How Gong Built a $7B+ Company in the Age of Vibe Coding

In this interview, Bendov reveals how a bad quarter at Sisense exposed that only 1% of customer information ever reaches the CRM, why Gong stepped on the gas in 2023 while its competitors were exhausted, what the "local optimum" trap is and how founders get stuck in it, and why he thinks the "SaaSpocalypse" panic is way overblown.

Hi, I'm Amit Bendov, co-founder and CEO of Gong. Gong is the leader in revenue AI, which helps companies maximize their revenue potential. We saw an opportunity to automate revenue with AI. That was in 2016, before AI was all the rage, and today we're one of the fastest-growing companies. Our growth rate is over 50% year over year. We're actually growing faster than the previous quarter for almost 12 consecutive quarters now. So, pretty incredible times.

I grew up in a very humble background. I mean, great family, loving family, but most of my family did not even go to college or university. So, from my family's perspective, if I were a seller at a retail store, it would be just fine. They'd love me still the same, because it's a job and that's what you do. So, my original dream as a teenager: I wanted to be a guitar player in a rock band. I was pretty good, to be honest, but not in the top 1%, and I was in a small country, Israel. Even if you're big, it's not all that big. That was a miss, but I started my career actually selling professional audio equipment at a retail store, like a Guitar Center. I was hoping to get close to the showbiz, but I realized, "Okay, maybe I don't want to play in weddings when I'm 40," to pay the bills. And I was really intrigued by computer science. So, I signed up. There was a fork in the road for me.

My first job was an engineer. I wrote code. Many people don't know that. Then I led an R&D team. Then I moved from Tel Aviv to the US to start my first company, ClickSoftware. It's now part of Salesforce. And I became the CEO of a company called Sisense, in business intelligence. We were growing pretty fast, and then all of a sudden, we had a quarter that wasn't great at all. I had no idea what was going on. We were analyzing a lot of data from the CRM that everybody says, "Oh, this is the way to manage customers. It's a single source of truth." There was nothing there. It actually relies on people telling it what happened, and that didn't make sense. Only 1% of the information ever actually makes its way to the CRM. It's all in people's heads. Now, what happens at companies: only half the salespeople are making their numbers. Nine out of 10 opportunities are not won without the company ever knowing why.

And I thought, "Wow, this just totally doesn't make sense." That was the year that AI was already beating the world champion in Go. You remember AlphaGo and DeepMind. And the thought was, "Can we create a system that will relieve people from the need to update CRM?" That was the idea. When I left Sisense, I thought, "Hey, maybe we could build a company around that." The idea was an autonomous system that automatically captures context from conversations and uses AI to turn it both into insight and action.

Are you using a meeting assistant today? We invented this. This was from my background in the music industry first. If we capture a meeting, meetings usually have multiple people, but we didn't want people to listen to 45-minute meetings. We wanted them to hear the interesting points out there. So, things like markers, topics. Today it's pretty standard, but that did not exist. When we first drew it, I plagiarized a little bit from SoundCloud: there were the play buttons, and you had those tracks. If you use GarageBand or something, you'll see there's a drum, a guitar, and a bass. So, if you look at Gong, you'll see those tracks. All the concepts, like the listen-to-talk ratio. It was the first application that measured how much you're listening versus talking, and how it tracks competitors. We invented this thing.

That was pretty innovative. And we went to raise money. I said, "Let's start a company." I was pretty sure it was going to be a walk in the park, but it was a far cry from that. It was very frustrating, and almost nobody wanted to invest in us. They obviously took a meeting, and another meeting, and what about this and what about that? People thought that sellers were going to hate Gong because they were going to see it as a big brother; that Google and Amazon were going to own this space; that the technology was not there. When we started, AI was nowhere near where it is right now. So, it was a hard job. We struggled to raise money, but finally we found a couple of investors who gave us the money, and that allowed us to start the company. And it was pretty smooth from that point on.

At Gong, we have our number one operating principle of raving fans. We're not happy with just happy customers.

Happy is like, "Yeah, I could be happy with a chair," right? That's not it. We want something people talk about, that is a level above. What if I sit on the chair, here in the kitchen, and all of a sudden it starts to give me a back rub, and it's just at the right temperature? What just happened here? You get people to talk about it, and it surprises you for the better. It does something that you don't expect. You've got to understand your mission: it's not a feature or AI or all of that. That's what people buy. That's what they ultimately want to accomplish.

