with very limited insight into what's happening. So they're chasing their teams for updates, and they're often managing their pipeline out of spreadsheets as well. But it doesn't have to be that way. With agent Forest sales, you can eliminate so many of those manual tasks because you're bringing in data together and it's making sense of it. So we're analyzing calls, emails, notes, all in real time to automatically generate consistent opportunity updates on stages like Field Next steps. So not only is it impacting efficiency, but leaders are finally trusting the data that's in front of them so they can forecast with confidence and then they can coach their teams more strategically. All in the context of Salesforce account planning. I used to dread those words, but let's be honest, it has become a critical piece of growing accounts and driving customer success. We also know that customers are doing this inconsistently today. If there even is a process to do it, they're often siloed in slides and not revisited. Again, with Salesforce account plans, we're allowing teams to scale and standardize how their teams plan and grow, aligning on objectives, stakeholders and data all within Salesforce and Thanks to Generative AI, account planning doesn't have to be a burden anymore. That's why we have introduced account research directly into the process. Now, with just a click, sellers can get answers to key strategic questions. They can bring in trusted external data and automatically populate fields without ever leaving salesforce. This is huge. This is saving our sellers a massive amount of time and it's helping them craft more relevant strategic points of view. And. So what this is doing here because we need to come to the table a lot more focused on what's important to our customers. Our sellers are becoming better solution sellers. They're differentiating from the competition and they're improving win rates. Another trend we're seeing in high tech is that companies are spending a lot more on enablement. This is for new messaging, product launches, competitive intel, and it's a lot. Our sales teams are struggling to keep up. Learning it is one thing, but putting it into practice is completely different. Coaching helps, but we all know that our managers don't have the bandwidth to coach every rep and every deal. Salesforce coach helps scale what managers cannot do alone. Deliver personalized objective feedback that's grounded in in business data and real deals. So reps can practice pitches, they can strengthen their approach, they can practice their objection handling all while in the context of real deals. And because the feedback is immediate and actionable, they can figure out what it is that they are doing well today and areas that they can improve. Ultimately, a sales coach is making sellers much more confident in what they're doing and they are able to differentiate from the competition. Last but not least, let's wrap up with one of the most complex parts of selling and that's going to be quoting. This is by far the top use case that we're hearing from our customers that they want to automate, anthropic included. You'll be hearing from them shortly and it's for good reason. There's a lot of moving pieces when it comes to creating a quote. There's different revenue types, complex pricing configuration rules around pricing, and then complex approval workflows. So sellers are often finding themselves in a back and forth with the deal desk and finance just to get a quote that is accurate and approved. Now that's what I love, what we're building with agent force for sales. Sellers can simply type out what it is that they want to quote. And what that means is behind the scenes an agent is working to apply catalog rules. They are enforcing the pricing rules, discount thresholds, and of course respecting those role based permissions. So so we're ensuring accurate and compliant quotes. Now, not only are quotes going out the door faster, but we're ensuring faster deal cycles, tightening compliance, healthier margins, and overall confidence in the business in every single quote that goes out the door. All right, and there you have it. We just walked through five top sales use cases that we're really focused on to make sure that sellers are confident and efficient at every single stage of the sales life cycle. Now pass it off to these ladies to talk about how Anthropic is bringing some of these use cases to life.

Yeah.

Amazing. Thank you, Brooke.

Hi, everyone. Ashley Moore. I'm a solution engineer here at Salesforce and I'm joined by Kai from Anthropic. I'm very excited we get to chat with you because I know a lot of these things are very top of mind for you. And so why don't you tell us a little bit about your role and some of what you're trying to solve at Anthropic today.

Yeah, definitely. Thank you so much. I'm so excited to be here. My name is Kai, Kai Kelly. And I'm responsible for business systems and related infrastructure. Anthropic. And one of our biggest challenges is that we are a very, very quickly growing company. And I mean, we went from having no product to, as Dario said yesterday, announced billions of dollars of revenue, and we're onboarding new AES constantly, which just means that there's not a lot of time to help them learn how to sell and to guide them and make sure that our Salesforce instance is easy for them to use. And so that's just like a big challenge that we're hoping to solve with agent force.

Yeah, definitely. Your growth rates have been massive, and I think it's still relevant for everyone because everyone is trying to expand what their teams can do and accomplish. And I'm curious, for your AI use cases that you guys are implementing for your sales team, where are you starting? What are some of the top ones you've been looking at?

Yeah, I mean, just as Brooke said, one of the top ones is the quoting agent. That's actually one of our most requested use case and one that we are actively implementing as well. And we're actually calling it deal desk agent as well. And it's more of a kind of sales ops deal desk help guider, like where somebody can ask questions. Is this deal standard? What should I do to sell that? That's another one, actually. Just was looking at the inbound SDR agent. That looks super interesting. We're gonna Be meeting about that later. But another thing that was actually implemented internally, not actually using AgentForce, but by some folks on the larger team is an account research and account planning, which is really, really helpful for the sales teams. I mean to have the next steps and prep for meetings. And so there's all kinds of really awesome use cases that are being worked on and it's hard to do them all.

