Name: Brandon Metcalf, Asymbl
Uploaded: 2026-09-25T17:30:00.000Z
Duration: 15 min 19 s

More from Dreamforce 2026

Brandon Metcalf

Founder & CEO

Asymbl

14:52

17:09

17:10

Info

Brandon Metcalf, Asymbl

Brandon Metcalf

Founder & CEO Asymbl

George Gilbert

>> Palo Alto studio, connecting Silicon Valley and Wall Street. I'm John Furrier, the host of theCUBE here with Dave Vellante, my co-host.

Alison Kosik

>> Welcome back to theCUBE's live program and community coverage of Dreamforce 2026. I'm Alison Kosik alongside George Gilbert, and we are here streaming live from San Francisco. It's a beautiful day and we've got, the CEO of Asymbl, Brandon Metcalf, who is joining us on theCUBE. Welcome.

Brandon Metcalf

>> Thank you for having me.

Alison Kosik

>> Great to have you.

George Gilbert

>> So Brandon, we gather from you that you've, you've managed to hire like 100 people in 100 days with an agent. That sounds like one of those success stories where, you know, we just want to unpack it because what steps did the agent— was the agent responsible for? How much human involvement was there? And then how much effort did it take to set that up?

Brandon Metcalf

>> Yeah, so we did hire about 100 people in 100 days last year. We had a large client contract that we were onboarding and needed to fulfill. We had one human recruiter at the time, so capacity to hire recruiters, human recruiters, would have taken longer than we had to fill the contract. So we've been building recruitment technology on Salesforce is what our company does for decades, and we've been in a lot of beta programs and using AI just in general in our company. So we decided, let's do this agentically. Let's roll out a digital recruiter that could work alongside and elevate our human recruiters to actually fulfill the request. So help us source, help us screen, help us move candidates through the process, really be the second set of hands that our human recruiter needed.

Alison Kosik

>> So what made you comfortable trusting an AI agent with the full recruiting process?

Brandon Metcalf

>> we're pretty forward-thinking. We really look at how do we have a human-digital workforce across the board, not just for recruiting. Half of our workforce right now is digital across 13 different functions in the business. So playing around in the early days with ChatGPT and then into Claude and Anthropic, and then getting into the beta programs with Salesforce for, at the time, what was Einstein Copilot that became Agentforce, we really started to develop a perspective of what's the value of not just generative AI, but predictive AI? So then as we were using this, it just clicked in my mind. I'm like, this is the value that I'm getting is the same value I get out of employees. So if I think of this as an employee, if I give it a job description, a job to be done, if I understand the motivation for success for it to complete its job, and if I give a human manager the ability to coach and manage it, what can we do? And from there we've just grown an army of digital workers, if you will, And then Rosa, the name of our digital recruiter, she works with our recruiting software inside of Slack and inside of Salesforce to help us automate and streamline the entire process.

Alison Kosik

>> Good.

George Gilbert

>> So unpack some more. You were talking about sourcing and screening. So what does that look like? And then how did you have to program Rosa to do it that way? And then what are the functions sort of in that funnel that the human recruiter who sounds very highly leveraged now. What does that recruiter, their role look like?

Brandon Metcalf

>> Yeah, it's really up to the recruiter, the human, as to what they want it to be. So do we want Rosa to go out and reach out to candidates and get them interested in jobs? Do we want Rosa to ask them pre-screening questions and move them through that process? Do we want Rosa to schedule the interviews with not only the candidate, but the hiring manager, and then manage schedule changes and all of those types of things? Do we want Rosa to help with onboarding and collecting documents and doing all of that? So Rosa can do all of those different types of things and it just depends on the need that we have. Hiring slowed down, so Rosa doesn't need to do as much as what Rosa was doing and the human recruiter is doing more. It's also interesting because Rosa can work inside of Slack. Managers can talk in Slack and have conversations not just with Rosa but also the other digital workers. And Slack has really become our communication operating system, if you will, across the company.

George Gilbert

>> What are some of those other digital worker roles? That sounds really interesting that you're that far along that you've got, you know, specialization already.

