E69: Stop Paying For SaaS: The AI Playbook Driving A $1B Trajectory | Ghazi Masood, CRO @ Replit

Ghazi Massoud is the CRO at Replit, where they've grown revenue from $2 to over $150 million in less than a year. And now they're tracking to a billion-dollar run rate by the end of 2026, all in the service of one mission.

We're democratizing software creation, and the next billion software creators are people who sit outside of the engineering and development organizations.

In today's conversation, Ghazi draws on his experience as a Vibe Coder, but also as a multi-decade revenue leader, as we talk about the best ways to think about build versus buy.

The big behemoth question is, hey, are we building our own CRM? No, we're not. But then everything we're doing on top of that, they're all built on Replit apps.

Plus, we talk about the challenges and solutions around talent selection, change management, and how to structure a team when you're growing at a truly insane pace.

There's a lot of firsts happening in this business because the business has grown so fast, so quickly.

Welcome to the Revenue Leadership Podcast. A quick PSA. There's a change coming to the Revenue Leadership Podcast. I've moved my show to a new YouTube account. So if you enjoy the show, I recommend you go subscribe so you don't miss any upcoming episodes and a bunch of other content that will be on there. And you can find the new channel at youtube.com/@Kyle-Norton. The channel name is Kyle Norton, the Revenue Leadership Podcast, and that's where we're going to be posting all the future episodes, clips, and a bunch of new stuff as well. And, uh, if you're wondering, the audio podcast will still be in all of the same places that it's always been, uh, on Spotify, Apple Podcasts, Snips, and so on. So if you listen in audio-only form, no need to do anything. But if you watch on YouTube, now you know where to find future episodes. And while I'm at it, I would love if everybody could support the show by liking, subscribing, reviewing, leaving a comment. on whatever platform they listen on. It helps new viewers find the podcast. It's really helpful for us. And I always forget to mention it. Uh, if you love the pod or even like it a little bit, uh, do us a solid and go review it, comment, uh, share it with a friend, and that would be most appreciated. And now let's get back to the show. Today's guest is Ghazi Massoud, who is the CRO at Replit. And Ghazi has spent the last decade building enterprise sales motions on top of PLG companies. Through hypergrowth at Auth0. He was the CRO at Retool and now at Replit, where revenue has gone from $2 million to over $150 million in less than a year. And Replit's now tracking to a billion-dollar run rate by the end of 2026, which is truly insane. And Ghazi and the team are hiring 200+ people over the next 12 months and rebuilding GTM from scratch in this AI era. We're going to dig into what it's like to build an AI-native revenue org. Replit, for those of you who don't know, is a vibe coding and agentic building platform where you can basically build anything. It was actually the first vibe coding experience I had ever used because I had done that 100 Days of Code course in like 2022. And so you could build anything and we're going to talk about what that means and what architecture looks like and why some of that answer involves scrapping the traditional CSM and how the org chart looks like. So I'm really excited about this one. This is going to be a ton of fun and thanks for joining.

Yeah. Amazing.

What did I miss in your background that's worth noting and maybe just give people like the snapshot of what, what Replit is?

The background is pretty spot on. I've been at this for a long time. I've been a go-to-market operator for going on 3 decades now. The first part of my career was like big, big companies. I was at, you know, did my stint at Avaya, Microsoft, Oracle. And then after Oracle, I've been in venture-backed startups, which some of the ones that you actually talked about. So I've been at this go-to-market game and I've seen different iterations of, you know, the on-prem to cloud boom, the dot-com boom, the financial meltdown. So I've seen this and, you know, being front and center in AI revolutionizing the entire landscape is like super, super exciting. And Replit is just a fantastic place to be. And for me, for me personally, I think from a career standpoint, it is a perfect, you know, natural evolution of my career as an operator of kind of being in this AI world of where I've been. For a lot of people, what they don't know is the very first company that I went to after Oracle, the big corporate America company that I left, was a company called K2 Software, which was in the low-code, no-code space. This is when no-code first came out and the art of the promise was empower knowledge workers to go build. So I've seen the evolution of low-code, no-code to now what is AI. So it's just a perfect trajectory for me personally. And, uh, super excited to be here.

Replit is low-code, no-code on super steroids.

So that's right. That's right. Yeah.

That's what I'm just explaining the business for folks.

Yeah. Yeah. I mean, that's the last company I was at Retool. One of the reasons I left is we could not pivot to AI fast enough, right? And yes, so with AI coming onto the scene, the relevancy of low-code, no-code is no longer relevant, right? You can't really— low-code, no-code is— I actually did a podcast on this a few months ago where I actually basically went on and said low-code, no-code is dead. Low-code, no-code is dead. And if you haven't pivoted to AI, you no longer have a business, right? So yeah, so Replit, what our mission is, we're democratizing software creation and we feel the next billion software creators are people who sit outside of the engineering and development organizations. And we want to empower anyone to go build irrespective of their technical background. In the previous world, if you had an idea, you would have to go take your idea to the development and engineering teams to go build. We want to remove that bottleneck and we want to basically empower the average person sitting in, you know, the business organizations to go build and create software. So that's kind of what we're all about.

24 months ago, you know, the mission was the same and people would've like sort of rolled their eyes at that and being like, yeah, okay. You know, Vibe Code, you can Vibe Code like cool throwaway software and like what is the real enterprise fit for something like that? And now you fast forward 24 months and a lot of the experiences I engage with day to day now are things that we've internally built. Which, which is sort of a crazy experience. Like people don't send a dashboard or a spreadsheet anymore. Like any spreadsheet is now an interactive web app. And, and like, that's just, it's just like, whoa, like why wouldn't I, I put this into an experience people can engage with that's more visual. It's like takes 2 seconds and that's the default, like already.

