Go back

Why Software Sales Playbooks Break in AI Markets with Alex Varel

1h 1m 47s

Why Software Sales Playbooks Break in AI Markets with Alex Varel

In this Revenue Builders replay, hosts John McMahon and John Kaplan speak with Alex Verrill, EVP of Worldwide Sales at Cerebras Systems, about leading sales in the AI-native era after scaling teams at MongoDB and Zscaler. Verrill explains that Cerebras built the world's largest chip—an entire silicon wafer—which co-locates memory and compute to eliminate the memory bandwidth bottleneck that limits AI inference. As the market shifts from training models to using them, latency becomes the decisive factor, much like the broadband revolution, and there is no market for slow search or slow agentic AI. Verrill admits to initial imposter syndrome but argues that five universal sales pillars—pipeline generation, courageous qualification, messaging mastery, dynamic ICP management, and accurate forecasting—apply everywhere; what changes is the speed of adaptability. Cerebras revises its ICP every few weeks, tracking funding flows to bet on well-capitalized AI-native companies while still pursuing enterprise accounts. Sellers must become value translators who connect technical speed to business outcomes, and those who train their own AI agents to automate administrative work will reclaim selling time. Verrill advises leaders and sellers to build personal learning paths around AI, stay curious, and attach themselves to their customers' biggest business issues, since human skills and execution fundamentals remain future-proof.

