GES C3 Clay Yash Webinar

So yeah, welcome everyone. Uh, today we're going to introduce Go-To-Market Engineer School briefly, and then Yash, who you all know, he's going to take over and lay down some knowledge, uh, and he's going to share a little bit about the workflows that you're going to learn in Go-To-Market Engineer School. And whether you join our course or you don't, you're going to absolutely learn a bunch of things today. Um, we're gonna do a little Q&A at the end of this that Matteo's gonna moderate. And so yeah, let's meet the hosts. So I'm Jared Waxman. I live in Silicon Valley, currently in an operating partner role in venture, but prior to this led growth at Auth0 and then led campaigns at Okta. Lifelong learner in some of what we're doing in the course here —was inspired by or modeled after great programs like Reforge Pavilion, Winning by Design, and all these programs that were really inspirational. So I'm fortunate enough to work here with Matteo getting this course off the ground. Matteo, you want to say hi?

Hey everyone, Matteo here, Italian in London, and been in growth for the last decade, now consulting in content and product marketing. Marketing for tech startups, Series A to C. And yeah, it's been super fun to be working with Jared on GTM Engineers School, which has been now just about a year, I think, which is awesome if you think about it. Yeah, we're super amped to gearing up for Cohort 3. And yeah, without further ado, I think it's— we can talk a bit more about the program. Yeah, yeah, yeah.

So, you know, again, what's going on here in go-to-market engineering in general, you know, a lot of people, they contact us and they say like, where should I focus? What— there's so many tools, so many workflows, like I'm lost. I don't know where to get started. So we all know that AI is kind of reshaping all the different functions from engineering, software engineering to go-to-market. And the teams that figure this out are winning. So it can be overwhelming. So we built this course to be really hands-on. We brought together pioneers, founders, educators, agency practitioners to guide you through. It's very project-based. You're going to learn by doing. There's like 16 mini lessons across the go-to-market stack. You're going to get some theory. You're going to get some framework. But mostly, it's going to be hands-on work. There's no prereqs. It's, it's 101 level. It's for RevOps managers, AEs, SDRs, content and growth folks, you know, anyone that wants to kind of, kind of stay ahead here. Be about 28 hours of live class and help sessions over 5 weeks. And in the remaining 3 weeks, we're going to have the option for you to do a capstone project, which we did last cohort, which was very valuable to to folks. So this is sort of the course schedule here. Week 1 is enrichment with, with Yash. We're going to start by building and enriching high-quality lead lists, automating personal outreach. You're going to learn how to combine enrichment, personalization, and deliverability best practices so you can scale your outbound effectively. And then in, in Week 2, uh, here with, with Zach, uh, from Octave, we're gonna, we're gonna step back and make sure that your messaging holds up. You're gonna define, uh, clear customer, uh, clear ICPs, ideal customer profiles, turn that into messaging that can— you can operationalize across outbound, inbound content. And you'll also learn how to score and prioritize your leads so that your effort really pays off. And then in week 3, in week 3 with Lorenz, we're gonna go through how to build scalable AI automations. We're gonna introduce some go-to-market engineering principles, some systems thinking. We're gonna apply them into some, some real automations using tools like n8n. You're going to create AI-powered workflows for things like email classification, routing, and automated responses. That's what we're going to do in Week 3. Week 4, um, Week 4, we're going to do Agentic inbound, uh, go-to-market workflows. And this is going to be with, uh, with, uh, Philip, uh, from, from Zapier. And he's going to show you how to design Agentic inbound workflows turning raw form fills into intelligently scored, routed, and personalized follow-ups using a mix of automation, AI judgment, and also human review. And then in the 5th week, we're gonna focus on AEO with— actually, with the slide is— needs to be updated— with Owen from Airopps. And we're going to teach there how to optimize SEO and AEO content, systematically repurpose content you have into short-form and messaging assets so the content becomes repeatable in your GTM engine. And then the final 3 weeks is going to be optional for you to do a capstone project. And in the capstone project, you're going to get some guidance along the way. This could be for your employer or, or, or showcase portfolio project that you want to put together, and you're going to submit a Loom video at the end of that, and you're going to get feedback from us. The, the, the pace of the course works like this: on Wednesdays, there's going to be a 2-hour lecture. It's going to be instructor-led. There's going to be a recording available if you miss it, and then Over the following week, there's gonna be live debugging sessions with TAs. What makes this course a little bit different is we, we do go through a variety of tools. As you've heard, it's very hands-on, led by experts. You get free credits for all the tools that we're teaching, so you don't have to worry about burn through your credit card for like the 2 months that you're in the course and learning. Like I said, there's a capstone project at the end, and we have a Slack community for building connections and and peer learning. So, um, uh, yeah, the, the seats are being filled right now. Uh, I think it's until February 9th that we're opening up, uh, um, seats, and then we close it a week before the course starts. Um, so that's it. And with that, I want to kick it off to Yash. I'm going to stop my screen share and, uh, uh, let's go. Sweet.