I am the perfect use case for that. So, you know, like in previous roles as a CRO, You know, you've got a big support team. We got a big enablement team. They're building content and decks for you, for your presentations and whatnot. I'm whiteboarding my own decks. I'm literally whiteboarding because I have ideas in my head and I can literally put my ideas on like, hey, I want a 10-slide deck. I want to go tell this story. This is the kind of one I wanted to go do. I literally do it from my phone when I'm sitting in my bed when I know I have to go give something. And then literally I've got a 90 5% ready stuff that I just take the last 5% and I kind of just make it my own. It's been working just great for me. In fact, I used it when I did our first annual kickoff at the beginning of this year. You know, we have a small team. I don't have the enablement resources or the people on the team that I had at my previous companies. So I had to just go do it and I just did it and very, very satisfied with the results and it's something I use all the time. So all my, all my decks right now and my presentation decks are built in Replit. Are built in Replit. I do that right now. I don't even use Google Slides. I can export to Google Slides and PowerPoint, but the actual building is in Replit.

From a, like a CRO tooling, stuff that you build and use yourself or automations or apps or workflows, what are the principles to do that really well? Feedback I hear so much is like, yeah, like I can try and it gets like 90% there, but it's not really like quite good enough. and I'm like, end up editing it a whole bunch. And for me, like I've just sort of organically learned how to do this on my own and I don't have a great answer for the principles of like, okay, this is how to think about the context engineering or, or what you should be doing with the specific prompts and the chains. But how do you think about advice for CROs who are like not super technical in terms of how they could be building the most powerful automations and workflows for themselves?

And let me put it in two buckets, right? So one would be a CRO that basically needs the tools to effectively do their job. You know, you need a view of your business, you need a view of pipeline, you need a view of deals, you need a view of customer health. You kind of need the entire gamut from start to finish. So from that standpoint, I can tell you like what we're doing here and how I've kind of morphed into that. So the big behemoth question is, hey, are we building our own CRM? No, we're not. You know, yeah. We had HubSpot. I'm moving from HubSpot to Salesforce. And that wasn't even a question internally in my business or for me, like whether I was going to go vibe code my own CRM. Just on that note, I do have plenty of customers that are vibe coding their own CRM and we can talk about that here in a minute. But for us, the decision was to not go build our own CRM. So we use Salesforce and we just migrated from HubSpot to Salesforce and that's still going on right now. So Salesforce is our system of record. But then everything we're doing on top of that, for me to be an effective CRO and get visibility into my business, they're all built on Replit apps. Mm-hmm. There are, you know, a variety of different iterations. We have something that we call a revenue copilot, which is my all-in-one dashboard that gives me everything from top of funnel to conversion rates, all the way down to deals, sales stages, closed one, everything. I pull gone calls into that. I can click on any customer logo, look on what it is. It's just, it's basically what I live with every single day when I wake up in the morning and what I look at and listen to. I don't even touch Salesforce. I just kind of go into that. Then for our teams that actually work with customers, we created basically an application that we, we have all these funny names for them internally. We call it Repleebook, but it's basically our version of Gainsight. And basically a customer health dashboard that basically gives all of our customer advocates and our field engineers complete visibility around every single customer, you know, their utilization, their customer health score, when's the last time they, you know, what the engagement process looked like, are they healthy or not healthy? So basically you can think of that. And then, you know, then we have other things like we use our, we basically built our own CPQ and quote the cash. Wow. Okay. All order forms, uh, that we send out to our customers, it's a Replit app that we go do that. Um, we're building our own version of forecasting. So think of Clary. We're basically, again, on top of Salesforce, we're basically building our own version of that. And then our BDR team uses a bunch of research and, uh, script writing applications. That they actually go do. We use Nuxt, by the way, but we use specifically for the dialer, but all of the other stuff we kind of just do on our own. So again, there's a— we can have an entire conversation on the AI GTM stack. Yeah, that we have. And in fact, um, there's a session at SaaStr that I'm doing where I'm gonna actually showcase and dogfood, uh, some of the stuff we've built, and we're gonna do it this way. So I'm gonna be— I have a session I think in the middle of the week to go do that, where I specifically talk about the AIGTM stack. And then, so then the other bucket is as a CRO, like things you do kind of every single day and like, what do you, you know, like, what do you use? And that's where, you know, so again, I'm, I don't want to sell Rappland, but the one thing is we sit in the browser, right? We sit in the browser. So it's just up in the day of the thing. So what the way I use it is if I have ideas or I need stuff done, I just go prompt it in Replit and just build it versus going to like my enablement person. Hey, go, I need this. I need this. I just go prompt it because it's faster for me. It's quicker for me and it gets me what I need. Yeah. My own terms in my own way, kind of right away. Right. I don't have to sit there and wait. I don't have to go explain it to anybody that I want. I just go explain it. I explain it to a Replit agent. That's actually what we call our product. And the agent knows my style, it knows my personality, and it actually creates what I'm asking it to go do.

And so for the personal side, like you mentioned, agent knows your personality and style. Is that because you specifically created a context file about your likes and dislikes in business context, or what's the scaffolding behind being able to quickly spin up one-time or like personal apps? What makes that go well or not well?

I think it learns and iterates on your previous engagements and what you've done and what you've asked it to do. And I think it just gets better and it gets smarter in realizing the context kind of actually behind it. So yeah, I mean, like this notion of like a one-shot prompt, I think has gotten a lot better than it— where it was probably like 12 to 18 months ago. And again, the more specific you can be, the better results you're going to get. And I've personally noticed that improvement as well.

How much other context do you have in there to help your personal apps do the right things? Like, for me, I've got, you know, a context file that's local and in GitHub so that my OpenClock can get at it, but it has like all the stuff about our business objectives and how we operate. And it's, it's built like a, an understanding of how I decision make and problem solve. Cause I fed it a whole bunch of content and, and like prompt it to be like, hey, analyze this stuff, come up with like a, a deep understanding of my approach so you could be the most effective AI chief of staff. And have you had to proactively do that or is it just about Hey, use the tools and—

No, I haven't gotten to that level of sophistication. I wish I had. I'd do that. I haven't gotten to that level of sophistication. It's me just being specific with agent on what I actually need and the agent kind of going back to my previous context of the interaction. Okay. Learning from it and just getting smarter.

That's a specific like feature of the Replit ecosystem is this like recursive learning then?

It's built in. Yeah, it's built. I mean, we're in Agent 4 right now, which is our last iter— that was our last launch that we actually did. And Agent 4 had a bunch of unique things that the market doesn't have where you can actually run multiple parallel tasks at the exact same time. You can go do that. You can have collaboration. There's a lot of different things on it. We came up with a canvas, a very robust mobile app thing. So, but yeah, it's basically the agent is just, just getting smarter. And I think what Agent 5, I think it'll get even better. Yeah.