Transcription

10106 Words, 54425 Characters

English
Speaker 1No matter the technology, we're going to have to go out and attack the market and generate a pipeline. We're going to have to have the courageous ability to qualify the pipeline. We're going to have to stay on top of mastery of messaging. We're going to have to manage the ICP, which is extremely dynamic in this space, by the way. We can talk about that, but I have to manage the territory and prioritize the territory. And then fifth and final, no matter where I'm at, I'm going to have to deliver an accurate forecast.
Speaker 2Welcome to the Revenue Builders podcast, a weekly show featuring B2B sales leaders and executives. Hosted by five-time CRO John McMahon and Force Management co-founder John Kaplan, the show takes guests in the barrel, behind the scenes with the people who've been there, done that, and seen the results. Revenue Builders covers best practices for scaling and growing your business while sharing the pitfalls to avoid. Enjoy today's episode.
Speaker 3Welcome to Revenue Builders, hosted by John Kaplan and John McMahon. In this Revenue Builders replay, we're revisiting a conversation from earlier this year with Alex Verrill, EVP of Worldwide Sales at Cerebris Systems, the company behind the world's largest AI chip, which has since gone public on the NASDAQ. Before Cerebris, Alex built and scaled sales teams at MongoDB and Zscaler through IPO and hypergrowth. So he knows the software playbook well, and exactly where it stops working. In this conversation, Alex breaks down why traditional software leaders get exposed in the AI native markets, why his team revises its ICP in weeks instead of months, and why the AE who can turn technical speed into business value matters more now, not less. If you're building your 2027 sales plan, this one is worth a second listen. Let's dive in.
Speaker 4Alex, dude, it's great to see you again. Thanks for coming back on. Johnny and I have been waiting for this one. I'm excited to be here.
Speaker 1It's great to see the both of you. Happy Tuesday.
Speaker 4Happy Tuesday. So buddy, how do you want to kick it off?
Speaker 1I want to talk politics. I want to talk religion. I want to talk narco-trafficking. I want to talk about tax dollars on hospice centers and daycare and learning centers. And then I figured we, no, I'm kidding. Oh, that's a good idea. Yeah. Figured that's not talked about enough. We might as well cover it.
Speaker 4Well, we know we got our audience in Texas covered.
Speaker 1Yeah. You might understand where I would come in on those topics. But so all I want to do just to open up quickly, because the only time that I've been on the show, I read a little bit of poetry and I thought that I would start some sort of pattern or tradition. I don't know if it took off, but I wanted to read something just really quickly and then we can jump to business if that's okay.
Speaker 4Buddy, I appreciate it. John, it might be putting like perfume on a pig for Johnny, but let's try it. Okay. Okay. Okay. Yeah. Shots fired.
Speaker 1So Joseph Campbell, the power of myth. He had an interview with Bill Moyers back in the eighties. He was a professor of mythology, extremely influential on a number of levels. And anyways, I'm going to give you this excerpt from the hero's adventure. He studies unbelievable observations around them. And he says in the hero's adventure out of the power of myth, this interview where Bill Moyers is just going, going deep down mythology with him. He says, furthermore, we have not even to risk the adventure alone, but the heroes of all time have gone before us. The labyrinth is thoroughly known. We have only to follow the thread of the path and where we had thought to find an abomination. We shall find a God and where we had thought to slay another. We shall slay ourselves where we had thought to travel outward. We will come to the center of our own existence and where we had thought to be alone. We will be with all the world. And if I turn anybody to Joseph Campbell's work, my job is done. And I hope that maybe caused a few people to lean into that. Okay. But we can get to business now.
Speaker 4Well done. Well done, buddy. Real quick, though, I would like to, I love how you start these off like that. Tell us, just synthesize at the top level from the heart, what's it mean to you?
Speaker 1Well, McMahon, I told you this at a dinner table a long time ago when I was at MongoDB. Much to the dismay of my dad, I decided to shift from being an economics major to an English major.
Speaker 5And my dad was like- In the driveway.
Speaker 1Yeah, he made me move the Cadillac that I bought right after I graduated. I brought it to the house. I pulled it in the driveway. I was so proud because I had this job in sales and I was making money and I was going to take on a big car note. And I had the Cadillac in the front, just real quick for your listeners. I came in, I was like, yeah, I was like, dad, dad, I bought this Cadillac and it was the big Cadillac DTS sedan. It was black and it was on these really shiny wheels. And he says, let me take a look. And I throw him the keys and he goes out to the front. He looks at it, comes back in like five seconds later. He throws the keys back to me. He goes, son, you're 23 years old. You have five seconds to get that car out of my driveway and move it somewhere down the street. And I was like, oh, okay. Because Alex went into
Speaker 5sales, what his dad used to always ask him is, okay, you're going into sales? All right. Yeah. Just tell me what restaurants you're going to bartend in.
Speaker 1That's right. That's right. What, with this English major, what restaurant are you going to work at? I'll try to get your table and tip you from time to time. He was really nervous, but I, I, um, uh, you know, have an appreciation, I guess, a Kaplan to answer your question for literature and for escaping a little bit in that way and slowing down this motor. And, um, it helps to, I think, offset my obsession with work. And so reading stuff like that and thinking on those things, why it's the motor a little bit. And it's helpful for me kind of like exercises. This, this gets me out of myself. So that's, that's the reason behind it. Love it. Love it.
Speaker 4And so this in the spirit of reflection, you spent your career at big powerhouse software giants like Mongo and Zscaler. And then you made a highly reflective pivot. So when I first started talking to you about like, what did you do? Like, you had been thinking about this in the spirit of your reflection. You had been thinking about this for quite a while. So tell us a little bit about the pivot that you made and, you know, why you're excited about it.
Speaker 1Sure. So I think I was fortunate to have really good timing, um, uh, and intersection with great talent at MongoDB and Zscaler. And then I had really cool exposure to, kind of the upskilling of human capital through a couple of venture backed companies in that space. Right. Like I, and I found that to be really fascinating. Um, in 2016, there was a book written by Kai-Fu Lee, uh, it was, it was written in 2018. In 2018, there's a book written by Kai-Fu Lee, um, that talked about the AI super wars and like the battle between China and the United States 2018. And, um, and the story goes quickly, sorry, cause this, this, this is the reason that the spark or when the spark occurred, maybe, um, in 2016, alpha, uh, uh, Google's deep mind, a company they acquired an AI company built a piece of software, a piece of AI called alpha go. And in 2016, that piece of software, that AI beat Lisa doll, the world champion of go, which is an extremely complex, uh, game, far more complex than chess and its configurations and possibilities. So this piece of AI beat the world's most popular, most complicated game go, and it woke China up. It was like a massive event in 2016. The number of viewers that watch this match, um, uh, is like hundreds of times what we would watch in a, in a premier league game or many thousands. Thousands times more than what you'd watch in a super bowl game. And China woke up and they started throwing an unbelievable amount of money into developing AI and new economies focused around AI. And I remember reading that story. And then I watched the video, um, of, of alpha go beating Lisa doll and how everybody freaked out. And I was like, this is really cool. So I started to lean in there. And, um, I saw maybe with the human upskilling, that that was a really important thing so that we could coexist and augment ourselves with AI and supervise AI. And then I got a phone call about this opportunity to be at the, um, the compute layer of this world and what's happening right now. And the more that I understood, um, what was invented and what was being brought to market, the more I leaned into it. And I think that agentic AI and its velocity will be the biggest creator of GDP growth versus, uh, any, any sort of software company or training of individuals and things like that.
