Gong's Amit Bendov on Powering Your Company Brain

Well, I didn't say it's a lie. That's a bit harsh. I said it's not the source of truth for what's happening with customers. There are things that are true in CRM, like who bought from you, who your customers are, what they're paying you. That is all true. So it's not a lie. But your interesting question, I was asking like, well, why are we losing so many deals? Every company's conversion is like one out of five deals or one out of 10, whatever it is. What happens with out of five? And why are half the salespeople not making their numbers? Is it them? Is it us training? And the answer is, we're not a CRM. So I said people think it is a source of truth, so it's more accurately a source of a truth, but CRM depends on people telling it what happened, and they don't do it right because there's not enough time and they don't always want to do it and everybody hates doing it. So the idea was, can we retrieve that information from elsewhere, taking people out of the equation? That seemed crazy. When we announced Gong in like 2016, I said this is the biggest invention since the introduction of CRM 20 years ago, and everyone said, oh yeah right, there's another delusional founder. But now this doesn't seem crazy. People see that the majority of the information does not exist in a CRM.

Yeah. I personally have not logged into a CRM like in years. All I need to do is, hey, what's the status of Zapier right now? And it tells me everything. Not obviously you talk about the transcript, but the emails, all the interaction with the account, everything that's happening, it's a much richer data set.

Right. Is that a trend you're seeing consistently across your customer base? I would have to think so.

Well, what we're seeing, and that's a good thing, people look at the personal habits right now. Okay, we're going to automate meeting prep and here's my agenda. That's all good. But the powerful thing is what Gong has always been about, to get the insights on a global scale, to understand what are we doing right and what could we be doing better. I'll give you an example. At Gong right now, we actually don't have people writing battle cards and competitive information, even product documentation. It's all done by Gong and it's better and faster. So if you think about it, you just say, okay, create a battle card against Evil Corp, whoever your Evil Corp is. Before that, it was a product marketing team or enablement team spending weeks picking one competitor and writing what they've heard. And usually those are okay at best, and by the time they're out, nobody used them. They're in some content management system. Now, create a battle card. It'll sift through all the deals that we've won, not hypothetical, real stuff, all the deals that we have lost, and it'll create a three-pager. Here's who they are. Here's their strengths, truth. Here's their weaknesses. Here's why we lost to them. Here's why we won. Here's three customers that switched. Here's what they're saying they're seeing different. Exactly the information that salespeople want to see. It is better, faster, and cheaper. So the ability to get content is insane. People are creating marketing content from that. We have people that are automating entire onboarding. So if you think, all the information of the organization already exists in the collective brain and conversations, if you put it together, it's super powerful.

Exactly.

Right. The other is real intel, real time intel. So we launched a couple months ago, three months ago, a new product called Gong Enable. It's an agentic coach that coaches people. It is trending really well. If you look at the revenue numbers, they're great. It's our fastest growing product. So that's good signals, but I'm worried about the customer experience. Does it really work? Are there problems? It takes time for people to deploy. So I wanted to see how we are doing, what's the experience. So I immediately put an agent at Gong. It says, okay, tell me every time, every deal that we win, tell me was there a PC and what was the experience. And I can see in real time. Before that, I would have to go and start asking people and ask them to write reports. It is so easy right now. It's amazing that a new product launch can be tracked in real time and if there are issues, we can correct them in real time.

Yeah, it's definitely something that we encode internally to capture the organizational knowledge, but also something that we're building into the solution. It is shared with customers so they could get better and better over time.

Well, each team did their own. So definitely everything revenue and product is from Gong, but we also have other systems. We use it in legal, we use it in engineering. Obviously you build skills that have the coding conventions and the practices. So that builds the organizational brain.

Oh, no. Well, that's the way I understood it. That's the problem. No, use AI a lot. There's a lot of hype that people don't understand. Every day there's an, oh, I just vibecoded Monday.com as a weekend project. This is BS. Or even a year ago, people said, we hired our SDR and they even had a person named Fire SDR team and said this is BS. We knew it's not going to work.

And sure enough, it doesn't work.

