Warm Introductions Built in Code

This is a working transcript with chapter timestamps. It may contain transcription errors. Private names have been removed; check the video before quoting a precise phrase.

00:00:00 Introduction

00:00:00 Awesome. Uh, my name is Jai Toor. I'm one of the co-founders of Deepline. Thank you all for coming up today. Uh, so we like to do these things pretty often, pretty straightforward. We're just going to do landing props. Walk through what people are working on so you can see what other people are doing, what you could potentially be doing, uh, and hopefully leave with something useful. Uh, so let's get into it. So our speakers today, we have, uh, Jen, Setareh, Brandon, Karim, Thomas, and Jack, uh, that's

00:00:34 a lot of speakers. So we're going to try to be a little quick. Uh, generally people are going to do a quick talk or demo, and then we'll save, uh, time at the end for one or two questions. And then if you have more questions, we'll try to keep them for the end. Uh, and generally folks are able to, uh, stick around for a bit. So we'll try to keep things moving, but hopefully get through everything. Uh, so before we go into that, uh, really quick, uh, again, who is Deepline?

00:01:00 Why do we host these types of events? Uh, Deepline is building composable tools for coding agents, uh, specifically for go-to-market type use cases. So if you're building internal tooling, if you're building systems where you're writing code for go-to-market tasks, uh, those are types of things that we're generally really good at helping with. Uh, we're very, like, SDK, CLI-native, uh, the goal is to make it as easy as possible

00:01:20 Why put GTM logic in code

00:01:25 to build bespoke go-to-market systems, uh, where you own the code, you own the logic, you're not putting it in a third-party platform. And so how we actually do that, what does that look like? It's integrations with both public and private data providers, like Parallel, uh, and then we give you infrastructure on which to build, uh, so that comes with, like, a workflow engine and database. So you've got to think about it, like, you go into Claude Code, you tell Claude Code what you're trying to accomplish, the outcome, uh, is what you're trying to build.

00:01:53 We work with every coding agent, [agent name unclear], whatever. And then when you get back to the coding agent, it's going to come up with an optimal solution. It's gonna try every possible solution to get to the best outcome for you. And so we've switched the problem from creating workflows to defining outcomes and letting the coding agent do the hard work. This is already what software engineers have been doing for like a year. We're just bringing it to go to market and you get access to a hundred plus integrations

00:02:22 with one API key. Everything is in code for you to review just like a traditional software engineer would. And so, I'll just do one quick example of what Deepline can do in text workflows. We have a couple of customers speaking to that, but I could go too deep into the weeds. But I think one interesting aspect of what I've taken away from our conversation with customers over the last couple of months is the point of everything that we're doing in Go Market Engineering or RevOps is to enable our sales teams to build relationships.

00:02:55 And when I think about what that actually looks like, it's a lot of operational work being taken off their plate, but there's also more we can do about what actually does a good relationship look like? How do you engage with your customers or potential customers more effectively?

00:03:11 Warm introduction demo

00:03:11 So a quick example here, I'm gonna show, so I took everyone who's attending tonight and I broke down, this is how an engineer would solve making friends. One, I broke down everybody by their jobs function. So we have around 101 go-to-market sales marketing, 34 GTM engineers with a specific title, and then engineering AI, 86 founders. All of the founders should try WorkOS. It's really good when you start.

00:03:48 And what we're trying to do here is try to figure out what does the room look like? And then I'm going to go down into, let's actually go to someone, who would be a good one? Let's actually just pick someone random. And then I'm going to try to find, so [attendee name omitted], and then I rank to everyone by the people in the room that they have the most in common with. And so this is the same thing I actually do for warm intro automation. Who has the most in common with a account

00:04:19 that we're trying to get an introduction to? Maybe they went to the same school, maybe they worked at the same job and the same function, or lived in the same city. So what we're doing here is they worked at, you can actually read on the side of the screen. [employer name omitted]. So they worked at the same company at the same time. That overlap may be useful context for an introduction. And this is how this is actually being populated

00:04:45 in the backend. We are taking the attendees from Luma that's getting typed in via web code. And then when people are getting added to the system, we reach them in real time using a Deepline play that's stored in a database. And then the scoring is a separate play, a separate workflow that happens afterwards. So at the end of it, right before I said, go score everyone, create a visualization. So this was all live coded like two hours ago. The takeaway is this probably would have taken weeks,

00:05:16 like a year ago. Figuring out what actually is relevant and which features we should care about. I did that by saying, hey, I know these two people know each other, go find the commonalities. I looked at our old warm intros and said,

00:05:30 Applying the workflow

00:05:30 how did the warm intros have worked? What were the commonalities? Almost always it was college and they worked together. So I skipped all the LinkedIn intro and are they connected in secondary connections? And I just said, who works at the same place and what's the sample? So if you are building a go-to-market solution, doing something this past, there's probably a tool that does this, right? But to do it for my exact use case would have been very difficult in the past. Today, this is pretty easy.

00:05:59 I can do it. I can do it custom for any other company and I can share the code that generated this and you can just work it and do exactly what you're doing. So it's not quite built in source, but it is pretty close. It's like open ecosystem, build what you want to build in the way that you want to build it. And we're just giving you the bits and pieces to make that really easy. Awesome. So I was considering just sending everyone an email with the introductions in here,

00:06:25 but I realized I should probably have people opt-in for that. So once I figure out that opt-in pattern, you guys are all getting spammed with emails. Awesome. So let's get started. Awesome.