Sometimes I wonder if I've like overcooked with my setup, like having OpenClaw and it's in this like DigitalOcean droplet and like the whole configuration. 'Cause I now look at the tools that Claude offers and like Replit's always been so such a, like a great building platform. I'm like, should I have just adopted one of these tools more holistically? And so that's sort of the advice I give to people is like, actually, my approach is probably, it's probably overbuilt because I wanted it early. Like I wanted to get my hands on this stuff as soon as I can. And I find it really interesting. It's a great learning experience to try to, to try to figure it out, but it feels like I could get most of what I've been able to build, which has taken a bunch of effort just like in a tool like Replit.

Yeah.

Try it out.

I mean, I tell people just try it out and then, you know, and see what it can actually go do for you.

The advice I would give to people is that if you're going into these new, new tools, just make sure you bring as much of your previous context with you. And there's prompts that you can run in your— in Gemini or ChatGPT or Claude. Just be like, give me everything that you know about me, download everything into a markdown file about how I think, my preferences, da da da. So you can at least start this new experience with a with a baseline. I want to just zoom out and ask about this experience as a CRO because there's only a small handful of people in the world that have been in a company growing at this type of speed. Like companies have never grown at this rate ever before in history. And now there's a handful of sort of Anthropics, OpenAI, Replit that are growing at like 50x year over year, like 10x year over year at like multi-hundred million dollar run rates, which is just mind-boggling. Like what is it like just as a CRO trying to do your job day to day with those like dizzying numbers and trying to hire 200 reps at a time? Like what is it like?

It's pretty nuts. You know, context switching of talking to a customer on a deal and then being in an interview and then looking for executives to get them right. I mean, it's, you're context switching so much every single day, every single hour. Right. If you look at my calendar, it's like a hodgepodge of, you know, so I just, given that we're building a small team, I had this process in place right when I started is I do all first screen of candidates right now, just because I don't, I don't have a leadership bench, right? Like I, like my org at Retool was 300 people. I don't, I'm not at 300 people right now where I've got like first line, second line. So I did all first interview screens just with every single candidate. And basically if they kind of passed my interview screen, then I would actually get them off to the panel and the team. From a day-to-day perspective in a hypergrowth company like this, as a CRO, you know, you're wearing multiple hats and I'm like rolling up the sleeves and I'm at the ground level. So I'm not a chair CRO that's looking at spreadsheets and, you know, doing orders. I'm actually doing the work, right? I'm actually doing the work and I'm actually just getting my hands dirty and it's actually like super, super fun. We're building, right? So we're building up, we're bringing on our first BPS sales here in a few weeks here in the US. You know, I'm opening up our London presence for the very first time and I'm bringing on a head of EU GTM, right? So we're building it and it's going to get better. But right now in this hypergrowth scale, I have to be involved and I have to kind of have the pulse on the business to kind of get it to a point where I feel good about it. So yeah, it's been, you know, we're basically building a B2B GTM company within this massive company. And it's phenomenal. You know, like everything we're doing for the very first time, you know, rolling out territories, very first time, rolling out comp plans, very first time, rolling out forecasting cadence, very first time, rolling out, you know, the post-sales motion. So there's a lot of firsts happening in this business because the business has grown so fast, so quickly. There hasn't been kind of this classic GTM playbook or structure that we all come from as revenue leaders. And I'm kind of incorporating that in this AI-first world, in this AI-forward way here at Replit.

And what do you bring over that is, you know, like similar to the previous playbook and what's totally different now?

Yeah, I mean, similar to the previous playbook is, wait, We need salespeople. We're not gonna go automate salespeople, right? I think what's different is the profile of the salespeople I'm bringing on board. Like in my previous 30-year career is different. You know, previous 30 years is like, hey, go get the top 1% seller that has, you know, blasted quota, comes from Oracle, Salesforce, has that pedigree. Just get them in because they're just a killer, right? I'm not optimizing for that right now. I'm optimizing for more people who are passionate about AI, that are passionate around our mission, and that really, really believe in what we're doing. And I've taken a chance on a bunch of people that are here right now that come from non-traditional backgrounds that have never sold before. but have expressed a very deep understanding of who we are and what we do. No, I'm not saying technical, but just they, they're really passionate about it. And I'm, and they're in sales roles for the very first time in their career and they're doing exceptionally well. They're doing exceptionally well. And I feel I can train them on the sales elements and part of it out. So that's the different, that's the, that's a different part of it. We, we invest in technical resources. because even though Replit is easy to go out and show, like, you know, we're not one of those companies where if you're a rep on a, on a deal, you don't need to bring an SE with you to do a demo of the product, right? We expect the AE to take the conversation almost to 100%, oftentimes to like 80, 90% completely on their own. And then once the customer starts digging into, you know, like data residency, security, deployment configurations that may be out of the depth of a classic AE, you know, we would bring in our field engineering organization. So I still have a technical organization, but unlike previous SaaS companies I've been at, you've been at, where AE would get on a discovery call and they would have their SE with them on the discovery call, like even if the SE said nothing, right? Or the AE said nothing, they would act. That's not our model. That we're building here.

Yeah.

AEs to actually be product experts and actually take the conversation pretty far because the product is just so easy to show.

And are you taking these non-traditional folks into AE roles? Like straight? Okay. Wow.

Yeah. Yeah. Yeah. I mean, and I did a podcast on this with Chris at Vibe Scaling. He asked me about this. Just a couple of anecdotes for folks that haven't or didn't listen to that. So when I got here, our number one seller in the company, and the person was already here, was, came from the military, was in the US Marine Corps, never sold a day in his life. Absolutely killed it. And now he has a much more passion for the technical side. So he's actually changed into a technical role, but he was initially doing sales. And then our, one of our top commercial sellers for our downmarket business. Who's still here in that role was a schoolteacher, never sold, doing exceptionally well. In fact, I just did a pipeline review this morning. He has one of the largest pipelines and forecastable deals for the quarter. So doing exceptionally well. So yeah, a lot of these non-traditional folks are in sales roles for the very first time. And again, I want to make sure people, that's not our entire org, by the way. There's a balance, right? I do have people who come from enterprise sales who've done enterprise sales before that have been like in the top 10% salespeople. We definitely have those as well, but we're balancing it out. We're not just optimizing for one or the other.