Yeah, I know.

I mean, speaking of that, how are you guys prioritizing? I know a lot of people have a lot of ideas. Some people just don't know where to start.

Yeah, I mean, so what my approach and I think the team is there are some kind of low hanging fruit, ones that might require less effort, a lighter lift, but have a higher impact. And so there are different kind of ways that we would be thinking about it. Additionally, we're very lucky because as an AI company we have a lot of folks who are really interested in getting hands on themselves. We also have a very open environment if somebody wants to play around and develop something in some kind of sandbox. We have a bunch of critical tests, we have a lot of implementation guidelines and if it fits, we will actually go live with it. But in terms of what we prioritize to focus our time on is generally highest impact things like quoting agent, you know, like account planning, things that we know will like move the needle in terms of revenue or just save a lot of time.

Yeah, definitely. I'm glad you brought up moving the needle because something I've also heard from some other folks is really wanting to understand the ROI on AI. What does this mean for this business and monitoring what's necessary and iterating on it. What are some strategies you guys have for that?

So it kind of depends on what you are trying to solve, like what aspects of the business you're trying to solve with AI. So for example, if you're doing something around deals, it is easy to measure the deal life cycle, how long you spend in each stage, how you can shorten that, how you can increase the number of quotes. That's definitely something that would be helped with a quoting agent. But then there's also just a number of times somebody on my team or I have to help somebody. And so that's another great way to measure how agents can be assisting because we were actually measuring and looking at trends and people reaching out to us on Slack or any kind of meetings or recurring questions and we can actually measure the decrease in the trends and also decrease the number of questions and areas and problem areas as they Go down as we increase automation or agents around that.

Yeah, I love that. I love that you guys are also looking at your Slack data. I know we've had a lot of announcements at Dreamforce around Command center and being able to monitor your agents, but there is a slew of data in your company that you can be using to figure this out.

It's very exciting.

I'm also curious in terms of change because as you're releasing AI, some people are faster to adopt. I know things may be a little different at anthropic in terms of AI adoption, but what do you recommend for people managing change and trying to encourage folks to take on more of those AI use cases?

Yeah, my recommendation would be to go with a crawl walk run solution. And that's just because not everyone will initially trust kind of any kind of AI or agents. And so for example, what we're doing first is we're implementing more of a walk situation in which is more of somebody who's helping and answering questions. And then the next step is the automation, automatically creating the quote, adding the products. And then to me, the third is the more proactive noticing trends in the accounts and this account's ready for an upsell, et cetera, and proactively creating the quotes. And so I think it's just easing your way through, building trust around there. And then another one, a really important thing is having beta program. You can have a small group of beta testers, get a lot of feedback, get people excited, and then after you've gone through that iterative kind of process, then you can release it to the larger group and by then you have a kind of more fully flush out, I guess, tool at that point. And so those are kind of some of the ways I would say it would be much easier. The other thing is, I know this is also covered a lot is having more automated but also having human intervention and I think having a nice balance of both because I think if you swing too far, like having all the emails sent out with any kind of potential for human to give some feedback or to say, yes, no, this is not in the right direction. I think it's easy to lose trust. And so having a good balance of both of those and as you train your agents to be smarter and smarter.

Yeah, humans and agents working together. And I heard you were working on an agent this week. It sounded like the crawl, perhaps.

Yes, yeah, yes. We were working on it last Friday and actually had been doing it this week. It's really fun. We are trying to automate a lot of our we have a kind of rev ops chatter where people ask lots of questions about things and it takes a lot of time and a lot of it is documented or referenced. And so what we're trying to do is kind of have an agent who can reply and direct people to write questions so that we can focus on the larger projects. And so it's been something that we've been working with Ashley and different members on Salesforce to build.

It's been fun.

And I think my favorite part is how long it took you to build the first iteration of this agent. Will you share it with us?

I think it took us around 10 minutes to get it going. This is amazing. The first iteration was like, oh, this is so easy. Yes. It was really fun.

Yeah.

And then there's the QA and the change management. But it's really exciting and it's great to see you guys following that crawl, walk, run approach. I think that's a great takeaway for us. I know we're kind of getting to the end. Thank you so much, Kai for sharing all your insights today. We're so excited.

I don't know if you guys knew.

Also agentforce runs on anthropic models.

Very exciting.

Yes.

So you have a lot of optionality there and we are so excited to be part the of partnering together. I know we're going to wrap up with a couple takeaways for you guys. I know there's been a lot to take in at Dreamforce. If you're looking for more, download our tech trends report. We've got this QR code up for you so you can see other trends in our industry and also be thinking about as you leave here what use cases are sticking out to you, what would be high impact for your business, what's important for your teams and work with your account team as well. We love to work together to help bring that to life. Thank you all.