Speaker 1I appreciate clarity of thought and communication over everything else. Someone we hired in the last couple weeks said, well, I'm really concerned about companies that are growing super quickly and whether or not they're going to hire too many account executives. And I don't yet feel confident that I know from perplexity whether or not you have a clear idea of how you won't overhire and how we'll all be taken care of. You can give me two or three of those questions. It's like, wow, this person is organized, has clarity of thought, is willing to communicate with me and be open. And it just implies we're going to have a fantastic working relationship.
Speaker 2Just wrapped up a great episode with Stephen Boone, VP of Enterprise Sales at Perplexity. In the conversation, we go through his non-traditional background to a revenue executive seat. I think it's probably a healthier mindset to think about
Speaker 1what you don't have with your background rather than it is like, what's the unique advantage that I have from it? Nobody likes the person, by the way, who walks in and says, well, I used to be an ex, so here's how you do it. Like a couple of years does not make you an expert, but I have some empathy for what they go through and what sitting on the other side of the table looks like.
Speaker 2We go through the different product mixes that Perplexity has right now, both on their application and their API, and how they think of consumer versus enterprise. People who were getting a lot of value out of Perplexity
Speaker 1were using it for work in some way. And so the idea became over time that if people are using this for work, therefore enterprise is a thing. Enterprise to us means B2B. Somebody who is purchasing this to get work done.
Speaker 2How do you think about capacity planning in a company that's like one of the fastest growing in AI right now?
Speaker 1Maybe this is another thing we've tricked ourselves into a little bit.
Speaker 2Welcome to the Crew podcast, where we have conversations with the top go-to-market leaders at AI natives and SaaS startups. I'm Chris Belestris, co-founder of Crew, a go-to-market recruiting, media, and investing firm helping seed through Series D AI natives grow. Now, on to the episode. Steven Boone, VP of Enterprise at Perplexity. Recently joined a couple months back. We have some Perplexity folks that listen to the podcast. So I was pumped when we got intro together. You looked into law. My wife's a lawyer. So looked into law school. Now I've flipped you into leading up enterprise sales. One of the hottest AI natives right now in Perplexity. Been a user for two years. I've written articles on how I've used it in the past. So this will be interesting to dive in more of how you guys are building out the enterprise practice and just Perplexity generally. But we had on Nick Bogaty, CRO at Vercel. Vercel, yeah. His first sales job. His first time was the CRO at AppDynamics, which that was the first time we had on somebody and that didn't have a traditional sales background. Stevie Case has said, I know you know from Twilio, has said there's going to be more and more of these like enterprise leaders that don't have traditional like, hey, I was in AE for 10 years. You have that with marketing, RevOps, you were chief of staff at Twilio to a CRO and then said yes to sales. What do you think in your background that you have that differentiates you heading up an enterprise sales team at Perplexity right now?
Speaker 1Sure. Well, both of those sales leaders are fantastic, either by knowing them or through reputation. Sure, I spent some time in a whole bunch of different roles, but candidly, I think that more than anything, those gave me empathy for other people's jobs. Nobody likes the person, by the way, who walks in and said, well, I used to be an ex. So here's how you do it. Like a couple of years does not make you an expert. And everybody I work with in marketing and everybody I work with in RevOps, for the most part, has significantly more experience than I do in those things. But I have some empathy for what they go through and what sitting on the other side of the table looks like. And so for me, I spend a lot more time worrying about what did I not get by the fact that I didn't carry a bag. For me, a lot of that was actually around trusting rep instinct of the most fantastic reps and trying to figure out what that means to them. And there's a lot of times that data will tell you one thing from my RevOps background. Yeah. And the best reps I've ever met are like, hey, man, you just got to trust me on this one. And so building up instinct was probably something that was really important. But yes, I think it has led to leading some pretty data-driven and oriented teams and one where we value, certainly in the organizations that I've led, relationships with our cross-functional partners more than anything. And I actually think a lot of companies have gone too far in this weird spirit of everybody's in service of sales. Like, I think that's great. Everybody in the company is in service of our customers, but that doesn't always mean that sales is at the center of some wheel and everything else is at the outskirts. Like, everybody is figuring out how to get it done.
Speaker 2Yeah, and you're wearing a lot of different hats, VP of Enterprise. You just said this right before. Sales, success, FDEs. You said a plus of having a unique background with doing a lot of things is this empathy, working with a lot of different teams. What are some of the teams that the direct sales team works with the most at Perplexity internally?
Speaker 1Sure. My organization also has support, both consumer and enterprise at Perplexity. As you can imagine, we have many millions of users. That support team is now a pretty fair size. We work a lot with support in the sales organization. You wouldn't believe how many great opportunities start from somebody hitting the limit of what they can do on one of our plans and filing a ticket. And it is great that an enterprise support rep will dig in and figure out how to, how to unblock somebody. But it is very often that the right next step is actually for them to sit down with a quote, salesperson after that point, especially if they're a very heavy self-service user, because the answer isn't, hey, here's what you should do in the product to unblock yourself and then you're done the way you might interface with a support agent. But maybe it's, hey, look, it sounds like you're about ready to do something really complex in the product and you should talk to sales. And so, yes, of course, we work with all of the teams that you'd think about, but I think a really tight loop between things that oftentimes don't seem like they're connected can create some really interesting synergies and flywheels inside of a sales organization.
Speaker 2You said you had millions of users. There's probably a lot of these support tickets. So there's a lot of people that are asking this, but then how do you bubble up which ones are most important? I'm assuming using AI, maybe this is like a good natural feed into using perplexity. How do you actually take all the support tickets and know which ones are worth the salespeople's time or not?