Brandon Metcalf

>> So we have Polly People Ops. She answers People Ops questions through Slack. So what are my benefits, time off, all of those types of things. We have Teddy, which we've talked a lot about with Salesforce already. He's built on Agentforce. He's our BDR agent. So he does outbound, inbound for prospects and moving them into sales cycles. We even have customer-facing digital workers, which we call digital workers and not agents because one digital worker can have 100 agents inside of it to complete the job and actually execute the job. So we have a digital worker named Ben. Ben's a business analyst. So half of what we do in our business is we sell recruiting software to staffing firms. We also implement our software and we implement Salesforce software. And then we also help companies set up digital workers to run their business. So a lot of that is we have statement of works and we have client engagements. So Ben, probably our most successful digital worker, is a business analyst that we assign to a client project. Clients will pay us for Ben as software, but Ben is a full-fledged BA, does every part of a BA's job for the client. So requirements, answering questions. Ben really understands the full scope of what our client's business is, and he also understands the full scope of what the project that we're doing is. So he's this on-demand resource for answering questions and doing things in the project and he then stays with the customer after the project. Like when we do a full implementation, he has all of this knowledge that now the customer gets to retain and interact with, which is pretty incredible.

George Gilbert

>> What are some of those artifacts that the business analyst produces?

Brandon Metcalf

>> Well, yeah, if you think about it, we have to understand who their customer is, what their workflow is, why their workflow is what it is, what systems they use, where the data lives. Connecting all the dots. And then we also have to understand what was the system, what was the statement of work that we were delivering on? Why did we deliver it in that way? Think about how many times you've done a project and at the end of the project the whole team leaves and then someone from your company that was working with that team leaves. And then now no one knows really what happened during the project. Ben solves that. Ben's there consistently for the client.

Alison Kosik

>> Let's talk about Teddy for a moment. I understand Teddy drove a big jump in sales outreach engagement. Which I'm curious to hear what changed in your sales process to make that happen.

Brandon Metcalf

>> Yeah, we had a small sales team and we co-sell a lot with Salesforce and we needed horsepower. We didn't want to hire a large human team to do prospecting and managing inbound leads and all of that. So we embraced Teddy, which was the Agentforce SDR Agent out of the gate, and used him to actually have the conversations and convince customers to move forward in the sales cycle and then schedule the actual meeting between the prospect and our sales team to actually move it forward. Teddy was our first real endeavor of what we were defining as a digital worker. So we decided to name him back then. This was a couple of years ago when we launched him and we called him Theodore Frank. And we called him that because if he's doing a good job, we call him Teddy.

Alison Kosik

>> If he's being naughty, he's being called out as Theodore Frank, right?

Brandon Metcalf

>> Like your parents. And what that did was humanize him.

Alison Kosik

>> Probably makes it easier for employees to work with these agents.

Brandon Metcalf

>> allows employees to conceptualize this as another colleague I'm working with, which made Teddy more successful.

Alison Kosik

>> Mm-hmm, Brandon.

George Gilbert

>> So it sounds like you've built these over a couple of years. How has the experience of building the agents changed as the tools have matured?

Brandon Metcalf

>> It's changing right now. Yeah. It's a wild ride. We were definitely on the forefront. Of innovation and rebuilding things as things changed. I think we're starting to see some stability in things. We've also learned from what we coach our clients on, not everything needs to be agentic. We've learned that the use of software and deterministic and really just compute is so important. So we created a sales analyst named Stanley, and Stanley was built primarily in Claude and would work with Salesforce pulling data and all of that. And we had a challenge with Stanley where every time he ran, He would struggle trying to do the process better than what he did before versus actually doing the core job. And what we learned from that is, well, why is he processing data? Why don't we have software, Salesforce, process data and then have him read and tell us what the data is supposed to tell us? So we moved a big part of him onto compute and that was a huge unlock. We're spending less on tokens, getting higher results and much more quality consistency.

George Gilbert

>> So Are all your agents based on Agentforce? Okay. When do you use Agentforce? And then, what was the effort to put Data Cloud in to underpin, Agentforce?