Yeah, let's jump into it. So I don't know, actually, I'm curious, like if everyone wants to do like a digital raise of hands, how many of you have like tried to use Play before or know how to use Play? Like raise your digital hands. You can click the little react button. Okay. A couple of folks. Nice. All right. So hopefully what I'm showing today then for everyone, oh good. Wow. More folks than I thought. Okay. Like half the room. All right, so I'll try and actually show you guys something that's more interesting than what you hopefully have done in Play. But if it's not, that's what I'm gonna leave the Q&A for, and then we can try and do something more exciting than that. One of the fun things about Play is that there are so, so many things that you can do in it. For today, I'm just gonna show you what I think is like the 3 most basic things that people need to do to get 60 to 70% of the value out of Play, especially if you're coming in like a RevOps or like an SDR AE background. 90% of those jobs the pre-AI era have been finding companies, finding the right people at those companies, getting information about those people, and then research on those people. The last mile of obviously doing the outbound to those people, which we can also even show you a beginner version of. So let's just show you how we can do all of that previously manual work in like a quick workflow in Quake. So I'm gonna go ahead and do this manually, but you could also, if you're playing with this on your own, work with our ever-evolving AI agents to kind of help assist you. But for now, I'm gonna come to a workbook. I am going to first start with finding companies. And then just to keep this simple, let's look for software development companies. Let's look for folks at the enterprise level. So I'm gonna assume headcount is, let's say, 500 to 10,000. I wanna keep it limited. I can see here in the top right corner, Or I should be able to see here. It looks like we're, oh, changing some things live. One of the benefits of being in the Clay team workspace is that the engineering team is always shipping things. So normally you can see right here in the bottom, see it's now, right? So 6,300 total results. And if I wanted to keep narrowing that down, I would go through, do a bunch more filters. I think for the sake of just this example, we'll do United States. Da da da. And now my list is narrowed down to 2,000 examples. Let's say I wanted to reach out only to companies headquartered in New York because that's where I live. Maybe I wanna offer an in-person meeting. Only 156 results. That seems good to me. I'm gonna go ahead and click continue, import it to a new table. And then once that runs, It's going to give me an option to pre-enrich some of these. I'll go ahead and just select the Enrich Company button because that's going to give me more information about all these companies. And then once we move through to this, I'm going to rename my workbook because renaming things is very important. Otherwise, we will lose things all over Clay. Cohort 3. And I can come back to my Find Companies table here. And so now you can see I've got these 156 software development companies based out in the US. I can start talking to Sculptor to enrich further details about these companies. I'm gonna go ahead and add this enrich company for the rest of my rows. And then what I'm gonna do is I'm actually just going to go ahead and try and find the people that I want to talk to at these companies. So if we assume that, let's say for example, I am selling, like, The people that I want to talk to at these software companies are likely going to be leaders in the revenue organization or the GTM organization. So to keep it relatively different, what I'm going to do is first look at their experience and seniority. So for seniority, let's keep it to rough VP and director titles since these are pretty large organizations. I could maybe even just keep it to VP and see how many results I get. Yeah. So even at the VP level, if I put in the job function as sales, or let's do marketing, or let's do customer success and support, the three pillars of GTM, I have 582 VPs at these 156 companies. Uh, and then if I just scroll through this, I can see that some of these titles might map, some of them might not, like regional vice president probably maps. You know, SVP of Marketing Growth, VP of Sales is probably more accurate to what we want. But also these people are scattered all over the world, which makes sense because these are large enterprise companies. So if I come down here to the person level as well, and I throw in the United States, that takes my list down to 419. Okay, these are looking better. I'm glancing at this table to see what else I might want to narrow this in on. 419 is still just a lot of people for this specific demo. So let's just take out customer success for a moment because sales and marketing is a little bit more direct to playing. 342. Stay in the bottom left corner. Let's take out marketing and just see what happens. 185. Awesome. So obviously if I was building this for like a full-blown outbound experience, maybe I would— I like to do 500 to 1,000 people in sub-chunks at a time. As long as you have enough similar characteristics between them that you can write really nuanced to sales copy. And I'll show you guys a little bit of what I mean there in just a second. So let's import to a new table here. And then I'll also just enrich these people. All right. So now I've got 184 VPs of sales at these 156 companies who are all based in New York. And this base enrichment that I have here, enriched persons, gives me all the information I could possibly get from their social media profiles. I have their titles, I have their awards, I have a summary if they filled it out, I have a headline, I can look at their previous education, all their previous job experiences, what they did in those job experiences if they filled it out. A pattern that I've noticed over the years being in VPM is the more senior an individual, the more likely they are to have a more robust LinkedIn profile. And so I already have a wealth of data I could use to start thinking about research here. But the biggest problem is I don't have a way to contact these people. So I need to first get their email. And you can see here we actually have an auto-suggested column that is going to add these emails for you via waterfall format. So if I just click the check button on that, it should automatically off of that thing to the rest of this table. So we can see here, here's this waterfall where we are using 