That's exactly the problem. That's what we knew. I can tell you how Elon and I discussed this. Should we be in this? And we thought, this is not going to work, at least for now. It might work in the future. But we always tried at Gong, when Elon and I started the company, it was an AI vision. We wanted to build an autonomous revenue management system. This is like a self-driving Tesla, level five. But we also wanted to know, when we interviewed a lot of people from the industry, what works today. We wanted the state of the art, but not a vision or science fiction. We wanted to go with something that people can use today. So we've always been on this edge or frontier, something that actually works well today and can create value versus something that is going to be three years out that maybe would work one day. So we always try to ship something that people can really use. And now there's this expectation. So that's the hype. You know, I've just vibecoded something like a Workday. And a CEO a couple years ago said, oh, I dumped my CRM and we're going to develop it, which turned out not to be true.

Well, it's not centralized. We instilled it in the culture and the work practices. It's built by design of the company. There are two things. First, we have a research team. They're really thinking ahead, but it'll never be foundation level like quantum physics that might work on a quantum computer. We don't do that. We're not IBM. There's nothing wrong with that, it's just not us. So we think of things that might not work, but if they do work, they could add tremendous value to our solution. So we have a research team and that's what they do. They test concepts, but it's not a free check or free license to experiment with anything. These are things that we have a hunch might work and they're hard and they might fail. The other is we always test with customers very early. A product manager will talk with customers and say, we're thinking about this, and show mockups. Nothing really works, but first to see if there is value assuming this works. If it's remotely interesting, then they would try to put some data. Sometimes we decide to kill a project or change direction. What we thought is interesting is actually not so interesting. That's the hard part, to know what matters to customers and what can actually work. Then when we ship something, we try to see if it actually works. We know there's a cycle. We think about it, customers see the value, customers tell you that it's great, but then until they see it on their own data and say, why did you choose this deal? That doesn't make sense. Or this is bad data. That's the time. So we take those small steps that we can iterate really quickly and understand what actually works.

Well, it could be the right approach. What I said is not all things that are saying, oh, we're going to lay off half the company because of AI. That might be the truth, but maybe not the whole truth. The truth is a lot of companies overhired and sometimes find themselves with bloat, pardon me for using that term for dealing with people, and they want to streamline the organization and adjust it to reality, allow more agility. I think there is a real opportunity to rethink structures with AI. Maybe better span of control for managers, maybe less specialization. If five people each doing a step of the process, we could blend it all together. Everybody's doing everything with AI providing the skills and the knowledge, and people have more opportunities to do more things. So there is an opportunity to rethink structures and become more efficient. But a lot of these layoffs have to do more with, it could be because of AI because we need more budget for AI, so we're cutting elsewhere and we're going to redeploy that capital. But it's not necessarily that AI has already replaced half of my jobs and I'm replacing people.

Well, I think we're going to see people that are more multi-disciplined. They can help more. There was a time where there were only salespeople or account managers. That was the two roles you had, maybe one or two. Then 20 years ago or so started the whole idea of hypergrowth companies. The business was growing so fast you just couldn't hire fast enough. So we said, okay, we can't train a person on everything. So let's build almost like an assembly line. One person taking the inbound, the other one doing the initial meeting, the other one deploying, which created a reasonable fast growth model. But the problem it creates is a fragmented buyer experience. People don't like, I just told you what I bought, why I'm buying, and now I'm getting a different person. I want that trust, especially in enterprise. You want someone that has the relationship. It doesn't feel good that you switch people. And second, siloed flow of information. But now with AI, you could have one person do more things because the skills are now encoded in the system and it can walk someone with a brain and a desire and passion to do incredible things. So I do see that the roles start to merge back, and it creates a better buyer experience and more flexibility for the company. Also, when you create silos, it's very hard to balance. Sometimes you have too much of this and too little of that. There's always these mismatches. So I think it creates a better model for companies, unified.

Well, it's complex. There are multiple reasons. First, salespeople don't like to do prospecting. That's a stereotype, but if it's right, they like to get the deals. Give me the qualified ones. Who likes to pester people over cold calling? Some do, but it requires discipline. People are focusing on current pipeline versus developing pipeline. Companies had lead generation Tuesday. On Tuesday you only do this. So they said, let's hire some junior people who could hammer the phone all day. That started the compartmentalization process. It is easier. Ramp time for a salesperson to do the whole thing, cradle to grave, could take six to 12 months. But to train an SDR could be six weeks. So it allowed for better scaling and faster scaling of companies. But now when companies might not be growing so fast and AI presents the possibility, you can have people doing more things.