Speaker 4So you wind up at, um cerebras and i'd say your timing's pretty good brother with the sales apocalypse and you know i think the way you were describing it to me like you know the software only world um just set the set the agenda for the topic that you know that we're going to talk about like you know and why it was good timing in your mind i'm i'm really fortunate um
Speaker 1to get the phone call from alan and then to get introduced to cerebrist systems um and i'll set the stage for the listener to understand kind of what what's what's going on here and um and realize that like i don't have any great i have to have this moment with you all of a little bit of imposter syndrome okay and like humility because i'm going to describe the inception of cerebrist systems and um and why that matters and how important that is um so i report to andrew feldman who's the ceo and co-founder he's one of five co-founders of cerebrist systems and i talk to him many times per week and if if you all sincerely haven't watched him on podcasts um with all in or with harry stevens and the the various ones that he has time to to do it's it's pretty phenomenal like the way he explains the ai landscape the world that we live in its complexities he breaks it down um and simplifies it and makes it relatable and understandable unlike any way that i ever could right and like and i'm surrounded by him and the co-founders and our engineering team and our product leads and these scientists and i get to watch them go to work and talk about this stuff and so um uh for me to describe it is like it's uh it pales in comparison to what they have but to go go back 10 years ago okay in 2016 the same year that open ai was founded the same year that cerebrist systems was founded and gary and sean and jp and michael and and andrew saw that ai was fundamentally a different workload than traditional computing okay this is 10 years ago and they saw that the math was simple but moving massive amounts of data required for the mathematical calculations was really hard but they realized that that was going to be something extremely important and so they set out to build the world's largest chip i have it down here this is the world's largest chip okay this is the entire silicon wafer it's like the size of a dinner plate if you can't tell tell by the scale of my large hands here and so why who cares what does it matter if it's the world's largest chip it matters immensely when you have a chip of this size you can co-locate the memory and the compute next to each other and that's what i'm talking about so you can move data between the two extremely fast you're not leaving to go off to thousands of gpus over a wire and so what who cares well the world has shifted and by the way this was attempted by the world's best and brightest scientists for 70 years to do that to use the entire silicon wafer and not have any failures across that way from be able to get yield out of it they did it they cracked the silicon wafer and they did it they cracked the silicon wafer and the reason that that matters now and the reason that their foresight is so brilliant and alan and i and the rest of the company are really lucky to represent this the world has shifted from ai training to ai inference right massive amounts of resources have been thrown at training the models that all of us use right now and it's like every week you read about one model beating the other and then you get open weight models out of china that come out that were apparently trained for pennies on the dollar that are far more intelligent and beaten benchmarks now the world's pivoted to using those models and everyone's using them now our grandparents our pets our kids and the requests are happening more frequently by more users and the complexity of the requests are are much greater and the whole bottleneck with ai inference is memory bandwidth because you have to move i'm really nerding out here but stay with me just to generate one token one one word one part of a word on a small model a 70 billion parameter model that is moving 140 gigabytes of data from compute to memory for just one word and you got to move all the weights and all that data again just to generate the next word and and inference is all about memory bandwidth and the way that they architected the world's largest chip it solves for that it has an incredible amount of memory bandwidth they eliminated the memory wall bottleneck and i started to learn about that and learn about these ai supercomputers that contain these chips and i leaned into it and i was like i gotta do this i gotta do this
Speaker 4well johnny you remember when we first started art well you started the journey first and then you were kind of bringing me up to speed up and then you were kind of bringing me up to speed up on stuff and you shared with me something on like where the cost of all this was going to go and and you you told me about like i remember like just how many billions of times somebody responds to a model that says thank you and the billions of dollars like i couldn't comprehend it when you first were sharing with me like the billions of dollars because i do it when i use the models i'm like hey that was great thank you and like you're wasting money in electricity good
Speaker 5morning to you chat gpt right i mean that's a cost of organization a lot of money so what
Speaker 4alex just shared with us like you were talking about hey john you don't understand like how much that costs and how much that you know electricity that uses and then when i started to dig in with alex over you know the last few months or whatever he's telling me about the you know the speed that is required and and what did you say to me about latency you said there's no what did you say to me oh no
Speaker 1yeah andrew talks about this this is why you know we we believe the world believes that you know i talked about the shift from ai training to ai inference right like you're using the models you have really cool applications on top of the models and using them and it's like the broad broadband shift that's what's happening in ai right now like the the the real engine behind the experience the speed of that engine is really what matters now right and like there's no market for slow search right what is the market for slow search there there isn't one and um the same will hold true for agentic ai and uh agentic reasoning our expectation as humans just like we expect with fast internet and fast search our expectation is to be able to interact with um ai in human clock speed responsiveness and that's what we do that's what we do best and no number of gpu's stitched together can match
Speaker 4that the reason why i like where we started here is because i think the average listener including me in my journey you know a couple of years ago or what have you would be i don't understand what's going on behind the scenes and the the why i love the way you shared this is when you know when you're sitting there listening to this a lot of people are enamored with models and i don't know what's going on behind the scenes and i don't know you know what i think how many times did you get a phone call and go alex what are you doing you're going to a hardware company so you're going from software company to a hardware company and the and the thought process behind that is why is alex going backwards yeah you know and it's not it's not even close to being this is where the real pain is in ai and the solvable pain right now is huge right here i don't mean to be
Speaker 1controversial but i i think that if you know if we used to look at like compute storage and networking as commodities and and then i don't know i used to knock hardware sellers i don't know if you all ever did you know like what box pushers or whatever i don't know i was one back in the day okay um and you know i mean look at the look at the models they're reaching like reasoning parity like they're they're converging onto each other and their intelligence and accuracy and reliability and so and then you have a a ton of ai natives that are building all these applications things are kind of crowded right now and um you know i think it's really interesting um if if that reasoning parity is being reached and these models and all these applications have a lot of overlap um to me the real differentiator left is the speed and scale of the engine and the sophistication of the workloads it can support and so i found the prospect of that to be really really interesting well you
Speaker 5talked about memory bandwidth what about power consumption that's a big issue right now with ai does the chip from cerebrus help with minimizing power consumption
Speaker 1it is the power draw is uh significantly more efficient significantly reduced um imagine that you know you don't have a requirement for thousands of chips that are stitched together you know running a bunch of data across across them back and forth yeah the the efficiency savings are very real and very material for sure.
Speaker 5What about cooling?