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, oh, 11 now. Wow. Up for a mount. Nice. 11 different providers, right? Like, this is my favorite part of Playful, even with all the AI features, because I really remember the pain of being an AE back in 2014, 2015. And like manually uploading and downloading different CSVs between different tools to try and find the right data. So here we're actually finding the work email for all these people in just the first provider, and then we're validating it. Excuse me. In the second. And the thing that I want to show you guys is possible. Yeah, what happens when we don't find the emails? So when we don't find an email, then we look on the second provider. And then if the second provider doesn't find it, then we look to the third provider and so on and so forth. So that maximizes your chances of actually getting the right email while minimizing your costs. You don't wanna have to overpay or double pay for emails that you bought with WhatMe. And we can see that tends to like really, really up your accuracy and coverage here. We have a lot of emails being found. And when, as I leave this table to keep running, I would bet that we would continue to get a higher and higher coverage percentage here. So you can waterfall all sorts of data points in play just to give you guys an idea of all the things you could explore. It's both powerful and overwhelming for many folks. I come in here to view all enrichments. I can look at the data at the company level, by country, by description, by domain, by social media platform. On the person level, I can look by mobile phone number, email, education, background. Job start date, title. And then we have a bunch of other functions in here to extract, normalize, and score your data. Because as we all know, it takes a lot of time to actually get all the data in the exact format that your team might need it for your CRM, which is the bane of someone's existence, or at least the bane of my existence. The one last thing I'll mention on enrichments before I show you the last thing that I've got planned for today, and then we can leave it open to Q&A. Is that a lot of people have questions about credits, specifically saying like, why do some things cost more credits? Why do some things not cost any credits? And the simplest way I can give to you all to understand this is that we only charge credits for actions that cost us money to run. So as an example here, if I come into enrich person info and you can see here we have all these waterfalls that use all these data providers. One of the big benefits of Play is that you don't have to go out and purchase two dozen different data provider subscriptions. We have centralized all of that for you because we have independent agreements with all these data providers to do this kind of data arrangement at scale and then resell it to you at the like transactional level. So the credit cost for each enrichment is essentially a reflection of the agreement we have negotiated with that data provider so that we are not losing money on every single call that we make to retrieve data for you. And then the other thing that you'll notice, a lot of functions are free. So for example, any function that has this Clay logo has no credit cost next to it because we're not running any API calls to go get you data. It doesn't really cost us a lot of money. It's like fractions of fractions of fractions of cents to run code in the background at this kind of scale. So we can just offer that for free because it doesn't really cost us money, so it shouldn't really cost you money. And then we talked about finding companies, we talked about finding people, we talked about basic waterfalls for enrichment. The last big piece of enrichment that I think is really useful to cover is AI enrichment, and that can be used in multiple ways, right? So there's 3 types of AI enrichment that I think about. Number 1 is I can use AI enrichment to transform my existing data. So a really common example of this is job title. You probably want seniority from the job title. But it's hard to write a regular expression or a formula that will match everything because the formats are all different. For this table in particular, I only selected VP, so it's a little bit easier, but I could just use a regular AI call, not a web research, but a create or modify content call using OpenAI, Claude, Gemini, whatever your preferred AI model of choice is to insert that job title as a field and then return to your audience. That's type number 1. Type number 2 is the creation of net new content. So I think of that as email copywriting, or I think of that as image generation. And there's a couple of integrations I'll show at the end that I think are really good for that. But those are much more dependent on how good your inputs are. A lot of people jump the gun to AI outbound copywriting, but if you don't feed AI good inputs on what to write for the email, of course it's going to give you garbage outputs. Think of AI like a really, really ambitious, all-knowing high school intern. It will do whatever you tell it to do, but only what you tell it to do with the input that you give it. So you have to be really precise with the instructions that you give, and then you can get outsized results, which leads to my last use for AI, which is information gathering from the web. And so that is this web research plagiat column, and this can be really, really really powerful because essentially between waterfalls, you can now get all the data that you're spending a long time manually gathering from your adjacent data providers through a much more quick, automated, consolidated flow. Then there's still what I call like the last mile of data. It's all of the data that you can't get from the data provider that you probably have an offshore team going to try and find research for personalization in your outbound copy. It's, you know, what kind of beverages do they like from their social media platforms or What is some thought leadership that they might be involved in? Or it might be an industry parameter that's not in a data provider that you would use. For example, like the number of beds in a hospital or the number of businesses in a warehouse parking lot. Those are all so unique that it's pretty much impossible for you to find a data provider that will give that to you. And so your only method is to send somebody to go copy paste that data into a spreadsheet or CRM unless you are able to intelligently prompt a web research agent. The one example I will show you guys using a Metaprompter is for thought leadership. Write a