Speaker 1Honestly, a lot of times it's actually still a manual flag. Okay. That you measure our support organization. Sure, we use perplexity a lot in a couple of different ways. One, we have a great support vendor. Perplexity is integrated into their product. Okay, that's great. Two, support agents, especially for complex tasks, are going to ask our product perplexity computer first. Hey, what's going on? Is this a product bug? Is it not? Is it a user issue? How might I resolve this? Before it even gets to that point, tons of tickets have been deflected to AI and we're always trying to figure out how that can be more, but only if it's a good experience but I would say that things that require salespeople are probably the same reason for things to require support agents and that those roles I don't believe are going to go away anytime soon because they're things that require judgment and taste. Yeah. Somebody wants a refund. They have feedback about the product. Somebody feels like whatever resolution we gave them was not the right one. It's time for a person to step in. Right. And oftentimes when I see the Slack channel that has, hey, I've tried to reach the end of resolving this issue or I found some really interesting use case that someone has, our sales reps are actually like farming that channel for their own accounts or for other accounts and because they have seen how wonderful the types of opportunities that can come from there or just the relationships they get to build from customers, they're often the first person to raise their hand and say, hey, can I help out with this?
Speaker 2Yeah. All makes sense. You dropped computer in there. So it made me think of this next question of what even is the perplexity landscape and where do salespeople spend their time? So before, I'd love for you to answer that a little bit of background. I've been using it probably for two years now. Thank you. It started with, hey, Claude, I was using ChatGPT at the time and Claude, like it goes back only to 2023 and it was 2024 that time. I was like, all right, well, if I want to see if the Jets won or lost last night, like there was no tool except for perplexity. So I would use that as the real-time search, whether it was the app and I was using that as my chat bot for more real-time stuff or the API when I was trying to like build apps and replit. To be like, okay, to write stuff, I'll use OpenAI and Anthropix APIs and then I'll use perplexities if it was like, I need something more real-time. You have the computer product now. So what is the product landscape at perplexity and where do salespeople spend most of their time selling?
Speaker 1Sure. You started using perplexity really the same way I did before I was at perplexity. I've only been there about five months now. It replaced Google search for me a long time ago. I liked the concept of real-time cited answers in a conversational format rather than ads, blue links, et cetera. Yeah. That was a very good start to a business because I think we've carried the same themes all the way through. Perplexity wants to push the frontier of AI by providing accurate products and to deliver orchestration across the world's best AI technology. So like, what does that mean? You interface with this search product that's now an API, that's the ability to go ensure that you have real information that is accurate and current across the web rather than like whatever an LLM has been trained on and then stopped getting information for after. That use case you just described of back in the day, you couldn't get real-time information from other LLMs or frontier providers. Most harnesses now have some sort of call, like an agent can go look at search results on Bing or something. And then there is now a whole space of companies who provide really good search APIs. I still think we have the best. what that turned into. too, especially when kind of this agentic era started to come about. And it seems forever ago, you and I talked about this, but like models got good enough that they could do real work on our behalf. Maybe nine months ago, from my perspective, we built a product called Computer. We built it actually internally. It was the idea that we have this ability to go get accurate information off the web. How do we build a really great harness across all of these new models? Because we had a belief that there wasn't any one model that was just the best. They all like to say they're leapfrogging each other, but whether it's open source models or models from the frontier labs, they're actually good at different things. And they've seemingly become more specialized over time. You may have the same feeling with, if you're using these singular models, you may go, I kind of like to use GPT for this, or, oh, I go to Gemini for that. I like the way it writes better. Who knows? But all of those models working in concert is this idea of organization. computer is, we like to say internally, everything is computer. It's kind of become the main interface that people work with our desktop web and harness application. So computer is a solution that breaks down complex tasks into its constituent parts and then ferries those off to the right models, frontier or not, and comes back and just gives you great results. Yeah. It has a series of tool calls in it. Our own search API is a tool call. Interfacing with connectors for all of your enterprise data are other tool calls. And computer is the main way that our reps interface with it.
Speaker 2I'm excited to talk more about how you guys are using it internally. We'll get to that in a little bit of like how you see the best reps using AI. One call out, perplexity was ripping from a revenue perspective for so long. Yeah. I kept being like, why have they not hired enterprise leader just yet? What did the market tell you guys that made your hiring happen when it happened and maybe not six months prior? I'm curious of like this move to enterprise now and like how you guys are thinking through that a little bit more.
Speaker 1I believe that there was not any one moment where you just said now is right. But I think there was a continued analysis of how people use perplexity. And sure, there's free plans. And then there were, you know, premium consumer plans. And there's always been an enterprise product. And we can talk about the delineation between the consumer product and the enterprise product. But the idea was that regardless of any of those plans that people were on, people who were spending a lot of money, or rather, getting a lot of value out of perplexity, were using it for work in some way that meant if I was on a consumer plan, and then upgraded, I was doing research for my legal job or my tech job or my consulting job. Yeah. If I was connecting it to internal databases, or like CRM or something, I was doing that for work. And so the idea became over time that if people are using this for work, therefore, enterprise is a thing. Enterprise to us means B2B, for the most part, it means somebody who is purchasing this to get work done. And I don't think what we were looking for was, hey, let's go scale enterprise with a capital E, we're just going to go hire a ton of enterprise sales reps and sell to the largest companies in the world. It was, there should probably be a different experience and one that is potentially for many of our customers, human driven. And therefore, we should build an enterprise team around that.