Speaker 1Another efficient benefit to the way that this was architected, it's liquid cooled. And the world is moving to that from air cooled data centers to liquid cooled data centers. You know, you can have that water more or less recycled constantly through the data center and chilled to the appropriate temperature. That is far more efficient as well.
Speaker 4And all of the new data centers that are being built just for the listeners and all the arguments that are going on on power consumption, like it's a little bit of hype in some cases because they're all being built liquid cooled, right?
Speaker 1Yeah. Yeah. I think that's a definite benefit. And, you know, I don't think there's a shortage of power. It's just, it's getting the power to the data center locations and doing so in an efficient way. I mean, there's plenty of power to be extracted from multiple sources. There's lots, lots of power and energy that's wasted every day. You know, the infrastructure has got to, got to be modernized and efficient and brought to the data center locations. But I'm going to outpun.
Speaker 5Companies like GE, Vrnova, Bloom Energy, Vertiv, all those types of companies, Constellation Energy, all their stocks are taken off.
Speaker 4Thank you, John McMahon. That's right. That is right. Yeah. I listen to you, buddy. I listen to you. All right. Yeah. Hey, you were talking, you were talking about getting controversial. Can I get a little controversial? Yeah, please. I love that. Let me get a little controversial. There's a provocative idea circulating that I hear people say that a significant percentage of software sellers today are really at risk, whether it's a year from now, two years from now in this new world. Like, first of all, what's provocative about that? I know that it's, you know, for somebody like me, it catches my attention. Where do you think that's going to be? I don't know. I don't know. I don't know. I don't know. I think that's coming
Speaker 1from one example right now, like Jimmy Lee, my RevOps guy. This is our third company together. You met Jimmy in London a while back. And, you know, he trained like five or six agents over Easter weekend while his family was asleep. He gets up at like four in the morning and just goes down a very curious path. And he trained four or five agents that are looking at things. He's got one for voice of the customers, some product feedback going through JIRA tickets and things like that. He's got one for some forecasting. He's got another one to monitor solution architecture work. And he's got six that are in the works, right? So think like 10 inside of a week that is going to augment his productivity immensely. They are going off and conducting tasks on his behalf. I'm going to build and train a few this week myself, get hands on keyboard. And I think the understanding how to augment your own productivity in your profession matters no matter what, especially for sellers. And if I can imagine like my daughter is a sophomore in college in two years time, I would really encourage her to go into an interview and almost like an artist would lay out a portfolio of work that they've done. Or if they're a model, they showed like the shots that they've done. I want her to come and show the agents that are multipliers of herself, extensions of herself, and show what she can bring to the table that equals the productivity of, I don't know, five or 10 humans, her plus her agents. And I think sellers have got to get really curious about this space. Those that are potentially at risk are those who don't have hands-on keyboard experience. And I think that's a really important part of this. And I think that's and don't know how to augment their productivity or supervise a genetic AI in their job, right? To just make them more capable. The great thing is, though, I think humans are going to crave, I just got back from London and Stockholm meeting with prospects and customers. By the way, the tech in Stockholm is incredible. I don't know. The minds there are brilliant. I don't want to call out specific logos, but what's happening in Stockholm is fascinating. Humans are going to crave sellers' interpersonal and intrapersonal capabilities and communication skills and just being face-to-face with them. The email is dead. It's just been taken over by AI slop. We're going to crave some human fallibility and some misspellings and custom-made emails. Emails and face-to-face meetings. I think it's two parts that really matter. Getting hands on keyboard, knowing how valuable.
Speaker 4This is where you wound up today. You've had discussions with me in the beginning where you would say, you called it, you said imposter syndrome. You said, Cap, I'm feeling exposed. Walk me through that experience and how you had to level set yourself. Because if you stayed in that exposure world, then. You would have just gotten, you know, and the people that are funding this and the people that are, you know, they have certain ideas about playbooks and the way software sales was done. Just give me your perspective on that. And I think you're calling it like a technical athlete is really what we all need to become.
Speaker 1I think that's it. I mean, there was a tweet. I don't want to name names, but there was a tweet that kind of rocked my world from a highly influential person who I love and admire. And he basically. He basically said that, like, I'm seeing these traditional software leaders go over to AI-native companies or AI-centric companies, and they're trying to transfer their playbooks, and they're just getting creamed. Like, I'm paraphrasing. And so I think there is, like, an adaptability and situational awareness when you come into an AI-native company or if you make a big pivot from, you know, being a heralded software seller into something like representing, you know, the world's fastest inference, the world's fastest compute. There's a lot to learn, and you have to. And you all encourage this in your work and, you know, in John's book. Like, you've got to come in and really assess the situation, you know? It's like John when. McMahon, when he talked about coming in and, you know, you assess the talent, right? Assess the situation around you. Have a really good understanding of strengths and weaknesses and vulnerabilities around you. What can you. What can you not affect? And I came into the company, and I'll never forget the first executive meeting. It's, like, four of the five co-founders, or they walk in. I'm just sitting in this big kind of room, and four of the five co-founders come in, and then there's Natalia, there's Angela, there's Jessica, these unbelievable product leaders and machine learning field leaders, and it's just really quiet. And I'm like, oh, my gosh, this is really intimidating. And I realized that there was a steep learning curve that was going to be required and that I didn't want to jump to conclusions too soon, right? And that I was going to have to go out and try to scale this thing and attack the market, and I was going to have to hire sellers for sure. But the sellers were going to come into the same situation as me because this is novel territory. I went and pillaged Snowflake. And I. Because, I mean, what an incredible company. Like, they had a very different novel architecture, and they trained elite sellers. I'm like, I love that profile combined with EMC because I love killers from Boston. I think the whole EMC sales culture is in, like, you know, the folks that have come out of EMC. But I found this sweet spot. I'm going off on a tangent of, like, Snowflake plus EMC. And then get in here because they have the intangibles. They have the skills that are transferable by pillars. They have to, no matter the technology, we're going to have to go out and attack the market and generate a pipeline. We can talk about that. But I'm going to have to manage the territory and prioritize the territory. And then, you know, I'm going to have to go out and attack the market and generate a pipeline. The fifth and final, no matter where I'm at, I'm going to have to deliver an accurate forecast.
Speaker 4Right? But I knew I was going to bring. Which is thousands of years old. So I love Johnny and I have these podcasts. And we say the more things change, the more they stay the same. But explain to us the nuances that basically if you have a static playbook with all of those things and you try to bring them into an AI-centric company, you're dead. Yeah. It's almost what you're selling. It's the speed of adaptability.