prompt that finds me any elements of thought leadership, ideally in the form of blogs, interviews, videos, podcasts, or anything else that you think is relevant for each of these individuals in this table. Also, just because people always ask me this when I do that, that tool that I just used is not Clay, it is Whisperflow. I think it is incredibly accurate for voice transcription and has changed the way that I work. Shameless plug for their tool. I do not work for them. I get no benefit. I just even love the team. And also I wanted something to fill the time while this prompt gets written. My one other thing I'll mention while this prompt is being generated, we're working on making this meta prompter even better. There have been a lot of advances in meta prompting over the past, I would say, month even. Where there's a few people now who are using MetaPrompts to feed data into a Playtable, check how accurate it is, and then have Claude rewrite the prompt itself and make it more and more accurate until you get the right data on each of your data points. If I had a lot more time, I would probably break this prompt down and like edit it a little bit more. Prompt engineering is an important part of this step if you really want to have high accuracy across your entire dataset. But for the For the purposes of a demo, we'll start with this. We'll use just a Clay Neon model. You can compare all the different models that are available to you here. Obviously the AI race is continuing and becoming ever so increasingly overwhelming. But we will keep it to Neon and I will save and run just 10 rows. That's the last tip I'll give you on building columns in Clay Tables is always test on a smaller subset of your data. Before you scale it out to the rest of your table because you don't want to be spending credits unnecessarily on a prompt that you actually need to refine across 200 rows. Why not instead test on 10, iterate, refine, and then scale it out to the rest of your table so that you're spending your resources as effectively as possible. And then the last, last thing that I'm going to show you here, just because we were talking about AI being all the way up to a sequence, is This is actually a little bit of a preview to another partner in the GTM course. We have integrations with Optive and Twain. I personally am a fan of Optive. You can see that we have this generate personalized email sequence option. And so now what you can do in place is I'm gonna do a very basic version of this, but again, the more you configure, the more inputs you provide, the better the outputs get from Optive. I'm gonna set a 3-email sequence. I want the average length to be short. I want the sequence type to be pure cold outbound. I don't know any of these people. The primary goal is going to be to schedule a meeting. Output format, I'm gonna want just raw text. And then I wanna generate unique subject lines. I could provide examples here, which would make it really, really good at tailoring to what I might want. The primary URL for my product is clay.com. The name is Clay. And then again, I could go through and add in the core value proposition, the key differentiator, all of this good stuff. But for the sake of simplicity, all I'm going to do is add additional information about the offering from this research column that I have here, right? Leadership insights. I'm just gonna put all of the insights in there and let it use the research to draft some copy. I'm gonna continue to adding fields and I'm gonna let this also run while the research is running. And so what you'll notice here is that these actions also get queued sort of on top of each other because I'm using the leadership insights in this AI sequence generation. It's not going to run until all the data that it is dependent on has been gathered. And then Octave is gonna take a second to run as well because it's going through and generating each of those 3 emails in an identical manner. Looking through all the information I've given it. It shouldn't take too, too long because I only gave it the research as an input. The more research and the more input that you give AI, as like a last lesson here, the longer you should expect it to take. The same way that if you gave your high school intern 10 pages of instructions versus 1 page, the trade-off you're making is it's gonna take them an extra 5 to 10 minutes to do the work, but ideally at an increased 30 to 40% quality level, and AI has much fewer diminishing returns on that time-to-quality trade-off. So let's take a look quickly at some of these leadership insights so you can see the result of this web scraping. So here for Ian Marks, we've got primary source with some LinkedIn profiles. He has some articles offered on LinkedIn, and he talks about ecosystem partnerships, digital self-care, collaboration. So this is good, right? Like, this is the type of stuff that I'm talking about. Where if I was going through and actually writing an email here, here's one of the modifications I would make. This research is correct. Correctly though, it's a little bit old. If I couldn't find anything else about Ian, I would maybe rely on a 2019 paper that he published, but obviously 2021 is better. The dream is something in the last year. If I can find some activity or some research in the last year, that recency alongside that relevancy, really the perfect combo for maximize chances of a response. And then I could go through and kind of look at that for the rest of these people, modify my prompts, keep updating to further the quality of what I'm doing. Let's also take a look at the outbound copy that I'm drafting. So we've got speed up partner onboarding. Hi Kyle, you manage partner load growth for a platform that must sell many COVID-15s and muscle signals at scale. Play that whole thing in the Nickman intent sources into one dataset. This is exactly what I'm showing you guys on this call, which is kind of funny. Applies AI cleanup and then runs conditional fallbacks. Are you open to a brief call? Right? And that's, that's the email copy from like minimal data inputs that I've put in there. And you can imagine how much better that copy gets if I give it examples, if I give it more inputs, and if I give it specific instructions on how I want it to write each one of these emails, then it also has a second email and the third email. and then I can throw all of these into a sequencing tool. We have a native sequencer in Clay that we could throw all these emails into. So just to show you the basics of how that works, we could create a Clay email campaign. Yeah, sure. What's the—