Speaker 2Yeah, makes sense. I think that this is why Anthropic has kind of ripped in the last year is that a lot of companies started with this consumer use case and really have been getting pulled into the enterprise more and more. The way that you define enterprise to is that mostly the API product that is being sold? Or what is the application use case in the
Speaker 1B2B enterprise for B2B right now? than the largest enterprises, is it means that the use cases can span a very wide variety of how people are using it. I like to think that, you know, Our enterprise product is, yes, consumed by solopreneurs and even SMBs, yes. They are often like building their entire business on top of Perplexity, whereas enterprises are often connecting computer to data warehouses and CRMs and a whole bunch of technology set that they've already purchased and built their business around. But look, there's website building businesses on top of us. There's people who've built entire content marketing engines that are highly verticalized for some specific use case. And those people can spend a lot of money on perplexity and probably want to talk to a human, but they may never grow past one or two employees because perplexity allows them to run a really lean business.
Speaker 2Yeah, I'm excited to hear more about like that use case you said of the two people spending like age or K on computer and like later in the AI and GTM, like what actually that means. There's a couple of things you said nine months ago, and this checks out to be the way that I see it, too. It's like Claude Cowork came out in January. we threw an Agents in GTM with Clay event in March because it was starting to become more common for the business user or the enterprise user or B2B, the knowledge worker to say, okay, I'm getting used to connecting all these tools. And it started in maybe Q1, Q2 of it just being a chat interface and you would just talk to it, but it was connected with a bunch of your tools and maybe starting to take a little bit of action on your behalf. It's interesting now to see what tools like maybe a Grokbot or computer, it sounds like it's doing, where we're getting more and more comfortable with them to actually like, hey, go do stuff for me and actually grab maybe my own instance. I'll log you in through like a separate browser or like this sort of thing. I really want to get into that in a little bit. I've like nerded out for the first 18 minutes of this on just like ProPlexi because I think it's really cool what you guys are building. I would be pissed at myself if we didn't jump back into sales topics because most of the people listening are BDRs, AEs, and sales leaders. You threw me off a little bit because you said, I saw PLG at Twilio and most of my friends that are in PLG, they never want to move off of it. They always want to stay in PLG businesses. You took the opposite approach by going and leading enterprise at like a company like Zora, where you were at before Perplexity. And you said, I want to learn top down and enterprise deals, which is like, that's great and all, but like, I feel like you don't get paid for playing hard mode is how I typically say it. What was the thought process around getting off the PLG train, going to Azora in order to learn a different skill set?
Speaker 1Yeah, I think it comes back to your, you know, can you go have a CRO type job without having carried a bag? The answer is at every, I'm looking for people when I hire to, by the way, and I really challenge myself for this. There has to be a new brick in the wall for every job that you're taking. Something unique that you are learning that will become invaluable and will make you a better, more well-rounded person. I like to see that in whatever role I'm hiring for. But for myself, it wasn't like I want to go do something that's hard. That doesn't I don't know if that makes sense. I think it's more how could you build a truly generational company if you haven't done all of the flavors? Because all of the biggest, best companies that I would want to be a part of actually have all of the flavors of go to market. I mean, especially if you have PLG, then you probably have an SMB business, then you probably have an enterprise business. And someday you're going to exhaust the amount of enterprises that are just going to sign up for your product without ever needing to actually go through a super formal evaluation cycle on the front end and get through security before they were ever to implement that product. And so you have to be able to do both. And, you know, the hard mode, do reps go to PLG companies or enterprise tops down companies and then never want to go back? I don't know if the binary makes sense. I mean, sure, it can be different reps, but if the quotas are the right quotas and the margins are right, then it totally makes sense to have both those types of businesses. It would be very weird if it was the same margin, the same level of stickiness, and product A just was harder to sell than product B. There are a lot of benefits to really tough to implement and to sell enterprise products, including that they are generally really, really sticky. There's a very weird inverse, by the way, of like the harder it is to implement something, generally the harder it is to rip out.
Speaker 2yeah, we're seeing this now with these companies going so fast. It's like there's always that risk of like if it's that quick to go up, you know, what's the way?
Speaker 1If it takes someone only a few days or minutes to make a decision that you're the right call. I think that that carries a little less weight sometimes than if it takes them six months to go through a buying cycle to make a
Speaker 2huge decision. You go to Zora and you have reps. You brought this up earlier. It was like, hey, one thing I needed to catch up on a learning curve perspective was trusting reps gut instinct. How did you then gain credibility with a sales team selling enterprise deals when you never held a bag, but you're learning on the fly, but you have to then make judgment calls as the leader of the revenue business unit? Like, how did you think through that of learning versus actually leading and. showing credibility yourself?
Speaker 1You have to start with places that you know you can add value that aren't outside of your wheelhouse. Everybody who's good at what we do can walk into an organization and has a set of skills that would allow them to have a pretty good conversation with other executives without any context of what the product even does, which is probably the case on day one, week one, whatever. And I've also never been in a company where I didn't say, I better get in front of some customers in the first week. And it's not that I'm going to walk in and be a product expert. And by the way, you should be vulnerable about that kind of stuff with your team. Here's what I don't know yet, and I'm definitely not going to speak up when X happens, but I can be empathetic to leaders. I want to talk to people with the following types of backgrounds. I'd love to be included in a couple of huge escalations related to product failure or support failure or something related to people because I'm already an expert in those things. You know, RevOps background going, to a company like Zora that sold monetization and you were seeing a ton of shift towards usage-based pricing, especially when I joined and still now. Okay, let me go talk to people who are thinking about implementing usage-based pricing and compensation plans because I just came from a company that did that. I can build trust with our customers on something that we are messaging without having to be an expert in the bells and whistles of our product.