Speaker 1That's right. That's right. Those are constant. Those five pillars, they persist no matter where you are. But the customer and how the customer buys is different. And you're coupled with a machine learning solution architect. And so you have to understand how you're coupled with that person. And I don't know if that's terribly different. You know, partnering with a solution engineer across a different piece of technology. But the way that these models come out every week. And the way that companies are formulating compute strategies and the way that they're formulating how they're going to leverage new models and what they build on top of them. The funding that is going around to these AI native companies that are disrupting traditional enterprises. You know, you have to think about, like, am I going to need, you know, the software capacity model where I've got to have this many bodies in the Northeast, this many bodies in the West, this many in North Central, et cetera, et cetera, and EMEA and APAC. You know, where really is my ICP granted all the dynamic, constant changing in this AI landscape? My inclination is just to go. Go attack the enterprise market with this crew and attack the Fortune 500 and the Forbes 2000. But what's really interesting are the extremely well-capitalized AI native companies that have just emerged in the last 6, 12, 18 months, right, have really come out and disrupted the whole scene. And so we have to keep our ear to the ground on the innovation velocity around different companies that are very well-capitalized and determined. And, you know, where to invest our time in them versus a Fortune 500 that's scared for their life but maybe moves a little bit slower.
Speaker 5Let's get more practical. So, I mean, whether you're working, you know, at your company, which is, like, on the leading edge of AI and things are changing every week and people are getting funded different ways, all of that stuff AI can help with, whether you're in something moving as fast as the company you're in or some traditional software companies or some companies that are just now starting to move. Right. Right. Right. Right. It's really important to accept, you know, what AI is doing to the sales. Because at the end of the day, today, still only 25% or 30% of the time people are actually selling face-to-face or selling to the customer. The other 75% to 80% of the time, people are not selling. They're in nonproductive activities. They're doing admin. They're updating the tech stack. They're researching customers. They're researching use cases. They're doing homework on what's the latest. They're doing the greatest news of the customer. They're coordinating calendars. I'm coordinating information with my internal and external stakeholders. I'm spending 80% of my time doing that crap that AI can help me at least cut in half, maybe even more, to help me spend not 25% or 30% of my time selling, but spend 50% to 75% of my time selling. That's where I think the leverage comes for anybody that's in sales and is trying. I'm trying to understand how is this going to affect me? I think it's going to affect you in a very positive way and give you a lot more time in front of the customer if you use the tools that really help you with these nonproductive activities.
Speaker 1100%. The RevOps team that I thought that I was going to need to build out, McMahon, 14 months ago versus what I actually need right now by using. Tools or training my own agents like Jimmy did this past weekend. You're right. And I think that's the evolution of sellers. As long as the agents that they train are sponsored and approved by their company and able to connect to the system of record and the artifacts that the agents would sit on top of, they can get a significant amount of time back now through great technology and specifically through agentic AI. That's a big point. That's something that I highly encourage folks to lean into. I don't have nearly, not to spook anybody, and I don't have nearly the software stack or the RevOps staff that I traditionally would have because we've found automation and agentic workflows that are really helpful.
Speaker 4So for the listener here that's listening to you, you said the kind of EMC background, the kind of snowfall. I want you to go a little bit deeper. A couple of days ago, you and I were talking about this, and you were talking about the intelligence and the common one that we talk about a lot is curiosity. And you started to talk about multiple levels and layers of intelligence when you think about you really need to have as a seller today in this type of environment. Could you expand on that a little bit?
Speaker 1I think that there could be an argument. Traditionally, in this industry, and even with some of the folks that have been here for a while, I don't really need the AE. Just give me a really sharp technical solution architect, and we'll go close the deals with product leaders, or we'll just do more founder-led sales. With the profile of AEs that I hired, I knew that no matter what, this is universal, that we're always going to have to translate. The technical capabilities of the technology into value, for sure, and that we were going to need to be really strong value translators. We're going to need to be able to whiteboard the tech, for sure, but the business win is going to be extremely important, even after the technical team wins the math. I have a fun, poor description of how great the technical team is. Our essays are brilliant, and they're extremely important, but we really have got to be able, in such a competitive world, to understand the buying committee, which is going to be extremely cross-functional. You're going to have infrastructure owners, you're going to have product leaders, you're going to have engineering, you're going to have technical experts all over the place, and they can all win the logic battles, but the AE has got to be an extreme orchestrator that listens. It listens and understands where the value is and understands how to command the premium of the world's fastest tokens. The whole multiple intelligence thing, that helps me get over my imposter syndrome in a highly technical sale. That's all Howard Gardner's framework. People can look that up, that human potential isn't just one IQ score. The AEs have got to go out and attack the market and prioritize their territory and beget opportunities, qualify the opportunities. It's like if you have an AE and an SA in the room, and let's pretend it's just SA's. Let's say that there's a founder-led theory that I don't need AEs, this is so technical, sizing workloads and all that stuff is so technical. Just give me an SA. We'll figure out the rest. Customer asks a question to the SE or the SA. They ask, hey, what models do you serve? What models do you serve? How's the SE or the SA going to answer the question? Black and white. Yeah, I'm just going to tell you the models that I serve. You ask the question, logically, I'm going to give you the answer. An AE is going to have the. The interpersonal intelligence and curiosity say, that sounds like that's an important topic to you, Mr. and Mrs. Customer. It sounds important to you. Do you mind? I'm going to answer your question. I don't mean to play games. Can you help me understand why did you ask? Why are you asking which models we serve? Okay, well, because I can't tell you six months from now what model is going to come out from where, from Anthropic or from OpenAI or from Grok or from. So, I don't have a permanent model strategy when these things are leapfrogging. So, I need to know, what do you serve now? And I imagine, Mr. and Mrs. Customer, it's probably important what we're going to serve in the future, like what our model roadmap is. Those kinds of things aren't most commonly natural, in my experience, with SEs. And pure sellers are kind of born and raised, or at least the ones I like to hire. Really qualify these things and go down and listen for that opportunity, listen for that intent. I think that's something I see quite a bit.
Speaker 5In that sense. Qualifiers and they're skeptics. Yes. The reason when the pure maniacal qualifiers and genuinely skeptical of every question a customer asks, they want to qualify it. Why do you ask it? How come? Is that a threat to me? Should I answer the question? Should I answer it partially? Should I answer the whole thing? I don't know. That's what's running through your mind.
Speaker 1Yeah, that's right. And we're taking on an 800-pound gorilla, right? Like, this is high-stakes stuff. We're taking on a massive incumbent. And we've got to have that awareness. And we've got to have some really good judgment and making sure that we understand pain and we understand the implications behind that. We're really good value translators.
Speaker 5Okay, so what tools have you put in that are AI? Like, you've bought off the shelf and you've put into your sales to help your sales force, the sales process.
Speaker 4These are his partners that he's selling to.
Speaker 1Go ahead. No, no, no, no. There's no circular stuff going on here. John, it's really limited. It's really limited. Like, I don't have the stack that I used to have everywhere else. Now, I don't think that'll stay this way, but the reason is –
Speaker 5No, it's definitely not going to stay that way.
Speaker 1Right. I mean, I think there's some tools that I want to take on, but, you know, like the example I gave you, we are training agents, the ones that have access connected to our systems, to pull from our systems of records and do some ongoing research and analysis and surface patterns and risks to us. So I have a traditional CRM, and I have content management. I have some automation in our outreach, but everything else, I'm using agents that we're training to pull the information and do tasks for us.