And so mission centricity and sort of a pull to the product is a part of it. What else did you see in those people that, and what else are you looking for that's most important for you?

The biggest thing I'm seeing and now just being in the trenches and, you know, listening on calls with them and being in meetings with them, the way they explain the art of the possible, that the customer really bought into our vision around what the art of the possible could be. And, you know, it's like this light bulb moment is like they've elevated to trusted advisor immediately. Mm-hmm. Like they are not viewed as a salesperson in front of the customer's eyes.

Yeah.

They're advocate. They're like their AI advocate. They're like their AI trainer. They're there. You know what I mean? That's the status these folks have elevated up to. And that's kind of what I look for is, you know, can someone, when I do these 15-minute screens, you know, the question I ask is, hey, you know, how would you explain Replit in like the first 15 seconds to someone who doesn't really know who we are and what we go do? How would you establish that level of trust and kind of explain kind of our mission? And usually you can kind of tell the people that, you know, are the product people that are just passionate about it because they've, you know, prior to coming to interview with me, they've built like a bunch of apps for their, you know, kids' Little League team to their hobby apps to personal website. So they know the flexibility of the platform. and the ones that are just passionate about it can just rattle that off like that and you immediately know, oh my God, putting them in front of a customer, that's going to be like amazing.

What's the shape of the team today? So, so you've got this like insane PLG business that, that's going exponential. And so how big is the sales org today? Where are you guys focused? Like what, what is the, the structure?

Yeah, we're building it. Like, so GTM all up, um, I just got the, we're hovering right around 60-ish right now.

Of all up.

So we've added a lot of people since I've been here and we have a lot more to go do. But yeah, the makeup, just to run it down for you, we're very, very high level. So when I first got here, I mean, in a PLG business, 90% of our pipeline was all inbound. It was all inbound people basically hitting our website, wanting a demo. And so we had inbound SDRs that will actually go do that, but I'm actually building an intentional outbound SDR team. So we actually opened up an office in Salt Lake City and we're basically building an outbound motion and an outbound BDR organization to supplement the inbound. And then eventually I want to automate inbound. So inbound will be like fully automated probably by the end of Q2 where we actually won't have humans triaging inbound. Inbound will be fully an agent that gets it, you you know, books the meeting, et cetera. And then we'll just have like an outbound motion. We'll still have to get inbound pipeline, but inbound pipeline will not be managed by a human. It'll be managed by an agent. So anyways, that's on the front end. And then on the field side, we basically have, we segment our business between commercial and enterprise. Um, and commercial for us is less than 2,000 employees. Enterprise is greater than 2,000 employees. And we basically have account executives for each one of those segments. And then I don't have, and I'm not building customer success here. I have a role internally that we call account managers, which is a pretty common term that many revenue leaders know, but they go to market and introduce themselves as Replit customer advocates. So we have a role, but we call Replit customer advocates that basically come online when the AE closes the deal to help with the customer on, to onboard the customer and then continuously plant seeds to go out and retain and drive future use cases of Replit within that account. And then the last role we have is what we call our field engineering organization. Think of them as like a forward deployed engineering team, if you will, that spans both pre and post sales. Spend the bulk of their time post-sales, but they're also available as pre-sales technical resources. So that's the makeup of our GTM team. And then I've got rev ops and enablement as well. Given how many people we have to go hire, we take enablement very seriously here. Yeah. You know, like we don't just hire and throw people at the phones or give them their books. They all come to Foster City, which is where we're based in California. They do the company onboarding. And then, um, the woman, Yo-Yo, who I've got here, who runs my enablement team, she's got like a robust GTM onboarding program depending on your role. So if you're a BDR, you go through a BDR track. If you're a, you know, AE, you go through your AE track. If you're a customer advocate, you go through your customer advocate track. If you're a field engineer, you go through your field engineering track. Um, so we do spend a good amount of time of making sure all the people we hire are properly enabled and set up for success.

The ARR per employee must be just bonkers. It says only 461 employees on LinkedIn, which is usually off by some, some of it, but it is bonkers.

I mean, the, the one thing that it's hard, right? Because we had to put quotas for the very first time. Nobody was on quotas up until January of this year. When I first got here, everybody was on full OTEs. So we changed that, put everybody on quotas, and it was kind of hard. To put a line in the sand as to what quotas to go use. So that's why I did half-year plans because I knew things would be moving so fast. And clearly with, you know, we're sitting here on May 1st with Q1 behind us, everyone is smashing their quota. In the TL;DR, and my team knows this, I can talk about it, is, you know, in the second half, we are going to raise quotas up and get them normalized to what it is right now. Obviously, that we actually have some baseline and some data against it. Yeah.

It's a law of startup physics. Quotas only go up. Yeah. Yeah. It's like gravity. And if quotas start going down, then that's when you got to work.

That's right. That's right. I mean, I still want to create a winning culture, right? I still want the winning culture. I want people making a ton of money. So we're still going to build it with that mindset. But we have to be judicious to our board and finance and the rest of the team and do the right thing.

Momentum is a powerful tool for turning sales and customer conversations into go-to-market intelligence. Using GenAI, it extracts, analyzes, and automates customer intelligence across your GTM org. The best part? It's not a new platform for your team to adopt. It integrates seamlessly with your existing stack, or it can straight up replace your conversational intelligence tool. At Owner, we ripped out Gong and went all in on Momentum. It writes back to our Salesforce directly, captures forecast and churn risk, autofills CRM fields, shares product signals, and tracks sentiment. Companies like Cursor, Zscaler, Ramp, and ElevenLabs use it every day. Check it out with a free trial at Momentum.io. When you grow at that rate, what's the first thing that breaks? Like, what, what are the things that are just like, can't keep up and you feel like you're duct taping together? The rev ops.