Speaker 2That's a whole other rabbit hole to go down with you. We used to compete with Zora a little bit. I worked at Metro as an early sales hire. Oh, okay. It's a great company. I think it was Togai that you guys acquired there. And yeah, now the market's ripping. I see Scott Woody from Metro and posting about how their revenue is inflecting because everybody's doing usage-based pricing and needs the infrastructure accordingly. Big thanks to our sponsor, Attention. Sellers were hired to sell, not to fill in the CRM. Attention records and understands every sales call, then automatically updates your CRM, sends follow-ups, and scores reps against frameworks like Medic, Bant, or your own custom methodology. Next best actions are: proactively generated from there. You get a clean pipeline, coaching on every call, and forecasting you can actually trust without asking reps for more admin. Turn your call recordings into a GTM operating layer with Attention. Try them today at attention.com. You said a quote that was interesting, this concept around you either productize it or you white glove it. Tell us more what you meant by that and how should people think through what is something that would maybe stay more in the PLG motion and productized, and then what should be more of a white glove in the AI era?
Speaker 1There are bounds in any product. There's a bound where the product is going to be really, it's going to be really tough to do it self-service anymore because the workflow or the use case becomes really complicated. The other is a place where a human can just actually provide a thoughtful person to bounce ideas off of or share insights based upon how other people have done it. What I'm looking for is the current bound of the product where if I let this customer keep going, they might become a churn risk, or they might get really upset, or they might screw up something for themselves if they don't have some piece of the product that they're really interested in. If they don't have a piece of knowledge. And then what that might be an example from Twilio was we would have customers who pretend you're scaling a social app and you use Twilio to get one-time passcodes to log people into said application. Well, there's something called toll fraud. We don't have to talk a ton about, but it can run up bills to be really high when you go into countries where there's less mature telephony networks. Okay. Azerbaijan is one of those places. A couple of really smart people on my team at one point said, "Okay, once you see somebody start to rip usage in a country that we know that there's high toll fraud in, send them the following content, get on the phone, help them figure out this problem." Because this is the same problem they would have on any one of our competitors, but if we show up as an expert there, you've actually filled that gap of everybody wants to say that they're a consultative seller and most people just kind of fub that. But if you actually show up in someone's time of need with a white cape and solve a problem for them, that's great. At Perplexity, if you connect your data, warehouse to Perplexity, you're probably about to do some real work. Yeah. And I think we have great connectors and the product gets better all the time at how it's figuring out how data is structured in these external databases, whether it's a warehouse, whether it's CRM. But as of today, there's probably a couple of things we gain from doing that, from trying to engage with a customer at that point. One is, "Hey, let me tell you how we've built a semantic layer on top of our data warehouse that allows us to pull the right information canonicalized at any time and the benefits we've gotten from that." Customers go, "Wow, thanks." Yeah. The other is, "Let's kind of learn how they were starting to struggle or if they were struggling and then give all of that feedback to product and then figure out how the next time that thing happens, can we get one inch further down the field without needing to engage a human?" Maybe there are things, by the way, that you can never automate away, but one of the best ways to get that feedback, find friction points, engage humans at friction points, decide if it's worth engaging humans always at that friction point because it leads to sales, but also go fix the problem so you don't need to
Speaker 2You know what I love about that story? It's your unique background coming out of what you said earlier. You have the RevOps background, so you're able to take all this data that's sitting in usually a CRM or support tickets and all that. What's the implication of it? Helps empower the sales team to take real trusted advisor or consultative action, and then it's fascinating then bringing in your ability to empathize, but also just cross-collaborate with multiple teams and then this is all like a sales brain too. This is like, I would want to work at a company like this that's doing all that because it'll make me more effective seller. So I love a story like that. And Stephen, one thing I'm getting out of you is like very numbers driven. I'm a former accountant before I got into sales. And then I was in sales most of my career before I started doing crew. You're very numbers focused. And one of the things that kind of broke my brain was this chart. I love thinking in charts. You brought up this concept of complexity on one access and then gross margin adjusted ACVs. On another, tell us a little bit more of what you meant by this of this kind of is a flavor of PLG versus enterprise a little bit. And what we were just talking about of when to bring in a human and how to effectively structure, uh, automation versus human led approach. Tell us more what you meant by this chart.
Speaker 1What I kind of imagine is that there's, you know, everybody loves these two axes charts, but definitely two axes. Yeah. Uh, complexity of sale on one and then annual contract value, but probably more effectively margin adjusted annual contract value on the other. And the, the slope with a line of one, what that would imply is like, no matter what small deal equals not complex deal and complex deal equals big deal. And I just don't think that chart exists out in the wild, or at least every single company and every single person's chart is different because the whole goal of what we're doing is to break the line. And like, if you work at a company where you break the line the long way, or even maybe think about it as even having a culture of breaking the line the wrong way. Yeah. Again, nobody wants to work at a place where super complex deals and they're small and have low margin. That would be like ultimate hard mode. That, that sounds not fun. And we shouldn't pay people more for that, or the company's gonna go bust or something. But I think if you get stuck in that framework of thinking, instead of going, is there a different thing to sell or a better way to sell? Or where could we allocate time where we're differently on the line? People will often like try to break the line even more, still the wrong way, which tells you why you're there on the line. Like we're really struggling with long sales cycles and w getting, getting technical validation in stage, whatever in your process. Oh, well then the answer is that we should do long workshops on site and bring more solutions, architects or services in, or another person. And oh, lo and behold, you actually just made your sales cycle even more complex and even harder versus saying, is there something we can remove from this process? Okay. And the best companies that, uh, you know, breaking the line the other way, part of it could just be the nature of your business. You have like, how do you break the line? The good way PLG is actually one of the ways. Okay. If somebody has already touched and felt and loves your product and is getting value from it and is happy, but there's something gated behind an enterprise contract, that is other features or discounts or access to people and services that can help them. Sure. Maybe that's easy mode and you can do some pretty big deals pretty easily. Other ways would be, maybe you should actually be ungating certain features in your product. Maybe you should figure out ways to reduce friction. Maybe you should combine roles so that you don't need multiple people in the same meeting to get something done. Maybe you should kill that first discovery call that you have, because if you have a big drop off after that, you've added no value. And you took 45 minutes asking somebody questions that the next person's going to ask the same questions for it. And so my challenge to people is figure out through your product, through your sales process, and then even your individual process and book of business, how do you start removing things to try to figure out how you live more time above the line than you do below the line?