Speaker 4I wanted to just follow up on this one. We'll come back to the stack in a second because I think that's very interesting and how we all agreed that for now and maybe what you think that's going to look like, but I wanted to make a point on this imposter syndrome because, I've been there before in my life, and I try to put myself in positions as much as I can to feel like an imposter. It means I'm growing and I'm learning, and, you know, if you're not growing, you're dying, so to speak. And, however, my experience is when people get into those types of environments, they try to learn as much as they can about the technology, and I've always felt like companies have the responsibility to really do their job. Right. To really help you with the knowledge of what that is, how it works, why it works, why is it a problem, how do you do it differently than anybody else. Companies owe you that. What you should bring to the table if you're listening to this is this intellectual curiosity, which is when you have an imposter syndrome, you have a tendency to try to be more interesting on a sales call or with people in a room versus, being interested. And somebody said that to us, Jonny, I think it was our selling DeVito friend, you know, the best sellers on the planet are the ones who focus on being interested, focusing on being interested in what the customer is saying versus trying to be interesting to the customer. And I just want to kind of highlight that I find myself in that scenario where I'm trying to learn as much as I can about the new world. That we're all in. And still going back to the basics of what problems does it solve. How specifically do you solve it? Like, how do you do it differently or better than anybody else? And where have you done it before? And allow people to go on a journey with you. I don't. Have you met anybody, Alex, that's expecting you to say, you should understand. Like, when you went to Stockholm, did anybody say to you, you should understand my environment. I don't. I don't think so. I don't think so. I don't think so. I don't think so. I don't think so. Right. So, I'm just expecting you to say, you should understand. Like when you went to Stockholm, did anybody say to you, you should understand my environment. You should understand exactly what we're doing. Or did they expect you to be interested in what they were doing? And then from that, from that interest, have a hypothesis about what you might be able to do for them. Does that make sense?
Speaker 1It does. I mean, the really cool part about the folks that I met with in both London and Stockholm, and I think even here is like, I get down. My strength is, I think, in the business acumen and understanding that if I, for example, provide the only engine that's fast enough to allow agents to self-reflect and collaborate in milliseconds, then I'm enabling products that were physically impossible two years ago or a year ago. And so, you know, when you think about like latent, so I get to ask a really simple questions, but when you think about latency. In a product, right? Imagine you're on with the pharmacy and they've built a new conversational bot or app. In fact, there is one right now that is extremely frustrating to me. It's like a two second SLA. I call into the pharmacy. I've low thyroid by genetics. So I got to take a thyroid supplement, violating HIPAA on the ship. And you call in and you call in to the pharmacy and they're the whole goal of these big pharmacy chains. It's like reduce the number of people like me calling, trying to talk to the pharmacist for basic things like a refill, right? So they built this conversational bot and it's on a two second SLA. I know I can tell, and it's just, it's way too slow. For conversational, you need to be like 500 milliseconds or less, right? But anyways, we empower instant AI, especially for conversational use cases for agentic coding, for deep search. And so I get to talk to people in Stockholm. In London, for example, about not just how much greater that is for the user experience, but what do you do to pack in value and revenue opportunity with that latency savings? Like how much more sophisticated did your product get? I ask them just point blank. If I bring it down where you're responding to human in 200 milliseconds and your SLA is 500 milliseconds, imagine the tool calling, the sophistication you can pack in. Between 200 and 500 milliseconds with me, can you help me understand where that might have an impact on your customer satisfaction, your revenue, your, you know, on and on and
Speaker 4on. Those two worlds, what I love about what you're saying here, and I just want everybody to hear, you know, go replay this story from Alex. There's a mindset that would say, I can do that faster and talk about all the things that you can do to do that faster. And you will be talking to people that are responsible for making that faster. You also need to be talking to people that talk about what they're going to do as business outcomes for that speed and those that are as old as dirt, that's Egypt old. It's like technical requirements create business outcomes. And it's just a little heads up because I think when I've talked to people that go over to AI native companies, they spend so much of their time trying to not feel, you know, and I think what, like you said, you just put the right technical people in the room and you are. The value orchestrator you're like, okay, I'm the, and I, I say this a lot. Sometimes I'm like, okay, let me just ask the, you know, the dumb question in the room or whatever. So what? So you can do it 500 milliseconds or whatever. What does that do? Who cares about that in the organization and depending on who you're talking to, you'll find that they might go up a technical chain and they'll talk about the technical people and SLAs and that, but there's a line of business that is, it's as old as dirt. It's kind of, I'm not saying you don't have, you have to be on your game, but don't get overwhelmed. That's right. Yeah. Bring your, bring your a game.
Speaker 1One of the companies, uh, McMahon and cap, you know, they have a flat subscription rate for their users, for their service. It's just, you subscribe, but you know, again, I get to be kind of dumb. And so, well, if I make the responsiveness that much faster. Do you think there's a market for premium users and you would charge a premium? Could you open up another tier, right? What's happening all over the place, right? Anthropic just did Opus fast, for example, uh, open AI as Codex and then Codex spark, which is powered by us and, and the guy was like, well, we never, we never thought we would get into premium tiers, but we, you know, we're just kind of one size fits all and I'm like, well, it's possible, right? He's like, absolutely. And I imagine it would pay for what you and I are talking about in droves. And he was like, yeah, you know, you can start seeing me, you know, selling a little bit. So I had to hold that back, but yeah, it's just where, where's the value of what's the art of the possible stuff like that.
Speaker 4Love that. So again, the more things change, the more they stay the same, it's like that didn't change when you were selling for the big behemoth software only companies. That's exactly what you were doing there. When you're going to AI native companies and now at the infrastructure layer in a highly technical environment, that doesn't change. You still have to connect those two worlds. So tell us a little bit about the, you know, the playbook side of it, as it relates to, you know, you're sitting in the seat, you're a CRO, you're sitting in this seat, you know, how to build organizations. What are some nuances that. Also have monster customers who are also partners. And so you're feeding, you, you can't, I can't imagine that you could build an organization with Salesforce. It just goes and calls on those monsters in that partnership relationship. What are some of the big nuances that you had to upskill yourself on, on how to attack that problem of going and building a go-to-market in this? I don't want to call it wild west cause it's not wild west, but it's pretty crazy.
Speaker 1Yeah, it is. I mean, I think, um, again, like the management of who our ideal customer is, our ICP is, is something that we have to like constantly pay close attention to. to and i i always um it's not a once a year type
Speaker 4thing hey this year it's this yeah
Speaker 1no no no no i mean we we revised this um not not in in months but in weeks like we we just stay really close to this because um you know we've we have to be uh very focused with our time and our energy of course i don't know if that's terribly different for others but like we've got to make we have these monster deals for sure um but i'm putting a ton of wood behind the arrow and just generating very healthy pipeline constantly but um there's a lot of companies that have emerged and that have been well funded and and it's hard to tell like who's going to survive what right like and so in some ways we're having to place bets with who we interact with right like genuinely who's going to go big with us who's going to have durable growing consumption of tokens um with us who understands our value and who's who's going to do great things and create new categories and new economies um and we go and and and build pipeline behind that it's it's different than me traditionally just going up market and attacking the fortune 500 and the forbes 2000 it's very different i'm watching the emergence of these ai natives and attacking the ones that we think are going to do extremely well for the long haul i've never had to do that if that