Stuff, right? We are doing so much change management, right? Territories, comp plans, quotas, systems, hygiene, you know, core, you know, the way we structure deals. We also went from seat-based pricing to usage-based pricing. So there's so much there right now. And with new people coming in, The legacy folks and, you know, ownership of accounts, that is, uh, right now our biggest thing we need to streamline and, uh, get better at. It's working, but that's— there's, you know, that is where I need to be spending our time and where the team is kind of spending their time just to kind of make sure that process— and right now it's the migration from HubSpot to Salesforce, which, as you can imagine, when you move from a system like that to Salesforce, it's rocky with You know, the duplication of deals, the account ownership, the lead routing, all of that stuff. So that's the biggest thing that we're spending time on. We have no issues on deal velocity, closing deals. Like I don't have a close plan issue. I don't have an issue of like, oh my God, this deal's been sitting stagnant for, you know, 2 months. What the hell are you doing? I don't have that issue that my many revenue leaders have had in the past. I don't have to deal with that. I need to go deal with making sure I hire fast enough, making sure I enable these people to be good at their jobs, and making sure that the actual foundation and the systems are kind of working in the right way to make sure that we're not missing a beat there.

I have the same opinion. Like we're not gonna go vibe code Salesforce if there's, if you're being choiceful about buy versus build, it seems like those core systems of record that manage governance and ecosystem and security, those are things that are best bought. And I think the announcements that Salesforce has made around this like headless architecture, or I don't know if they're calling it headless or Salesforce 360, whatever they're saying, seems, it seems like the direction you guys are going in as well. So like, what do you think is the role of CRM in this whole thing?

I mean, you still need a place to house your customer information, right? So that doesn't change whether you use Salesforce or some of the other tools out there. That's kind of up to every organization. So I think you still need that, but then making it your own, interacting with it in a way that's kind of useful for you and your business in a day-to-day. That's where I think it's changed from how it used to be, right? Where you would just go get a tool for whatever you needed on top of that. Like you, they were even like RevOps tools that were there that would help you with annual planning and forecasting. I tried some of them out and never ended up using them. You know, then you have like forecasting tools, customer health tools. So I think that's the world that's evolving right now is you need the systems of record for sure, depending on whatever it is for your business, but then making it your own and the flexibility to extend that and the extensibility of it. That's where like solution like ours can really help kind of play that game because we're fast, we're quick, we're pennies on the dollar in what you would actually do. Like it, like, you know, I think our RepliBook example that I mentioned to you and again, tune into SaaStr where I'll probably be giving a demo of that. I think it took our, our field engineer is the one who built that. I think it took him less than a day, less than a day to probably put that together, framework. And we continue iterating and doing that, you know, modifying it as we continue growing. So I think that's where, that's where it is right now. But, you know, the other, just on this topic, I do have several customer examples, right? Um, specifically SMB customers where I've seen SMB customers like smaller startups and like small organizations, ViP code some pretty cool stuff. I've seen them actually ViP code CRM applications. I've seen them ViP code like another HCM type solution. They've actually done employee onboarding, um, ViP coded. Um, so I've seen that as well. And, you know, upmarket, obviously I don't see that. Like a GPMC is not going to go get rid of their ERP and their Salesforce instance, but I've seen the extensibility even happening upmarket.

What is that mental model for you about like what are the things, you know, for a company of meaningful size, like 50, 100, multiple hundred millions in ARR, what should people think about not building themselves? And like what is the decision?

Yeah, I mean, if you're a big enterprise organization like that, you're not building your own CRM, you know, you're not building your own ERP, you're not building.

And so like, what— why though? Is, is it because like, it's— it is— it, is it because the governance and secur— or security? Is it like the stability? Like, what would be the— what would be—

I would say, I, I would say in an organization that's as large as that, you have so many touch points with billing, finance, accounting, true-ups, delinquencies, you know, the whole integration to your back office, to front office is pretty dialed in. And if you, you know, like it, it's just, it's just really, really tight. And to go do that in a vibe-coded fashion and connect all the— could you do it? Yes. Would you want to go invest the time in doing that and re, you know, basically get, getting rid of your plumbing and getting a vibe-coding solution to reinstitute a whole new plumbing, I think that is a huge uplift for any organization, any enterprise organization, right? That has been around for decades and decades to go do that.

So like integrations and sort of like the centrality of the tool, I think that's a good heuristic. Governance and security, I would think is another good heuristic. Like if part of that platform's value is security, stability, governance, like buy versus build is probably smart. Any other like key heuristics? 'Cause what I'm, what I'm trying to, I get asked this question so much and, and I'm trying to think of the model and actually I might, I'm, I'm planning to talk about this at SaaStr as well and I'll, I'll drop my like SaaStr link in the show notes 'cause I'm gonna come see yours and everybody should come see your, uh, your demo too. It's like at what point down the stack, like should we vibe code sales loft? Like it's a, it's something that we've talked about a whole bunch. You know, I don't think we would vibe code a calling tool. Like you could vibe code your own thing and just API into Twilio, but like, what's sort of the point? If that goes down, it like your whole team is offline. And so the, like, I'm trying to come up with the heuristic of like how far down the stack and like, what is the trade-off question? Like if this, then that of like where you stop. Building, like we're gonna, we're building our LMS, like we've rebuilt a learning management system. We've, we've built like a recognition and a contest system. It's because they're not that connected. They're not that central and important. Stability, if it goes down for a couple hours, like you're not totally losing your mind if your contest tool, but any other heuristics that you think are helpful?

I think you have to have a discussion as an organization and as a group. To actually look at the feasibility, right? And just have some common sense behind it. The way I think about it is just have some common sense of like stuff, you know, that you're going to go do. Like, we're not going to go build an outbound dialer. Like, I'm not going to go build a dialer, right? Like, there's other things that are like more important for me to go do. Could I? Sure, I probably could. Like, but am I going to go do that? I'm not going to go do that, right? So yeah, I think every organization is different. I would just say, just sit down and have an open discussion.

I think the cost benefit is the other thing there. Like, like what is the incremental benefit above the quality of the tool today that you unlock by doing it yourself? And like, that might be a determining factor as well. Like, you know, vibe coding your own dialer, like how much better are you gonna make a dialer?

Yeah, exactly. Exactly. Exactly. Really?