Speaker 2All right, y'all quick break. If you're getting value from this, do me a favor and hit the follow button on YouTube, Spotify, and Apple. It helps way more than you think. Now back to the episode. I love this concept of removing things that you'd brought up there. And you brought up the story of you know, the two person shop who was spending like 800k with you guys. We're seeing this a lot too with these metrics of like, you know, the one entrepreneur company that's doing the leverage point almost that starts to exist in this AI native era. But then it got me thinking like, I do want to pivot into like, AI and GTM, dogfooding, how you all are using your product internally, very AI pilled, using your own product to create leverage. How do you see the best teams, the best sellers specifically using AI in GTM?
Speaker 1The first is a specific thesis we have that we actually want to provide a lot of flexibility to our AEs and how they consume AI, even at a company like Perplexity, where everything starts in AI. And what I mean by that is, build the scaffolding and the building blocks so that people have access to the right tools and skills. We have one way that we want to run a pipeline cleanup system. skill in our organization. We have a very specific way that we want to measure when it's time to engage with self-service leads and think about, you know, moving them to enterprise contracts. But how that works its way into somebody's day and the way that they consume that information or the way they engage with it, I want to remain as flexible as possible versus there just being some, this is the GTM agent and UI that's now some new dashboard and your leads are here and then you click this button. Like, it's not the point of AI. If somebody wants to consume their meeting brief in the morning through a podcast or a video or text or in email or in Slack, I don't care. Go run that automation however you want. But the way that I find people use AI best in sales is stop doing the point solutions or the quote, the agent for this. And it has to be a full loop. What does that mean? Like, how do you follow the virtuous cycle of self-improvement in sales? We think of it as one process for a rep to go into a meeting. And before that meeting, they should have been prepped by AI. They were given the right amount of information, not a 10-page brief, but like the actual bullet points that are relevant. If they wanted to produce a demo for that customer, AI should do that automatically. By the way, with a product like computer, sometimes the best people to ask how to demo computer is computer. You know, produce me a demo environment that has 21 use cases for Chris that are relevant. Oh, and by the way, make sure these three use cases are relevant. And then you can do that automatically. Cases are included. I had a conversation with him that you can point to the recording here, but also make sure it looks like this and is on brand. Okay, that gets done. But then after the meeting, the follow-up has to be, okay, based on what we heard in the call recording, here's what stage this should be. And here's the recommended notes, but the AE overrides AI. Then I want the rep to get its call torn apart by the AI. Like, hey, by the way, well, it's still fresh. You did not set up a follow-up meeting. Shame on you. Other calls that are in the same stage or our disco calls have the following things happen. Here's how much you talked versus not. Okay, great. Now they're going to get a little bit better. And the suggested follow-ups happen. But you have to always be completing the whole cycle of self-improvement. Sure, we do that inside of computer. But the most important part for us, again, is the flexibility for the individual's specific cycle, but that all of that stuff is timely and that they don't have to go into 10 tools to do it. I think you had asked earlier or started to think about, like, what our tech stack is.
Speaker 2Definitely, yeah. I don't think
Speaker 1AI has completely changed it. Like, we still have a CRM, by the way. We still have tools that allow people to email people. We still have, we have Apollo. We also have Zoom Info. We have Salesforce. We have loops for, you know, broader emails for the life cycle of our millions of consumers. We also have the search API to go figure out information from the live web. But the thing that I care the most about is that when a rep wants to know contact information for someone, that they don't have to leave one pane of glass and go into three other ones or make a call. They just go, here's this list of people I want information about, and here's this outcome I want to produce. I want to email them or I want to tag them to be in my territory or whatever. And I want behind the scenes something to decide which one, which tool needs to be called to get that information. You need CRM, Snowflake, Momentum is our call recording software. You need Momentum Info. You need product usage, info, and you need email and contact information info. You should just be able to ask a system that information and it comes back and gives you an answer in whatever format you want. Versus the biggest thing we think about in our tech stack is how often I have to use the technology. I just don't want to have to think about the technology or ever log into it. And those are the best vendors. And it actually has impacted how we pick vendors. I don't care what your UI looks like. I hate being in UIs. So if it has a really great API and MCP connection, RIP's super fast. And you have a support model where I get to escalate to a human or if I need help really quickly and you'll just solve it, then let's go. So people who like have a Slack channel with us, they allow computer to be in that Slack channel with us and watch and learn and listen from issues. And I can pull good information, then I'll buy it and I'll never turn off of it. But it needs to be consumable in a way that I don't have to log in.