Brandon Metcalf

>> Data 360 has been plugged in since the beginning. We really look at it, what's the job to be done? It, again, goes back to why we think of them as digital workers. Like, what's the job to be done? Where are we going to get the data from? What tools does this worker need to actually utilize to complete the workflow. Agentforce for us has been brilliant for execution, process execution, process automation, those types of things. We've leaned heavily on Anthropic and Claude for more reasoning and evolved thinking. Claude's also helped us think about what we should put back in compute and build back inside of Salesforce. So we're so ingrained in this, we actually have a role you might not have heard of before called the Chief Digital Labor Officer. His name is Shiv. He's running around somewhere. His whole job is to deploy digital workers across the company. So they're in 13 functions. Last year we had about a $5 million productivity impact from them. This year it's about a $13 million productivity impact. And it's really just changed. We're a small company, but we can act like a very big company.

George Gilbert

>> Jensen Huang has talked about HR onboarding human workers and digital workers.

Brandon Metcalf

>> Part of my head of people ops, her V2MOM, which is what we use internally to manage, half of it's written around how does she enable our human team to work better and more consistently with our digital team.

Alison Kosik

>> So beyond reducing headcount, which I'm assuming that's happened as this transition has happened.

Brandon Metcalf

>> Headcount's changed. So I don't think the story about AI is going to replace all the workers is true. I think everyone evolves. I wrote a blog post about this called The Talent Paradox, where I think talent's going to be harder to find. Because the talent's going to need to understand how to manage this new coworker. So we haven't really lost headcount. Headcount shifted and formed in different ways.

Alison Kosik

>> Okay, so beyond saving time then, perhaps, has AI helped Asymbl grow faster or generate more revenue? And where is that? Talk me through the measurability of that.

Brandon Metcalf

>> Yeah, it has saved us time, so I don't want to neglect that, but it's helped us innovate product faster. We rolled out a new process that we call OLE, which is outcome-led engineering, where instead of the traditional, you write a scope and you spend all this time developing and then you go into QA. Now we're just consistently prototyping. So what would take us a 4 to 8 week sprint is now a 1 to 2 day sprint for us to ship product that fast so we can innovate faster, we can sell faster, we can lean on these digital workers to take up more of the work that we don't want to spend our human judgment, our human taste, our human time on. And that's unlocked a lot, which is why we have them in so many different functions across the company and continuing to invest and grow. The secret, though, is you have to stay on top of them. You have to manage them. You have to coach them. Again, in my brain, this is an employee. The more I invest in making sure it's doing a good job and helping it do a good job, the better the results are.

Alison Kosik

>> Yeah, there's trust and accountability in this, right? do you manage these agents like employees and then are there goals and performance goals and oversight?

Brandon Metcalf

>> There's two sides of it. So you have the technical side, which Shiv manages from a tech standpoint, but then all of these digital workers are assigned to a manager, a human manager that rolls up into their department. So they have performance metrics of what they need to achieve in order for that leader to actually hit his or her goals of what they're focused on.

George Gilbert

>> We talk about unlocking $5 million and $12 or $13 million. So is it that you're growing and you're growing without having to hire as many people? And then individual contributors are now managers of digital labor and part of their development, professional development, is providing the right goals and measuring the output of their digital labor. And so part of the question is how do you upskill the employees to be managers who were individual contributors? And then how do you give them the infrastructure to measure the output? And so that their agents, they give them their objectives, but then their digital workers can learn how

Brandon Metcalf

>> to—Yeah, it's a lot of enablement. And it's also not just teaching them how to manage something. It's also teaching them how to manage the tech. And then we do have an engineering team that can help them with the actual prompting and coaching and doing all of that. But the big takeaway from the question is think about this, think about, so we had, when we rolled out Teddy, our digital SDR, Mitch, our human SDR, managed Teddy and Mitch had to get really good at teaching Teddy what good is, what's a good email, what's a good next step. What's a good phrasing to use?

George Gilbert

>> Yeah.