go ahead. What's the sender settings with the native sequencers?

So when you say sender settings, what do you mean?

Like what, which email do you send it from? Domain? Yeah.

Yeah. So that's, I'll show you right here actually. So let's call this campaign GTM cohort 3. Let's add 3 messages. I will just use rows from following table. I will come back and do this later. We'll cancel. I'll save the campaign. And then if we come into the settings, you'll see sender setup right here. So essentially you can connect any of your inboxes here. You'll see that we have a recommendation on the volume you should use for that inbox just to comply with safe sending limits. We also can tell you if the inbox is being warmed or not warmed, we'll do some of that natively for you. And then we're working on the ability for you to just purchase more inboxes directly in Productivity. Thanks. Yeah. And then you guys are familiar with Sequencers, I assume, so I'm not gonna go through the whole process of like putting all of the text that we generated in the Playtable into the Sequencer. But that is pretty much in what, 15, 20 minutes maybe, an end-to-end find companies, people, research, and email flow all in Play. So I'll take a pause there and I see a a lot of messages in the chat. I minimized everything 'cause I'm only on one screen. I'll stop sharing for a second actually. I can reshare if helpful. And it looks like Matteo and Jared are doing a wonderful job of answering all the questions in chat. I'm curious if anyone has questions that they wanna shout out live, things that they want to ask about the workflow that we just built, things that they wanna see as other functionality in play, or is everyone else's?

Yeah, feel free to come off mute and ask your question or tap, type it in the chat if you'd like.

So will we have the opportunity to learn this in the course of the course?