Speaker 2I love that. What is the pain that you see people using? Is it mostly Slack? Like what is the way in which an A or are they using, just the in-computer, it's computer like an actual app that that is the chat interface? Like what is the common interface that you see AEs using?
Speaker 1Yeah, computer. For many customers, you're going to be in the web app or the desktop app. Yeah. The same way that you probably are interfacing with a lot of other AI tools that you use. We have seen our usage shift more and more towards Slack as the most common one internally, partially because we have a culture of transparency across all of our teams. And it's actually the way that the computer was started. I don't think I went into this, but at the onset of computer, when we were figuring out what it was going to be, everybody in the company could only interface with it in one Slack channel. And so you would say at computer, I need help with X, Y, or Z, and then you'd create a thread. And we created the web app later. But so that has really carried through to where the vast majority of conversations are the computer bot is watching. And if you're working with someone and somebody says something that you don't understand, or you need to validate with data or some other source, you say, Hey, at computer, help me with X, Y, or Z. And then that will create a thread and multiple people can, you know, do teamwork together to try to figure out how to get to the bottom of it. But you can engage with computer and email in Slack, in the web app, on your phone, any of the mediums.
Speaker 2It all makes sense. You brought up demoing with computer. It got me thinking about how you think through team composition at Perplexity. You said you don't have sales and engineers, but then you said, and I'm assuming that means pre-sale, because then there's support, which is like post-sale. Then there's FDE that you, as part of your remit. How do you structure the team if your statement is like no sales engineers? And is that only relevant in a company that maybe has an app and an API versus if you went and worked at like inference, you would have to have more of a sales engineer? Like, how do you think through team composition for you guys?
Speaker 1First off, I don't think everything's binary forever, but it is true that we do not have sales engineers. We don't have sales engineers in the traditional sense of, okay, I'm going to split out this role into a technical, commercial seller who showcases the product and another person who doesn't bear the burden of figuring out the product in a deep manner and is there for the commercial half and the relationship half of the sale. And while I think that makes sense in some products that are so complex that they're impossible to grok, I think we've kind of overdone that in go-to-market. What I think about in terms of selling a product like Perplexity Computer is what you are doing is helping people figure out how to use computer better rather than needing to always deeply understand their company and exactly how they operate. Sure, there is stuff when we engage with law firms or professional finance that's a little more specific. We can talk about verticalization some other time. But if you are at Perplexity, you are probably ripping enough usage in computer that you use what the average even AI power user outside of Perplexity uses from a credit consumption in like a day versus a month. And so if that's true, roughly 30 days of AI experience, which is like longer than most people have even been engaging with AI and six months in, you have many, many years of AI experience. And so we just want the AE to come in and lead with curiosity and in a non-meme-ish version, be a consultative seller. Oh, well, this is really cool. Well, I asked computer what use cases it thought would be relevant to you based on research. It sounds like some of those are relevant. Why haven't you gotten this from other AI tools? Where are you struggling? Oh, that's cool. I'm not sure. I'm not sure. I'm not sure. I'm not sure. Cool. Let me show you that. Or here's where I struggle with that in my day-to-day and here's how I've changed the way I operate. I think that with the power of an AI product that can build its own use cases and do some of the more technical stuff, the setup of a demo is often in most enterprise products, by the way, is more complicated than actually showcasing it. I believe that a person who's a really great consultative relationship building seller, plus the power of our own product, I just don't want to add another role in because I think it's confusing to people.
Speaker 2Agreed with the computer use case. I was also thinking, what about for the API perspective? The reason I say this is because I was an app seller for a while and then I started interviewing at Stripe and Plaid and I couldn't get through and they kept saying I wasn't technical enough. And that was way before you can maybe have a chat to catch you up to speed maybe a little bit faster. I think that no better person than an AE who uses computer to be able to like, hey, this is how we use it. You guys sell an API and you sell computer on the app side. How do you think through interviewing salespeople then? Sure. I think learned experience is a very good thing that gives you a leg up versus people who have other learned experiences. Yeah, we're looking for people who are like, okay, you're either on the cloud team or you're on the API team. If you're on the API team, we want people from Twilio. If you're on the cloud team, we want people from Slack. So app layer versus more technical. Do salespeople sell both of those? Do you separate them out when you interview? Like, hey, this person's our API, AE and this person's more of our computer AE and they have different technical backgrounds. Like how do you think about interviewing salespeople in that regard?
Speaker 1Sure. Yeah, we're looking for a little bit of a different profile on the API side. But I also think that the promise of like dev products and APIs, most of them are supposed to be self service in nature. Yeah, maybe this is another thing we've tricked ourselves into a little bit that certain things or certain three letter acronyms have to be less accessible than other names or acronyms. I think that the first time I've ever done something was the first time I ever did it. Yeah. So like, I don't think all the API sellers have deep engineering degrees. No, I think they were just at some company and that company took a bet on them for the first time. Exactly. And so what were the core things that led them to be so smart, so compelling, so endearing that they were going to learn the thing? I'm a little less concerned about exactly what product they were selling before and more that they have a high slope versus Y-intercept on my specific product. The companies, the Venn diagrams that overlap of like computer versus API usage, there is overlap, but it is not one for one. I have a separate team selling API. Makes sense. And they can come in as an overlay in super complex deals. But back to the, you know, if you have a product feature, should you productize it or not? Every time that somebody in my generalist AE population struggles to sell something in API, we should enable everybody to get past that issue and then do it better the next time so that you don't need more complexity in this motion and then reserve only the most complex API deals for small set of sellers.