Imagine an assistant that always whispers in your ear, here's the next step, next step. And you need to apply judgment and taste and accountability. You have a person in the chair, almost like an autopilot, and you just need someone with the right brains and common sense. So you could hire people out of college that will do that. They share those traits if they're smart and hungry. Why not? And the system kind of knows what's going on. So definitely exciting possibilities.

Fortunately for us, first we were deemed dead in 2023. People said Gong is dead. We're now one of the fastest growing companies at our scale.

It was a rough year. You might remember the post-zero interest rate hangover. Our growth definitely slowed significantly. No more seats, everyone's not hiring, so no seats to sell. Everyone's firing. That was a tough period. And then ChatGPT came out. And everybody said, oh, now I can take the transcript and load it into ChatGPT and tell me how to win the deal. We've been accelerating every quarter now for the last 10 quarters since that moment. So it created a huge tailwind for us. Now every company needs AI, so they're calling us. Second, it made the product better. Things that we could only dream of in 2016 when we launched Gong, now they're a possibility. It's telling me how my product is progressing in a market. This is unthinkable 10 years ago. So we built the infrastructure both in terms of the company and the product. We always envisioned a system that has situational awareness, what's going on, the context today. Second, the insights layer. And third, the application or user interface. And we're just able to bolt on the new technology. We had machine learning, we had small language models before that. This is not new. Large language models are the new exciting thing that enables a lot of the growth right now. But what we have done right, not necessarily just because of AI, is to focus on customer outcomes. As we grow, we're not the only one in the field. We had one or two competitors, and now probably a hundred. And I'm sure in three years there'll be a thousand, which I love as long as we lead. So we need to always be ahead. But you can't get into all these features. You just say, there's one thing: Gong drives sales productivity, which means we help you win more deals and do less work. It's very simple. That's who has the best record, and we try to be the best version of that for our customers. Technology is a means to an end, but also the way we deploy it, support it, and partner with a customer is the key. In the early days, we just threw Gong at people and they would figure it out. That works to a certain extent, but to get to the next level, you really need to do a lot more.

Well, first, if you want to record calls, you don't need Gong. If that's all you want to do, you could get it for free on most video platforms. We knew it when we started the company. We didn't even plan to record anything. We saw ourselves as dealing with the data that recordings provide. But because people didn't have anything, we said, okay, we need to develop it. Gong will never take off, but one day it goes away, that's fine. But the real reason is, first, calls alone are not enough. We want to capture what's happening in an account or an opportunity. Most of our customers are in B2B, so it could be 10 people at Gong talking to 10 people at Amazon. It's a bunch of emails, in-person meetings, calls. Everything needs to be associated with the right account, and you need to create a complete picture. Emails contain about 40% of the signal. If you just listen to calls, you miss that. Second, you need to look at the whole thing. You need to capture everything, fanatically everything that happens with the account, because otherwise it's the one thing you miss that could have an email saying our CFO is putting the budget on hold. All your calls aren't going to help you. So you need to capture everything. That's a constant evolution for us as we had more and more signals. Second, it's one thing to capture my own meeting. You could use a tool like Granola. But to do it in an enterprise or even a midsize market at our size, you need compliance. Why is my call recorded? I didn't consent to that. Gong can prove back. You need access control, data retention policies, organization access control is massive. Privacy is a massive issue. For example, if I send an email as a CEO to one of my board members who's also on ServiceNow as a customer, and Gong captures that email, and that email says, I'm looking to replace one of my executives, can you imagine the disaster? Gong has fine-tuning of access of who can see what. That's the hard part. Capturing a meeting is super easy today. Capturing all the organization in ways that are compliant and can actually work and not miss a bit and not miss my most important call, and run at high availability, security, compliance, that's incredibly hard. We call it the revenue craft, associating the right opportunity with the right activities, controlled with workspaces.

It's very hard to run any organization beyond five people to do it in a way that is legal, secure, and works, not to mention scale and availability challenges. That's why we're doing well.