Speaker 4makes sense yeah are you using any ai or any of your agents that you're building a lot of people are talking about in this world the icp is changing under our feet sometimes are you using anything could you give any heads up to the to the audience on you and your smart mr lee your smart rev ops people yeah are you doing anything to stay on top of that more than you have
Speaker 1done before um we we are like we are we're tracking the flow of capital we're tracking um raises we're taking data from multiple sources and um and synthesizing it and um and constantly updating our top target lists for sure um and and so um tell
Speaker 4me how that affects the seller workflow like so you got the ability to do that like what changes for the seller
Speaker 1um it's the the tiering i guess the prioritization of the accounts that they attack and put into their their pg cadence changes right they will introduce new companies into their tier one which commands the majority of their time and attention um based on market signals um you have really interesting companies that are attacking health care problems that are disrupting what lawyers do what accountants do and so um we we have data sources and signals that flow in and um we have scoring of course and then of course that'll change the way that they prioritize their time and who they go after to try to create pipeline i don't mean to be too vague i'm a little hesitant to give you like exact vendor names maybe i can do like a follow-on blog or something to tell you exactly what's in my stack but i would just tell you like it is um it is a shorter stack than i've ever had with less less humans behind it than i've ever had and
Speaker 4you're not buying technology just to buy it or to have relationships you're you know what what i people are going at use cases and they're going at seller workflow so use cases for the technology that they're interested in and does it matter in this example does it match immediately to a high value uh sales workflow that makes sellers more productive i think if you start there the advice for everybody out there the closer you get from the farther you get from inspection to adding value in the engine of pipeline generation and qualification and making it easier for those sellers with these highly uh intellect intellectual curiosity and the things that we were talking about the human factor that's going to be very very hard to replicate getting it in the hands of those people uh seamlessly and you know less painfully is really the name of the game i think so i'm looking right to the blog
Speaker 1yeah i will uh i'll put i'll put some thoughts on paper and and and reveal some of the secret sauce but i i would tell you that like i gotta i'll just give a tip of the hat to the leaders that um took it took a bet on me to come over here that were in really great companies really great high growth environments uh griff paul jason mark suve um jimmy the crew uh others like they they went out and you know like i learned from you all like you hire great a plus leaders they're going to go out and hire great people um you know they're obsessed with our work and um they are um they were able to to recruit people before we had a really large inference business or as it was emerging they made the bet with me and i'm extremely grateful for that
Speaker 4and for that amazing amazing hey alex before we go advice to sales leaders and advice to sellers thinking about everything we just talked about and then thinking about alex two or three years ago okay what is the best advice you could give to sales leaders listening to this and sellers so i
Speaker 1would go just april april first april fool's day um roloff brotha if i'm pronouncing his last name sequoia had a conversation with jack dorsey and jack dorsey wrote a paper about i think roughly it's like hyper intelligence over hierarchy and um his vision for the this intelligence layer that's going to traverse the entire company and that there are three layers of humans that are interacting with that intelligence layer there's the ic the dri and and then the coach conductor orchestrator um having an understanding of where potentially the world is going with agentic ai and having an extreme central brain that traverses the entire company and how you would interact with that i'd go dig into that thing and listen to that i would also prompt your favorite model to build a learning path for yourself so that you understand um ai training inference um agi you understand prompt engineering you understand the concept of tokens and you understand maybe how tokens might um uh be assigned there might be a token budget per employee in the future right there kind of is one now but it's it's good that will be a thing there's a labor shift not just a software shift i would just get really in tune with that and understand it and like have a self-learning enablement path and um you can shine with that stuff and the things that are in john's book and the things that are in this podcast are universal truths that will travel no matter and apply no matter where the tech takes us and i tried to hit on a few of them and so um you know the the the level of human talent the level of of of coaching your people and developing your people and being maniacal about execution um being extremely curious and driven like those things are still going to apply and i think coupled with a really great learning path um you know you're you're future-proofing
Speaker 4yourself uh potentially in a really good way i love that i what i've been saying alex is and i've been saying it for decades but or a couple of decades at force management for me now more than ever attach yourself to the biggest business issue facing your customers and employees and i think that's a really good way to do that and i think that's a really good way to do that influence their decision criteria with your differentiation now those are a mouthful yeah those are absolutely a mouthful but if i'm a leader or i'm a seller and i'm looking at those two skill sets those for me are are at the top of the list right now thousand percent agree
Speaker 1sincerely even in this new world that i'm in that is a hundred percent so true and that will remain true that will persist amazing
Speaker 4i feel like i gotta go read something so i feel like i gotta go read some poetry johnny what are you reading for poetry johnny what are you reading right
Speaker 5now me
Speaker 4yeah come on exposure
Speaker 5leonardo da vinci right now the biography yeah by walter isaacson he's oh that's a big one jobs he wrote about elon musk he wrote um einstein i love his writing yeah
Speaker 1i i read two i don't know if this is gonna make the cut for pro-spot but over christmas break uh christmas break i'm gonna read a little bit of christmas break kormac mccarthy no country for old men and the road um and those those will blow your hair back those two are really cool too for
Speaker 4for fiction well done mine was my two are uh well my one for sure and i keep going back to reread it is the the founders about the guys that started paypal it's fascinating to me i just found so many and what they did after paypal and how they built companies and and uh the culture associated with that and how they messaged and created boards and i just found that a great read and then a personal one that i have is i'm i'm on a i'm on a big uh just coming off the holiday here i'm on a big um investigative path on the shroud of turin and we'll have to do another podcast on that one but the when you talk about technology and technology advancements are actually blowing my mind on what they're revealing on that so coming to a theater near you very cool well done buddy so great to see you so happy for you you've earned it and uh can't wait to have you on again when you're talking about you know foreign full scale what this thing has turned
Speaker 1into thank you congrats john i mean man it's great to see you yeah i i appreciate you very much john cap and i appreciate you very much and um i wish you the best uh thank you so much you got it
Speaker 5brother and thanks to everyone for listening to another episode of the revenue builders podcast
Speaker 2thanks for listening to today's episode if you enjoy the content please subscribe rate and review the show to help us reach more people this show is brought to you by force management where we help companies improve sales performance executing the growth strategy at the point of sale check out forcemanagement.com for more information