First, we focus on value, not just features. So, we ask, "What does it do?" We have two things that Gong tries to do. First, 75% of a seller's time is non-productive: updating CRM, doing forecasts, doing training, sifting through a list. What can we remove from that?

Second, when they're talking with customers, how do we help them be better at their job?

So, these are the two North Star metrics. People move from somewhat reluctant to use it, "Ah, is it invasive?" to "Oh, wow, she just did that." So, you get that wow all the time, and the product surprises you in what it does. It needs to work well, and it needs to be above what you expect.

We did have a rough patch in 2023 because that was the post-COVID growth, right? The hangover. A lot of our customers had over-hired people and got into financial stress, so we saw higher churn. We also had competition that went in and bought a bunch of competing products, and they were going to our customers and consolidating. They said, "Oh, we're going to do all three or four products in one." The products were not very good, but at that time, customers were just looking at a spreadsheet and said, "Okay, yeah: one, two, three." And so, our growth slowed down substantially in '23. I wouldn't say existential, but it was definitely not fun days. We raised a lot of money, but we barely touched it. We just saved it for if and when we needed it. So, that gave us a lot of confidence and the ability to execute. And we knew that if we were struggling right now, all of our competition was struggling. If you're in an early market, which we are right now in AI, there are so many possibilities. Why would you try to one-up a competitor? Just try to go in a different direction; there are so many options. It could be way better.

I'm stealing from Jeff Bezos, who said that most people say they're obsessed with customers, but they're not. They're really competitor-driven. They look: "Oh, they did this. I'm going to do that, too. And I'm going to do it better." That's the wrong thing. So, what do you do? We stepped on the gas pedal. We knew that everybody was going to struggle. We had the cash, and we were going to invest in engineering. I'd much rather invest in people and engineering than in office space. I remember in '99, I was stuck with an office that was all steel, and plastic bags all over the area, and chairs, you know, millions of dollars.

It was giving goosebumps even to watch it and say, "Why did we do that?" We spent money on real estate and not on people. Is there anything that I could have done that would have changed the outcome? So, we only do two- to three-year leases. I think now, for the first time, we did a four-year lease on an office, but generally we've been at WeWork since day one, and we still are. Offices: either they're too small, you grow so fast that you need to move out, or if you don't grow, then you're stuck with a lot of space. So, we built two more products. Now we're at about four. Then we started consolidating, and now they're behind.

I'm an amateur cyclist. When you're climbing, everybody's exhausted; then you attack. You just give it all your energy. Psychologically, it breaks your opponents. If you're good, if you're able to accelerate, then everybody just gives up, because they know that they're barely breathing and now you're charging ahead. And that's exactly what we did.

Competition should be more like an Easter egg hunt than a football match, where you've got to match every play that the other opponent is doing. It's not a zero-sum game in the early stages. There are more out there. Just go and keep looking. Why go where all the other kids are going?

A lot of companies stall not because of external conditions. It's because maybe the market isn't large enough. In math, there is this thing called a local optimum: you're climbing a hill, and at first it looks very exciting, and you reach the top, but now you're not getting to the mountain.

The most important thing is the market, which market you're in, and which market you play in. That's more than anything else. If you get it right, then you could be okay on the others. There are two choices. There's joining an existing category. The advantage: it's already established, people are buying, and you just need to explain why you're better. There's already a budget for this. Starting categories is very compelling to people: "Oh, I'm going to start a whole new category of music or whatever." It's also hard because people don't have a budget, and they don't know how to think about it. So, there's a lot of education. And if you want to be precise, you can't start a category.

Categories are, by definition, a number of companies doing the same thing. If it's just you, you're not a category. You're a unique product. So, you start by creating a unique product. If you're not successful, then nothing happens. If you're successful, it's guaranteed that others will try to either copy you, do something better than you, or do something different. That's when you start to have a category. It does take time. Today, Gong is considered, or something like Gong, revenue AI; most companies say, "We need that." But it wasn't a must-have when we started.

People had to find a budget for it; it didn't exist. It wasn't like a CRM or a contact database. So, creating a category sounds very exciting, and I have no regrets. But if you're an entrepreneur, don't just think, "Oh, I need to start a new category." First, understand if you're stuck. Understand if it's intrinsic to your business or something that's happening right now. There are a lot of companies that say, for example, "Oh, we're developing solutions for SDRs." Don't do that.