So that might be the other thing. Yeah. Um, okay. It's interesting. It is interesting that a company like Replit is like replatforming to Salesforce. I think it says a lot and it sort of confirms my prior a little bit. And, and you guys are ahead on this like headless architecture of saying, all right, you're actually not going to live in the tool that much, but it is this like hub and connection point where everything like builds off of.

Well, yeah. And like it was a Salesforce, actually, We had a call with Salesforce just this past week where they showed us kind of their stack and we've kind of showed them our stack of how we're extending Salesforce just from a visibility standpoint. We even, we just came up with a Slack connector as well with Replit. So now within Slack, you can actually spin off a workflow to build an app all within Slack, you know, directly within Slack. So again, the, so we're all about the extensibility, right? How can we, use a technology like Replit to go extend and kind of just make things your own and do it in a way that serves you, we kind of show up every single day, right? I live in Slack, right? So if I can go spin off a workflow in Slack versus going to Replit specifically, right? That saves me time, right? That saves me, right? So it's, again, use cases are kind of all over the place.

Hiring at this pace is extraordinarily difficult and it's hard to keep, keep a bar. So how are you thinking about, and especially like if you want to hire 200 people at some point, you're not going to be able to do the screens. And, and so like, how are you thinking about handing off the playbook to, to somebody else who's going to be new to the business? Like new VPs, new GMs. I, I at least have a bunch of people that have been with me 3 years now as we're, we're probably going to hire like 100 people in sales this year. But, you know, there's so much, there's so much more institutional knowledge. It's a little different going from like tiny org to what you're trying to do. So how are you thinking about the hiring system first? And then I'd love to unpack the profile.

Yeah, I mean, so the system in itself, we actually have a fairly tight hiring process that I actually designed and I kind of created just to kind of optimize for speed and making sure we're also not sacrificing on quality and the people we're actually bringing in. So our hiring process is like very simple right now. Like, as I mentioned, I do the first screens and I was doing 30-minute screens. I actually changed from that to 15-minute screens. So I do 15-minute screens because I've got a good read on people and I can usually tell in the first 15 minutes whether they're worth a another conversation or not. So I do 15-minute screens with everybody and I, I'll talk a bit about how I'm gonna change that. And then usually if they pass the 15-minute screen, they get invited to a mock panel. And our mock panel is we give them a prompt, um, a, a very, very specific prompt depending on their role. And then we usually have 3 people in the panel, usually be like a rep, a customer advocate, or a field engineer. We kind of change it up all the time and the candidate will present to the panel against their prompt. And coming out of that, we make a higher, no higher decision. I eventually am going to get out of doing the first screen. So I'm like really excited to have my head of sales starting here in a couple of weeks and they'll kind of carry that torch. And my head of Europe who's going to start, it will carry that torch internationally as well. I'll still be there to help close candidates and do whatever it is that I need to go do to bring people out there. But I don't think it will change the process in itself because it's pretty quick and dirty right now. Our biggest thing is making sure we go back to the first principles of like some of the characteristics that we actually want in the people that are coming here. And that's really what the prompt really helps us validate is, you know, outside of their customer-facing skills. You know, I talked about their knowledge of the AI ecosystem, their knowledge on Replit, their knowledge on the art of the possible, you know, how they can go articulate that. Obviously, if they're technical, we actually give them a technical exercise to go do. So we just make sure we're good there. And the other thing I'm also, given how fast I have to go hire, I make this call right in the beginning, is we want an 80% bar. So we're— we are not striving for 100% or like if they don't meet this 100% bar, we're not. If it's 80% and we think somebody can kind of meet that 80% bar and we know that with the right level of enablement, we can kind of get them that last 20%, get them to 100%, we're actually optimizing and pulling the trigger on that person just because we have to go higher super fast. Because if I take like a fine-toothed laser focus of like, oh my God, they must go meet, I just never can hire as fast as I need to go higher right now. So that's kind of what we're doing. And then I'm spending, like I mentioned, a lot of time on enablement because I think that's going to be a huge differentiator for us as well.

And so that last 20% is either like the product knowledge, the sales craft, and you feel like you can just like, you can fill those gaps with your system. Exactly.

Yeah. It's like someone may be lacking on, they may be in the prompt, they were really, really good at the art of the possible and the product, but maybe they didn't do as well on discovery in that prompt. Right. And I think I can teach them how to dig in deeper and go do discovery, you know, in a much more diligent way. Or yes, the sales aspect or whatever it ends up being. Right. We want people that can actually show up, get in front of a customer. And we know that we can actually trust them to go out and explain our value proposition.

What's most important for you in that profile? So the, the mission centricity, You want to see that they understand the ecosystem, the tools. What are the, like, the raw traits? Because it sounds like you're making, like, traits-based decisions and then teaching a lot of the craft and specific knowledge.

You know, like, in the 15 minutes I get with candidates, what I want to be convinced of is their communication style and skills. How are they communicating with me? I do ask about kind of their aspirations of why sales, if they haven't been in sales and what, you know, why they want to do this role. And then I take the back half of the conversation and I talk about, I go basically, hey, if you look at the AI world that we're living in, stack rank yourself around your knowledge of AI between a 0 to a 10, kind of everything that's going on with the AI world from inference to LLMs to app creation. Like, where would you stack rank yourself from your knowledge? And then what's your awareness of kind of where Replit kind of fits into that ecosystem? How they respond to that. And then based on that, that kind of gives me either confidence or not confidence that, you know, this person can actually be effective in the role that we're actually interviewing them.

What's the most important thing for you in that 15 minutes?

Look for communication style, because if they can't communicate well with me, how are they going to communicate What with the customer, you know, the people that are either nervous or, you know, kind of struggling with that, you can kind of weed that out pretty quickly.