Absolutely. There's so much nuance because we have so much scar tissue from real customers, thousands of customers across all sorts of industries, to nail it. I don't even know as the founder of the company all the details that go into that now. There's so much. I used to know every nook and cranny in the product, and now it's impossible for any one person. You can't vibecode that because you don't even know what the spec is. You don't even know what to ask the system for. That's the hard part.

Well, people always asked us in the early days, why do I need my CRM? When are you going to kill CRM? I said we're probably not. First, it's hard. It's very easy to put the front end, but there's a lot of plumbing and messy stuff just like Gong that goes behind the scenes. Second, it's not very attractive or lucrative. Cloud was an IT revolution. Instead of using Siebel, I use Salesforce. So now I don't have version control and worry about installing and infrastructure. That's all true, but it is 6% of an organization's budget. Payroll for most companies is the larger thing. Our enemy is not the CRM. Our enemy is drudgery and ignorance. Drudgery is the amount of work that people are doing. 75% of a salesperson's time is not selling. If you measure, look at their calendar and see how many customer meetings they have. That might give you 10 hours. Let's say they have 45 hours a week. That shows you their productivity. A lot of it goes to internal meetings, updating systems, researching, whatever they do, it's not customers. So 75%. And half of them don't hit their numbers. If you look at the market size, there are roughly 10 million B2B sellers worldwide. Let's say they're costing 250,000 a year and selling 600,000 at quota. That's six trillion in revenue, two and a half trillion in potential. 75% of it goes to waste. Why would I try to replace a CRM that is hard? We outsource the drudgery, the work that they don't like doing, and the ignorance of why they're losing deals. That's a way bigger opportunity. Now the world sees, we said it 10 years ago, this will be bigger, and now the world recognizes this is not such a crazy idea. People talk about service companies, this is it.

Absolutely. There are certain tasks. I don't think we could 10x human productivity. There are areas where we could, depending on the nature of the task. Even we see it with our engineering. There are some tasks where you see very little, maybe a 10% improvement, because of the cognitive load on people. Versus stuff that are mundane, like update this import/export script that is very repeatable, there you could see 10x improvement. But if you look at the mix, it's hardly 10x. In salespeople, they're researching for a meeting. Sometimes I get pulled into a customer conversation. It used to be hours spent with the team writing a brief for me, probably having another session to go over it, prepare, and I can ask some clarifying questions, and then have the customer meeting. Versus all of that is gone. Five or six hours of very expensive people went down to zero. For reps, definitely meeting prep, writing emails. Gong crafts all my customer emails. I don't even bother to correct them. They're not 100% what I would write, but the corrections are not worth the effort of editing. So I just say, good enough, send. So if you look at the 75% again, measure how much customer time you have for your own salespeople, you'll be shocked. Most companies know it's very little. If you could cut it, I don't think you can reduce that by 100%. There's a limit to how much people can absorb. But if you could cut it by half, that's massive for companies. And if you could increase win rate by 5-10%, that's massive. Just think about it. You're talking massive ROI on a very small investment. You could do more than that. There are two levers for sales. For an engineer, it's how much and how quickly you run code, how much code. Again, it's a bit of a novelty metric, but let's say you look at that. For sales, there's how much customer time, and then the customer time you have, are you being effective? Are you saying the right things and driving the right behavior to win more deals? So these are the two levers, and you put them together, it's very substantial.

I always review them. I always review them, but again it takes a second. It's a quick read and good. I don't think AI is at a place where it could fully be trusted.

Yeah, because it only takes one crazy email that'll drive the wrong impact. So I think it's a good habit. Again, it's not negative in AI. It does an amazing work, but you should be reviewing that. I wouldn't let it send a proposal.

Well, I like LinkedIn. I use it a lot to get up to speed. I obviously use, I don't want to say which one because I'm going to get in trouble with either one of them, but obviously I use the chat applications.

Well, they all use Gong, so whatever I say is... But I do use them. Sometimes I switch between them. Let's say I run a dialogue with one of them. Now I want to get an independent opinion, not being aware of my context, ignoring it, and I want to get a truly independent one. So I do use them both, and it's great.

Right. The beautiful thing about them is that they have memory. So they might remember something I said three months ago, and it provides context for anything that we discuss. But sometimes I just want a fresh mind. So I use them for different things, and sometimes on a project I would switch between them.

Well, it was a fun conversation, and congrats on your success and growth. You're building an inspirational company.