Speaker 2There is an elitism a little bit of like these technical sellers. Like there is this like, sort of magical skill that until it's like, well, how did anybody get an opportunity unless they were a software engineer that had a background in it? I think there's a lot of those traits that you kind of were talking about, like, oh, I, you know, I want grit or ambiguity. I think this is just a good segue into like thinking about how to interview salespeople.
Speaker 1Isn't the world just full of people who are given opportunities to do things that they weren't qualified for? And haven't those in most people's experiences turned out to be the best, the best people at whatever? Yeah. How do you interview? I have, no issues with how other companies do it, but I do think there's a lot of terms that we've started to use. You know, if you say the word so many times, it eventually loses its meaning. Like, I don't know, grit is one of those things, or you tell people, I need you to be able to thrive in ambiguity. Yeah. Like, yeah, that's true, but sometimes it's hard to evaluate for those things, or we try to evaluate for them in a really nebulous way that pulls you out of your interview and asks for past stories and stuff versus just like, maybe we should keep it super applicable. The job you're interviewing for right now and how you would think about applying those concepts to this job. Um, what we interview or what I interview specifically for by the time somebody gets to me, I like, I appreciate clarity of thought and communication over everything else. Yeah. Like that, that is the job people who think quickly and then communicate those ideas clearly. And then of course, follow the virtuous cycle of self-improvement we spoke about. Maybe this will sound almost a little bit different, but I think it's a good idea. I think it's a good idea. But a lot of that is to try to understand how people got into the room and get, get that forced them to be very explicit about how they make decisions that'll often start with something along the lines of, oh, I'm really excited to, you know, be interviewing a perplexity. I, I love perplexity and I think it'd be a great company to work at. Oh, that's great. So what were the explicit criteria that you decided upon that would make a company a good fit? You know, even just that question right there, like generally we'll throw somebody off because, uh, well, you're super fast growing. Okay. That's fine. That's a part of the list, but you give, you give people some leeway to, oh, okay, well, that's great that we're fast growing. What other criteria did you use? And then eventually good people will get to, because they're clear thinkers and communicators. Well, I was looking for A, B, C, and D. They like PLG too. They've used the product. Okay. Dig into one of those. Oh, you've used the product. That's great. How did you think it was different than other companies that was good enough to move it up in the criteria set on your list? Because you can PLG some of our competitors as well, or eventually pushing them down the line into, okay, that is fantastic. And now you'll get my favorite interview question. You've now laid out the criteria for how you evaluate companies. How explicitly are you evaluating me against those criteria? What did you know you needed to get answered for me? And how am I being graded against the criteria that you set out? Yeah. And most people will go like, and it's totally fine. I mean, what are the best answers I've gotten recently? Someone we hired in the last couple of weeks said, well, I'm really concerned about companies that are growing super quickly and whether or not they're going to hire too many account executives. And I don't yet feel confident that I know from perplexity, whether or not you have a clear idea of how you won't overhire and how we'll all be taken care of. Great question. And if you can give me two or three of those questions, it's like, wow, this person is organized. Has clarity of thought is willing to communicate with me and be open.
Speaker 2Totally. There's a lot of you said there. I love the clarity of thought thing. We just wrote an article. We run a recruiting firm as well. So it's like we put out content of what are the best candidates doing? They are bulleting out their answers. Hey, why do you want to work at Prolexi? There's a lot of reasons why this I'm really excited. Here's a three main ones, one, two, and three. And there's like a hack a little bit of you just want that person to think generally you want to be a clear figure yourself. I mean, that benefits everyone, but you want the other person to think that you're definitely a clear thinker. And the way that you do that is through, Hey, these are the three things. And the other person can like easily comprehend that. And then I love like in a PLG business, you can almost show receipts like, Hey, I use the product. Here's actually, thanks for asking. Here's five things that I ran up against that is differentiated that you guys should do that. I liked about your competitor.
Speaker 1Oh my God. Nothing, nothing gets me more excited. In an interview than somebody telling me that we're doing something wrong and having something to back it up. I don't even know if you're right, but I find it super compelling that you cared enough to figure it out before we got in the room. And I love that you guys are giving that potential candidate. We put it in an
Speaker 2article and then we did it on a podcast and we just send them that we're like, you read through this. These are like the things, and it's not like you're a bad candidate. And then you read that. And then all of a sudden you can gain the system. There are just really good candidates that don't even know that they should be doing that. They have that in them. They just didn't even know that they're getting graded against that. So if you don't know, you don't know sort of thing.
Speaker 1We're also slowly hurting people by having them engage with tools that make them think the opposite way, which is the AI slop cannon produces a lot of words. And my experience is most executives, which is who we sell to and who's going to hire you think in a five paragraph essay or a bulleted way. You're going to tell me what you're going to tell me. And then you told me super clearly. And then we agreed on what we're going to do next. Let's move on to the next thing. And, um, more people would benefit if they got the advice you just gave.
Speaker 2The new M dash is just a lot of words. So it's like, if you, you, you like send over a doc, it's like, Oh, I did all this prep. It's so long. You're like, okay, like you clearly didn't do this. Wait. So the overhiring question is it's one that we get a lot, but you had said capacity planning and AI. Like if you're somebody that's like, I have a really rigid plan. It's like, you got to throw that out the window. How do you think about capacity planning in a company? That's like one of the fastest growing.
