The AI OS for Revenue Teams — Amit Bendov & Seth Marrs

You go first, Seth.

Yeah, I think this is the year we start to... I agree with Seth by the way. It's a big focus for us, but also I would add like we start to see something that looks like an autonomous system that runs operations.

Well, that reminded me of that. I almost forgot. So, I was a poor penniless computer science student back in the days and I had a job as a cashier at a parking garage in a basement. And then a friend told me, 'Oh, I'm doing this training job. They're preparing people for SATs, kind of things.' Which I liked. And so this is kind of cool. I was working minimum wage and this was paying higher and I was working 3 hours a week, so it was pretty cool. And you work with very smart people doing puzzles all day. I don't think that had a lot of impact, but in the summer when people were not studying, I was doing cold calling. I was doing the same company, calling education program. And if you want a hard sale, call on school master, try to sell them something to improve people's brain. That was a pretty good experience. I was an SDR back in the days.

Yeah, so let's fast forward a little bit. So I think it was 2021, you guys did an event at the Chase Arena where the Golden State Warriors won. It was really cool the way you had it set up, you like half the arena and then afterwards everyone could go and play basketball on the court. And I remember sitting in the stands watching it and you had someone on stage and they were talking, they're like, 'Hey, you know, I use Gong, I love Gong, and I listen to calls on the way to work.' And my heart just sunk because as an analyst I was covering conversation intelligence, fledgling category, and I saw all the features you had. And I'm trying, I want more vendors to level it up to make the category really special. So, I was waiting for something special to come out of it where you can talk about all the other cool features you had nobody else had to try to spur the market forward. And no one else probably noticed it, but for me it was like, 'Oh, I wish we could see more.' But you've kind of since then, you seem to have really embraced what you're different at. That you being Gong being different has become that advantage. And I've watched it over the years just become more and more pronounced even in your product releases. When you release a product, either enablement, forecasting or cadences, you don't release it as a me too. You release it and your focus isn't even really on the base. It's on like, 'Here's these cool things that we can do nobody else can.' You talk us through that journey around how you think about those things when you're releasing these and trying to bring these products to market because I've seen the evolution over time.

Very interesting. You know, Gong was never about the calls or listening to calls or recording calls. It was about the data. It was about the data. So, I actually personally, I'm an undiagnosed ADD. So, I hate listening. I like everything needs to be in 30 seconds. At my previous company before Gong, we were recording all calls, but I just didn't want to listen. I just wanted the data: what's going on, what's trending, why are we losing deals, why are some people winning? And the recording we kind of had to do because there's no other way to capture the information. But we didn't even plan a feature to let people listen to calls. We just wanted to show the executives some dashboards with insights. But we saw people really love it. So we started developing. We capture the data, that is the foundation, and people enjoy it and they use it and prepare for meetings and now they do a lot of other things. But the vision for Gong was an autonomous system that runs the revenue operations. We realized that CRM does not have enough data. It tracks customers but not the revenue, not what's happening with the interaction. And there's a way better data source in conversations, in emails, in calls, in videos. And if we could use AI, and that was in 2015, not today's AI, machine learning predictive models, if we could capture all this data and turn it into both insights and actions, we're on to something good. So that was the idea. That's why we did not move on. This is so fundamental. Gong's strength is in the data that we get. We get whatever data we can get from CRM: who are the customers, their accounts, and a lot of other things. And then we add third party data and data from the internet. And today people understand this is the context for AI. Without data there's no AI. You can have the best models. So that gives us the advantage. If you do forecasting, it's not based on just what people's sentiment is, it's what customers actually say or don't say. If you do engagement, it's how customers respond. If you do enablement, it's based on real customer conversation, not hypothetical product marketing scenarios. So that's a real strength and that's why we're doubling down on everything.

It's psychology sometimes called denial. You downplay something because you don't think it's... It's very deceptive because it seems easy. 'Hey, I could probably crank up a call recorder, two people in a garage for two days.' Okay, that's one thing. You can record calls. But if you think about the complexity of building something like that, the system is really three layers. There is the data which is the revenue graph, the insight and AI layer, and then there's the application layer and orchestration layer. These are the three layers. So anyone who would want to build something like Gong would need first to create the revenue graph, which means mashing all the data together, matching the right activities to the right accounts and contacts, which in itself is a really hard problem. There are many opportunities, many accounts, it's a complex relationship. Just match that together at scale for large organizations. Now legal compliance kicks in. 'Did you get permission to capture this?' And prove it to me. And access control. You don't want people... you need access to the emails of the company. People could be searching for a job, there could be lots of other things. I could be messaging my wife. So who has access to what? You need legal compliance, security, access control, international legal compliance. We deploy in 120 countries. So first you have to get that right and accurate. If you don't get that right, there's no AI. Now if you built that, you need to build the insight layer. Even now, people say ChatGPT could do anything. No, it can't. Since 2023, Gong is accelerating every quarter. So if I had a penny every time I hear 'Gong is going to die'... because it's the fastest growing company right now. Since the introduction of generative AI, the insight layer uses LLMs for some tasks, but there's a lot of predictive models which LLMs don't do. There's a lot of small language models that are faster and more accurate for a lot of things you have to get absolutely right. You can't have hallucinations. So that's a pretty substantial layer. And then you have the need to have the applications. If you build the first two, you need to build the application tailored like a glove. Sellers love Gong. The user interface is easy. It's tailored like a glove. So there's a lot of work. It's not something you do overnight. It's a lot of work. And we've been building it for 10 years and continue to grow.

Sorry, Seth. I still like you, though.

That is absolutely the time. And again, it's not a new idea. I have sketches from 2017. The thought was earlier, but the earliest I could find is early 2017. We showed kind of the dashboard of a seller, the cockpit, where they have 'Here's my goal, here's what I need to do today, here's advice.' Even before co-pilot was a thing, we had a co-pilot idea. The technology didn't exist. But we focused on building it from the ground up. Remember the three layers: the revenue graph, the insights, and the applications. And this year we're building the cockpit for the sellers, also for their managers, also for enablement, also for revenue operations, and for CEOs. The primary interface... we always had the vision that revenue AI becomes the centerpiece of the revenue operations, collecting the data and providing... but it's a counterintuitive path because you took the insights path and now almost last you're doing the 'this is the place of work.' Most took the path to 'I want to consume the place of work' and then insights are commoditized. But you're going the other way around.

Yeah. Our goal isn't to have people spend more time. There are a lot of tools that I use where I can see how the product manager is being measured because they say 'There's something waiting for you' but to know it you have to click. And I know that they want to get that page view. So it's not our mission. Our mission is really to help sellers and to help companies sell more, be more successful. That's the mission of the company. It's not to get page views. Now realistically, pragmatically, today Gong is most of the information for them. Our DAU/MAU ratio is one of the best out there and it's increasing. The more AI that we add, the better it is. But if in 10 years people don't log in at all, I don't see it as a sad thing. If the work gets done, I envision a system that is omnipresent, that oversees what is going on, listens, reads, needs, looks behind the shoulders, and does its things without me having to go. So I think that portals are kind of like they belong to the '90s. In the first few years, for sure, this is the... we built what we think is a better home for people to spend their time than running through a bunch of other tools. But again, it's not the mission of the company to count logins.