If you live in the Bay Area, you think everybody has SDRs. Companies in Wisconsin don't. So, this is an example of a small market. You can get to 50 million, 100 million; you're not going to reach a billion. So, understand if you're stuck because you've reached the local optimum of your market. Also, if it's a small market and it's very competitive, that's the worst thing. It's like five hungry dogs fighting for a very small piece of meat. For us, we understood: "Listen, we have a trillion-dollar opportunity. If we don't succeed, it's not the market, it's us."

The noise that's happening right now with AI. I know it's very confusing. Just make a bet and keep going. Ignore the noise. Right now, people are very emotional. It's almost like when people fall in love with someone: it's more the hormones thinking than the brain. The dust will settle. At some point, it'll be clearer. So, if you have a direction, hopefully it's good, just keep going, keep plowing, do your own thing. Look a little bit at what's going on, but don't be enamored with the things around you. If I have to distill it to one thing, this is it: who do you cater to, and what do you offer?

Let's say it's a SaaS apocalypse. There's some truth in that, but it's a way overreaction of the market. Right now, investors are worried that if AI has this power of God, anybody can do software. So, what is the moat?

It's a natural question, but it's way exaggerated, because most companies could probably open-source their code, and it would still be difficult for someone to compete with them. The code is just a part. What about running it? There are security vulnerabilities that you need to patch in a package or a library. Who do you call at 3:00 a.m.? Just writing the code is one part of a software company, but actually running it, supporting production, and driving adoption is extremely hard. So, even if Salesforce's code were out there on the internet, it would still be extremely difficult for anybody to compete with them. CRM and HR are incredibly complex.

People just underestimate it. You can't vibe code it, even with 100 engineers. Otherwise, why are we not seeing one right now? Why are we not seeing a competitor? We have to understand that sales is an asymmetrical engagement. Buyers don't necessarily tell you everything. They might even lie to you: "I don't have a budget." Or, "I am the decision maker," even though you're not. "We're looking at a few of your competitors", they might not tell you what they are, and they might not be looking at anything. So, AI needs to have excellent context or data. When you negotiate, you want to lower the price; you want to increase the price. So, that's one thing.

Second, at least for us, we help mostly with B2B sales, which are complex. It's multiple people selling to multiple people. Understanding the power dynamics, the political power structure, all of that is something that AI will struggle with. Ambiguity is not where it's good. Reading the room, judgment, creativity. AI is very straightforward: it's trained on being plausible. Sometimes the creative moves aren't the ones that are plausible, it's doing something that seems unreasonable, but is still a winning play.

When you're alone, nobody cares. Once you're ahead, there's a crosshair on your back. Everybody wants to kill you, and we take it as a compliment. We've built the system over 10 years, and again, there's a lot of complexity.

Our vision was that we thought of it as a self-driving car. When we started Gong, we wanted to build an autonomous system. Our inspiration was a Tesla. A self-driving car has five levels: five is full self-driving, and one is getting lane-keeping alerts. And we thought, if you want to build a self-driving car, and this is a self-managing customer management system, then you need three things.

First, you need the sensors to capture all the data. In a car, it'll be the radars, the lidar, and the cameras. This is our revenue graph. Second, you need to make sense of all of it. You see a red light blinking: is it a traffic light, or is it a car or motorcycle? Is it accelerating or decelerating? This is the AI level. And then there is the application layer. So, we always thought about the five levels. And we said, "Okay, today, this is like 2017, we're at level one. And when can we get to level five?"

So, we built it even though something like a lidar didn't exist, or the AI wasn't strong enough to be fully self-driving. We built the framework so we can plug in the technologies as they mature. If you could vibe code it overnight, then we have a problem. But it's not easy. There is a lot of nuance and detail that goes into it, especially when deploying at a large scale.

So, I think it's an underestimation of what it takes to really run a company. All right, not investment advice, but I think it's a good buying opportunity for the companies that are real. From the outside, when you see a company growing, it looks like this. But from the inside, there are multiple revenue sources, and you need to layer them in a way that looks smooth from the outside, but each one is kind of S-shaped.

At some point, it'll hit the glass ceiling. You asked me about people getting stuck because you're doing one thing, and that's it. It's not growing as fast anymore. So, you always need to layer those things. You need to think about two or three years in advance, because when you realize that you need it, it's already too late.