Speaker 1Yeah. Well, one at every corner, you should probably challenge yourself to figure out how you become, you stay just as efficient or more efficient than you were the other day. So there's always got to be a set of stuff over here that says, what are the levers we have to improve productivity? And it's not just reps, but of everybody. But I think about it less in some spreadsheet that works backwards from, it would be really nice to grow X. And then I go through annual planning. One time rep productivity is some number. It takes this long to ramp or rep. We're going to have this much expected attrition. And therefore you need to put this many people in seats by this time. There is some version of that, but you should be comfortable with that changing all the time. And I think the world is moving a heck of a lot faster than that. Do you need to hire ahead of potentially current pipeline or something that makes you feel wildly comfortable all the time? Yeah, you have to because AI is moving so fast that the next incredible thing that we or someone else builds that will generate a ton of demand. And that's what we need to do. We already want people in capacity to be able to go sell that thing. But the way we think about it a little bit differently is put a really far out goal to grow 10 X or a hundred X from where we are. There's a lot of humans that you need at that super far out number. And instead let's build a machine that can put more people into it, a recruiting machine, a hiring machine, but at every next person, stop and take a breath and ask yourself of all of the things that you're doing that the last person is becoming productive that you're ramping. If the end plus one hire is going to be productive and think of that on maybe like a green, yellow and red light kind of spectrum. For us, that's, I like to say a couple of cool metrics. Been here five months and hired a lot of reps and every single person is close to deal in their first 30 days. Okay. So what needs to happen to have reps close a deal in their first 30 days? Well, you have probably got to be taking sales meetings at the end of your first week. That alone doesn't tell you your business is ripping, but it's a good sign that people can get on the board. It's good for sales team morale too. It's just like, is this stuff that's not on
Speaker 2the spreadsheet, but it matters.
Speaker 1Yeah. Yeah. I mean, it doesn't, it doesn't define overall productivity, but it says like, I, you know, this person was, was in front of enough people that they got one, then they know what our process looks like now. And maybe we even optimize for that being a little bit of a smaller dealer or a quick win or a PLG conversion deal. But okay. That's, that's a green light for me. The next cohort of reps, how are they doing when I'm hiring rep 105, 106 and 107, you're ramping rep 71, 72 and 73. And if everything looks golden for those people go hire the next set of reps too. But when you wake up and you do wake up in these companies every day and something freaks you out and it's yellow, well, I don't know if we're doing as well in this industry anymore, or this crop of reps, I'm a little bit worried about that. Or our tooling is breaking because we just hit a certain scale. Go optimize. For solving that problem and get back to green really quick. But if you get to red, be comfortable slowing down. that's just life the plan doesn't change once a year or even once a quarter it changes every day and hopefully be prepared and have an engine that can allow you to accelerate the plan even the other way if you need to yeah like hey maybe you need to be hiring five times as many reps do we have a plan for what that looks like if the metrics told us that we could
Speaker 2yeah you should
Speaker 1have that contingency plan ready to go as
Speaker 2well yeah i love that framework of yeah getting the green but then when yellow trying to get back in the green i was always at sales orgs that were in like the yellow red uh it was bad company selection and that's what i'm trying to do now is help empower reps to make better decisions sales mafias in like the sas era and the ai era that you admire and ones that you haven't
Speaker 1worked at the app dynamics crew and what they've all gone and achieved recently is really impressive and good people come from there mongo i think continues to produce some really fantastic sellers who are good at usage-based plans and i think they're going to be really good at that and i think they're going to be really good at that and i think they're going to be really good at that and i think they're going to be really good at that and i think they're going to be really good at that plus you know the yes they work with customers who just rip and it's really sticky but they have real good process and i gotta tell you a seller who has worked at snowflake for a long period of time yeah they're hitting outside of their weight class i've been really impressed with those folks
Speaker 2all makes sense and i've heard good things from those three as well um what about from like an ai native who from the outside looking in whether it's a certain team that you've been some of the people you've had on this pod and we'll probably have on soon by the way i have deep respect for
Speaker 1them but i think they're really good at that and i think they're going to be really good at that and i think they're going to be really good at that and i think they're going to be really good at that and i think a lot of us talk about the term go-to-market engineer because of versell i think they are fantastic and gene is an incredible c-r-o-c-o-o you should because maybe i'm a little bit biased here for other reasons but i also think the team that they're going and building that's a lot of ex-twilio folks over at fireworks yeah is really fantastic too what was even the chasm for
Speaker 2you to say yes to going into a sales or you were here chief of staff under a cro maybe it was at that point or um and i think that's a good point i think that's a good point i think that's a good i just can't picture you in a role outside of sales just based on your overall or this last hour so when did you say yes
Speaker 1so at that point i think i was pretty confident i was going to go back in into sales the hop right before was i went to a seed stage legal tech company to go run go to market which was you know one a year something and one marketer at the start so that was the first time i i directly managed people who had it back but for me it was i got really lucky in one of those product marketing roles where it was early enough that you spoke to the customers all the time there were definitely coaching sessions i had with my cmo and other folks within the organization were like hey you probably should spend less time in front of customers you're not a sales engineer you're not a rep and i thought well that's a really good personal signal of what i enjoy and where i like spending my time and i always sat with the aes and stuff and so i think it was good that i did those things before but i think there was a moment when i just knew all paths were going to lead me to sales it wasn't a compensation thing or how high you could go it was just what a dynamic fun job and man sales people are fun to work
Speaker 2with 100 i could have seen you as a like a top litigator as well kind of like tom cruise and a few good men i can see commanding a courtroom but you got out of your 1l and then you were yeah you're like i got taller taller and less attractive he's definitely aging well i keep seeing him at different memes but steven this is this has been great man thanks so much for coming on the crew podcast i appreciate it it was such a pleasure thanks for hanging with us if you like this one follow the show on spotify Apple Podcasts, and YouTube, and we'll see you next week. Cheers.
PodcastTranscript AI