Brandon Metcalf

>> So as Mitch coached Teddy, guess what happened with Mitch? Mitch got better. Mitch then got promoted to an account executive. In his first month, I think he closed like 13 deals. So it elevated him because it teaches the employee, the manager, what good really is because they gotta get the agent to do it, gotta get the digital worker to do it. So from a metric standpoint, probably the biggest metric that we use is a metric that we call PPW, profit per worker. So we look at profitability across all of our workers, humans and digital, and that helps us contain costs and also look at actual profit to say what kind of value are we truly getting out of this hybrid workforce.

George Gilbert

>> So I'm just curious also about the technology infrastructure to enable all that. What sort of platforms did you need I mean, it's all over the map.

Brandon Metcalf

>> Shiv could definitely answer the question more elegantly than I can. Right now we're moving and have been moving to use Agentforce to help us manage all of these digital workers that are spread everywhere. But I think from my perspective as the CEO, it's what's the business outcome, what's the cost, and how do we manage it and govern it to make sure it's safe and secure? So a lot of our partnership and friendship with Salesforce has helped guide that. And it's been successful.

Alison Kosik

>> Thanks so much for stopping by theCUBE. Brandon Metcalf, CEO of Asymbl. Great to have you on theCUBE.

Brandon Metcalf

>> Thanks so much.

Alison Kosik

>> And you're watching theCUBE, NYSE Wired program and community coverage of Dreamforce 2026. We'll be right back.

Brandon Metcalf, Asymbl

George Gilbert

>> Palo Alto studio, connecting Silicon Valley and Wall Street. I'm John Furrier, the host of theCUBE here with Dave Vellante, my co-host.

Alison Kosik

Brandon Metcalf

>> Thank you for having me.

Alison Kosik

>> Great to have you.

George Gilbert

Brandon Metcalf

Alison Kosik

>> So what made you comfortable trusting an AI agent with the full recruiting process?

Brandon Metcalf

Alison Kosik

>> Good.

George Gilbert

Brandon Metcalf

George Gilbert

>> What are some of those other digital worker roles? That sounds really interesting that you're that far along that you've got, you know, specialization already.

Brandon Metcalf

George Gilbert

>> What are some of those artifacts that the business analyst produces?

Brandon Metcalf

Alison Kosik

>> Let's talk about Teddy for a moment. I understand Teddy drove a big jump in sales outreach engagement. Which I'm curious to hear what changed in your sales process to make that happen.

Brandon Metcalf

Alison Kosik

>> If he's being naughty, he's being called out as Theodore Frank, right?

Brandon Metcalf

>> Like your parents. And what that did was humanize him.

Alison Kosik

>> Probably makes it easier for employees to work with these agents.

Brandon Metcalf

>> allows employees to conceptualize this as another colleague I'm working with, which made Teddy more successful.

Alison Kosik

>> Mm-hmm, Brandon.

George Gilbert

>> So it sounds like you've built these over a couple of years. How has the experience of building the agents changed as the tools have matured?

Brandon Metcalf

George Gilbert

>> So Are all your agents based on Agentforce? Okay. When do you use Agentforce? And then, what was the effort to put Data Cloud in to underpin, Agentforce?

Brandon Metcalf

George Gilbert

>> Jensen Huang has talked about HR onboarding human workers and digital workers.

Brandon Metcalf

Alison Kosik

>> So beyond reducing headcount, which I'm assuming that's happened as this transition has happened.

Brandon Metcalf

Alison Kosik

>> Okay, so beyond saving time then, perhaps, has AI helped Asymbl grow faster or generate more revenue? And where is that? Talk me through the measurability of that.

Brandon Metcalf

Alison Kosik

>> Yeah, there's trust and accountability in this, right? do you manage these agents like employees and then are there goals and performance goals and oversight?

Brandon Metcalf

George Gilbert

Brandon Metcalf

George Gilbert

>> Yeah.

Brandon Metcalf

George Gilbert

>> So I'm just curious also about the technology infrastructure to enable all that. What sort of platforms did you need I mean, it's all over the map.

Brandon Metcalf

Alison Kosik

>> Thanks so much for stopping by theCUBE. Brandon Metcalf, CEO of Asymbl. Great to have you on theCUBE.

Brandon Metcalf

>> Thanks so much.

Alison Kosik

>> And you're watching theCUBE, NYSE Wired program and community coverage of Dreamforce 2026. We'll be right back.