Yeah. Interesting. So what, something I built a few weekends ago was the, uh, like a deep analysis of our entire XDR hiring process. And so I pulled in all of the scorecards from Ashby, all of the meeting recordings from MetaView. All of the in-role performance from our BI tool and had— using Claude— had Claude do this big analysis of like, what are the common traits? How are like, who's good at identifying talent and who's not? What are the things like, where are we most off? Because what I'm trying to build as an output, which I have it as an output, I'm still like iterating on my level of confidence, is a copilot tool that can take in every single interview afterwards. And give us, give a synthesis on their likelihood of success. If nothing else, it's been a really interesting gut check. So I'm the finalist. I'm the bar raiser interview. I'm the last sort of 30-minute interview in all of the market-facing roles. And so I'll do the interview, then I'll just go to my terminal tab and be like, hey, run it on XYZ person. It yanks in all of the information. It gives me this output and creates a Notion page. And it's really interesting to see like the, the distillation of all of the interviews, cuz I don't read the other interview notes. I don't watch any of the other calls, but I get this distillation as a second, as a second guess. And so far I've done like maybe 5 Bar Raisers and it's been aligned with all of the recommendations, but like as I add more and more data to it, I'm, I'm, and my confidence level grows. Like I want to see if that score, the AI score is going to beat or improve on the human scores. The interesting thing is like I was churning and churning and churning on it and I was like, ah, I don't know. Like I just was a little, I was second guessing the analysis and then I, then I put in that Andrej Karpathy auto, auto research. GitHub repo. I'm like, hey, like, do the— because I have the outputs, output metrics, I have all the input data on the interviews, and I have the output, uh, data on, uh, performance, in-role performance. So I was like, oh, that's a good auto research opportunity. And then it ran for, I don't know, like 45 minutes, and it was way better after then. Yeah, I'd love to see that.

I'd love to see that. That's, that sounds pretty cool.

That, I think, is like one of the really interesting last frontiers. I've met some founders that are building in this category, but it was always like a little thin and a little like AI slop feeling.

Yeah. Yeah. Yeah.

You can't get confidence without like a massively deep understanding of, of like the in-role performance and this scoring. And so we, I ended up building it. I think it's, I mean, I think it's usable data now. but I, I needed to, I need to see how these scores translate to enroll performance over the couple years.

Yeah, that's awesome.

You don't need to do the 15 minutes. You can just take the whole output and, and have your brain in it, like layered over the whole interview process. And may— maybe other than like helping close candidates, like you don't need to be in it.

Yeah, that'd be pretty cool. That'd be pretty cool. You've been in this business now, uh, when, when did you start Replit? I started officially in December, but I was actually with the business early on and like last summer, I'm doing some stuff kind of behind the scenes.

What have been the biggest surprises and what do you find yourself changing your mind on as, as you're in this like fully agentic business that where you can build anything and do anything?

Surprise it was, I knew the GTM, the B2B GTM org was small. But I was surprised to see the lack of a lot of common infrastructure that was, that, you know, that I thought was already here that wasn't here. So that was surprising. That's fine. I mean, it's, you know, we just build and we grow. So that was definitely surprising. The other aspect of it is, you know, given how many users Replit actually has, I was surprised that from a headcount perspective, how concentrated we were in the US and didn't really have dedicated headcount globally across the company, right? We do, but not to the level that I would've thought. You know, like everything else around the product that I kind of knew coming in, there's this cult following, right? There's this cult following of Replit where, I mean, I cannot tell you the amount of times I'm in an airport with like either a Replit jacket or a Replit shirt, people, random people just go, oh my God, I love Replit. I'm like, you know, like, you know, there's not a lot of brands where you get that, where people see it out there. So that's like super, you know, you just impressive around like how far we've gone and how far we've kind of built the community.

So there's so much conversation like, oh, like, you know, sales, sales roles are going to get automated or BDR is dead, yada yada. You know, you've talked about automating inbound, but growing your outbound function, how do you think the shape of go-to-market teams are going to change? Like, and how are you thinking about building yours different?

So I, so there are certain roles I do think will change. And by the way, I've tried, like I tried the automated SDR, uh, last company and, you know, it didn't really work for us. And I've talked to a lot of other peers where it doesn't really work. So I don't think the automation of BDRs completely is going to go away anytime soon. I think inbound is a different thing because in the thing about the inbound motion, the inbound motion is taking somebody who's interested in seeing your product and getting them to go talk to somebody who can show them the product, right? That workflow I do think can be automated in the right way if you actually ask the right upfront questions in a forum and with an agent. So anyways, that, that's the model we're working towards. So like quotas and how we comp, I think there's going to be like different flavors of kind of tweaking that model in the old world, in the new world. And then I'm also dabbling with this experiment, which is going to be coming online here very soon, is I'm going to give for the very first time our customers to actually go buy Replit Enterprise in a self-service fashion. So, so today if you go to our website, you can sign up for our free tier, our core plan, and kind of what we have is our, call our pro plan. And that's where the majority of our usage is right now. You know, is these hobbyists, students, average person who actually wants Repli.ai, they just go sign up for one of our B2C self-service plans. Yeah. But I'm going to have, we're going to run it as an experiment initially, but we're actually going to have a flavor Basically not a flavor, basically our enterprise version available as a self-service option. And what I think that's going to drive, and my hypothesis behind that, is I think there's a segment of the B2B enterprise segment that is small enough that can really just go self-help themselves without talking to somebody else. For sure. Yeah. Right? And basically get the features of that they actually go want and, you know, small ECB, small deals. So the idea would be is like all the small ECBs, like say sub $20K, $25K, those are all just self-service. Then I've got my sales team kind of working on the higher construct of deals. And what that will probably do is that will change the makeup of my commercial team, right? Maybe I have less commercial sellers and I have more enterprise sellers as an example. I don't know. So that, that's one experiment that we're kind of going to be running with and it's going to be coming online here very, very soon. The other way is like the way we're building our RevOps team. Is totally different, you know, like the old world where, you know, you scale RevOps with the number of salespeople you have. We're not thinking about it that way. We're going to be opening up a role that is the classic role that everyone loves to hear about, which is an AI GTM engineer. We're going to have one too that sits within the RevOps team that kind of helps with everything around automation and what we want to go build. And then, you know, using a lot of the stuff we're already doing, like, you know, our own CPQ, like, you know, DealDesk analysis and things like that to just run a leaner team in a way that we can go support it. So that's definitely on the horizon.

So what sits under RevOps today? Like, because there's RevOps, there's GTM engineering, there's like applied AI, which you could say is different. There's data, there's BizOps, like those are all functions that overlap so much now in, in so many ways. So how are you thinking about that architecture?