Podcast Summary

Key Points:

  1. Alex Verrill, EVP of Worldwide Sales at Cerebras Systems, describes moving from traditional software sales at MongoDB and Zscaler to the AI infrastructure market.
  2. Cerebras built the world's largest chip, a full silicon wafer that co-locates memory and compute to eliminate the memory bandwidth bottleneck in AI inference.
  3. The AI market has shifted from training models to inference, where latency and memory bandwidth determine the quality of user experiences and agentic AI applications.
  4. The five universal sales pillars—pipeline generation, rigorous qualification, messaging mastery, dynamic ICP management, and accurate forecasting—persist regardless of technology.
  5. Traditional software leaders who transfer static playbooks to AI-native companies without adapting get exposed, so speed of adaptability becomes the core differentiator.
  6. Cerebras revises its ICP every few weeks, tracking capital flows and funding signals to prioritize well-capitalized AI-native startups alongside Fortune 500 accounts.
  7. Sellers must become "technical athletes" and value translators who connect technical speed to business outcomes like premium tiers, customer satisfaction, and revenue.
  8. Agentic AI can cut non-selling administrative work dramatically, and sellers who train and supervise their own agents will gain significant productivity advantages.

Summary:

In this Revenue Builders replay, hosts John McMahon and John Kaplan speak with Alex Verrill, EVP of Worldwide Sales at Cerebras Systems, about leading sales in the AI-native era after scaling teams at MongoDB and Zscaler. Verrill explains that Cerebras built the world's largest chip—an entire silicon wafer—which co-locates memory and compute to eliminate the memory bandwidth bottleneck that limits AI inference. As the market shifts from training models to using them, latency becomes the decisive factor, much like the broadband revolution, and there is no market for slow search or slow agentic AI.

Verrill admits to initial imposter syndrome but argues that five universal sales pillars—pipeline generation, courageous qualification, messaging mastery, dynamic ICP management, and accurate forecasting—apply everywhere; what changes is the speed of adaptability. Cerebras revises its ICP every few weeks, tracking funding flows to bet on well-capitalized AI-native companies while still pursuing enterprise accounts. Sellers must become value translators who connect technical speed to business outcomes, and those who train their own AI agents to automate administrative work will reclaim selling time.

Verrill advises leaders and sellers to build personal learning paths around AI, stay curious, and attach themselves to their customers' biggest business issues, since human skills and execution fundamentals remain future-proof.

FAQs

The episode discusses how traditional software sales playbooks must adapt in AI-native markets, emphasizing the need for speed, technical fluency, and human curiosity.

Attack the market and generate pipeline, courageously qualify the pipeline, master messaging, manage and prioritize the territory, and deliver an accurate forecast.

He saw that AI inference is bottlenecked by memory bandwidth and that Cerebras's wafer-scale chip solves this, making it the key differentiator in the AI era.

The chip is significantly more power-efficient because it eliminates the need for thousands of stitched-together chips, and it uses liquid cooling, which is more efficient than air cooling.

Sellers who try to transfer a static playbook without adapting to the dynamic AI landscape and without hands-on experience with AI tools are most at risk.

Sellers should train agents to handle non-selling tasks like research, admin, and coordination, freeing up more time for face-to-face selling and value-based conversations.

Chat with AI

Loading...

Pro features

Go deeper with this episode

Unlock creator-grade tools that turn any transcript into show notes and subtitle files.