Yeah. So, so in, in our world, um, it's very simple. So within RevOps, we have systems, we have a deal desk, and then we have kind of what we are calling like a generalist ops role, right? Where it's, that role would serve sales. Post-sales partners, right? And then we are thinking about this layer of like an AI engineer, and that's kind of it. Data doesn't sit in our world. The data actually sits in the product team right now. And then we have a BizOps function that is kind of like a BizOps GTM liaison that sits in finance.

And so when it comes to like lead scoring, so figuring out what leads to route to what people, is that in the data team or is that in systems?

It's in— it's a hybrid right now. So that's something we're actually working on right now, like lead scoring, the whole PQL intelligence. So I think that'll touch multiple people, right? That'll touch data, it'll touch marketing, and it'll touch systems as well as our BDR team, right? So it'll touch a bunch of different people there.

And that type of initiative is run out of the data team that's under product?

No, like the whole lead routing thing We in GTM, in collaboration with marketing, would kind of run that initiative out.

It's funny how much these, how much overlap there is between these roles. Oh yeah. How fuzzy it gets. Cause especially now that the technical barriers are so low, you know, like you don't really need it to call in a technical person from some other team to build something. So like who owns that today is a big question. Do you see the core role of a sales rep changing? Like, Do you think that AI will fundamentally change how things are bought? Like there's a big question around like, will agents just buy from agents? I don't think so.

No, I don't think so. I, again, like the whole customers doing their own research, making their own decisions and procuring something on their own, like kind of this self-service motion. Yeah, I do think that is here to stay for sure. Right. And there'll definitely be vehicles for that to happen. I, I do think that, but I think, you know, the consultative conversation, the relationship building, all of the things it takes, like on classic enterprise, to negotiate complex deals with complex regulated customers, you know, in FinServ and CPG and healthcare, that's not going to be automated with an agent. The salesperson's job is changing, right? Like the amount of research and information gather. I mean, that's all automated. So I think they can now spend their time in front of customers where they should be. Versus doing the back office crap. You know, the, the, the teams we've managed in our previous lives, like Salesforce hygiene and people spending more than half their week updating Salesforce. And I mean, those days are, those days are gone, right? Those days are gone. Like you can utilize a lot of automation to get and do all of that for you if you do it properly.

It already seems crazy that we did it.

Yeah.

Like it's, It's funny how much things change in such a short amount of time. Like it lit— like it seems insane now to think about how much time people spent building a lead list to call and updating their next steps and moving stuff around in a dashboard. It's crazy. It's a really crazy concept. Okay. I want to hop into the quickfire. Before we wrap. So, um, what do you think separates a good CRO from a truly great one?

Customer-first mindset, right? Um, customer-facing, lead by example. Don't ask people to go do something you're not going to go do yourself. As a CRO, you're not just running sales. You're managing the entire lifecycle from start all the way to finish of how people get in the door and how they kind of hear about your company from the very first gate and then how they kind of get value out of your product. So looking at that entire landscape and just having a viewpoint on it and being front and center, I think that, that's what makes the difference.

Yeah. What's the most common advice you give to like newer leaders? First time CRO, first time VP of sales, big mandate. What's the advice you give to those folks?

So the ability to scale down irrespective of your title. And irrespective of your remit is extremely important. Especially when I hire leaders, like the two leaders that I hired, that was a very direct conversation with them is, yes, you're going to have a big team and you're going to have layers of management eventually as well. But when you first come in, you got to get your hands dirty. You got to get your hands dirty and you got to get in the weeds and you need to kind of understand how things are kind of operating there. So that's The piece of advice I give all the leaders coming into my business right now, going out and doing deals, you know, getting in the trenches with the reps, that is like table stakes. And we ruled out a lot of candidates in this search that were very credible executives, right? Very credible seasoned executives, but I, you know, questioned whether they could scale down.

I think that it's way more important in this AI world where things are changing so fast. Like you can't rely on working a couple deals when you were a manager 18 months ago to apply today just because everything's different.

Everything's different. Yeah. Everything's totally different. It's the way we price, the way customers buy, how they evaluate. I mean, it's totally different.

What's the hardest lesson you've had to learn in your career?

God, hardest lesson. I mean, the hardest lesson, I mean, I had one company where it was my one and only, from a career perspective, the wrong decision was vetting the company, vetting the founder, vetting the CEO, doing it in a diligent way, right? I think that is just like how people interview us, we gotta go do interview them. And do the back channels on them because these are the people you're gonna be working with alongside, right? For a very, very long period of time. I had one, like I said, misstep in my 30 years and I think I can take a single bogey, which is fine. But that was a learning lesson for me. Yeah. I'm making sure that I double-click just as much as people are double-clicking on me. I'm double-clicking on them before I actually make a decision.

You only need to learn that lesson once. That's right. Yeah.

One time only.

Yeah. And then you start to do a different level of diligence. Last question. What's the best thing you've read in the last 12 months?

I'm an avid pilot in my personal life. I've been flying for 26 years as a hobby. And I do a lot of nonprofit flying these days where I do with Angel Flight West. But the biggest thing I've reading right now is the hourly cost of operating planes and whether I'm better off in flying out of my co-op where I share my plane or whether I'm better off having my own and what the math and the unit economics tell me. And there's a whole process and a book and a methodology that actually allows you to go do that because it's, they call it the total weighted hourly cost of flying. And that includes everything from fuel to hangar costs to pilots and whatever. So anyways, I'm reading a bunch of books that are online publications to help me decipher that in my—

Very cool. That is a very niche answer and one that I probably won't hear again. So that's good. Ghazi, this was a ton of fun. It's really cool to hear about hearing the journey you're on and just like, it's so unique in so many ways. So it's fun, fun to unpack it and hear some of the very like different approaches that, that you're taking. And this was a blast and I appreciate you joining.

Yeah, no, I appreciate you. Thanks for having me on.

Thank you for listening to the Revenue Leadership Podcast. If you enjoyed it, don't forget to subscribe and you can find a link in the show notes and be sure to leave a 5-star review. Share it with your network, and please join me next Wednesday for another great conversation.