Yes, yes. I mean, Yasha is going to go through over 2 hours introduction a little bit more slowly. Obviously, we're trying to condense this to a webinar, and then you're going to have um, in our learning management system, Loom videos, um, on specific topics, kind of breaking it down in exercises for you to do asynchronously, step-by-step instructions. And then you're going to have live help sessions with TAs during that week where you can pop in and say, hey, I got to step 6, or my use case is a little bit different than the exercise, how would you think about it for my specific situation? And so, yeah, you're gonna go through this. Specifically in the enrichment module, we've got async exercises for TAM sourcing, for waterfall enrichment, for intent signals, for CRM integrations, setting up sequences, account and lead scoring, and then AI prompting and Claygent.

Thank you, because I don't think I could do all that just having him see it, do it so quickly and so well.

Of course, but like, keep in mind that, you know, like on Tuesdays we'll have— Yash will deliver like a 2-hour deep dive. You'll come out of that with a bit more familiarity on the overarching product, underlying product, and then we'll give you one simple exercise or one basic exercise and then one more advanced exercise for the ones that are more, you know, want to get a bit more sucked in. And, and yeah, and you know, you can submit those to us and by just doing that you'll work towards getting the certificate. At the end of the, of the course.

Thanks, Robert. Question: would you say that this is similar to go-to-market execution? My background is in product marketing, and I'm curious to know how go-to-market engineering and go-to-market execution are similar and/or different.

How do you define, uh, go-to-market execution, Whiteness? Great question.

So as a product marketer, my job is to drive adoption for products specifically, and when developing the strategy, positioning, messaging, value props, channel strategy, all that good stuff. Now someone has to actually deliver on that work and put it out there, right? So the question is oftentimes with product marketers, are you the one that's just going to come up with the strategy and then ship it off to someone else to execute it? Or are you going to be the person that can execute it? Specifically in the startup space, you're looked to be a full stack strategy and execution. So I'm curious to know how does go-to-market engineering with the inclusion of AI workflows almost help expedite that execution, which is oftentimes the heavy lift aspect of strategy?

Yeah, I can, I can jump in on this because actually, uh, recently we've been building a bunch of these flows for our own product marketing team, right? So it's the way I think about like data marketing engineering in general is just really the biggest thing is that you can engineer systems that will transform data in rapid ways across your team. So wherever data is a blocker to go to market execution for you guys or for product marketing, there's probably a more creative way for you to build better systems. Two concrete examples that I'll give you are number one, our product marketing team is constantly asking our salespeople, what are the best customer use cases we have at the enterprise level? What can we talk about? Do we have logo rights? Do you have really cool tables that we can turn into collateral for the new sales team to go sell new deals with all these other enterprise customers? And then we have to go do enablement for that. And that's a really manual process. The things that we can't change are the enablement training side, just takes time to train humans. We have a little bit of like our own async resources for that. And then obviously we use AI a little bit in the process of creating those materials. The hardest part is actually sourcing all the tables for that, right? It's hard for salespeople to then take all the time out of their heads to give it to a product marketing team and then actually remember what all the most unique cases are when they're handling 30 to 40 accounts. What we have now set up is every time a salesperson or a customer success person hops on a call, that gets recorded via Gong or Attention. It gets stored in our CRM, and then we are constantly parsing all those transcripts with AI for a couple of things. We're looking for competitor mentions, we're looking for keywords, we're looking for expansion opportunities, and then we're also looking for unique and interesting value props that product marketing could use. Each of those, when detected, gets sent to a different Slack channel. In our company Slack, and then each of those teams gets to scale it up much more quickly using their own AI workflows or manual work from that instance. Does that make sense, Wadness?

Wadness, and yes, that definitely does make sense. Thank you for breaking it down. I'm looking forward to connecting with you when I want to learn more.

For sure.

To jump on that, actually, like, I mean, I'm a product marketer by trade, so like, I can definitely relate to your scope. And I think like, what's the beauty of like folks like us that are used to act like cross-functionally between sales, marketing, product, and design, you know, like products like Clay that are very high ceiling, they really give you like the ability to understand what could be done and they expand that horizon for you. So, you know, like, and again, like that goes back to what Yash was talking about with creative use cases. So, you know, like I, for example, you know, like I never really sent emails with Clay, but I used it to create competitors pages programs at scale, you know, like, so I use it for creating like 1,600 competitors comparison pages and push those to Webflow, right? So like, that's, I think, like a very, you know, like a unique use case that I've never seen, you know, like from many people using Clay, but, you know, me as a product marketer or, and content person, you know, like it was really useful and Clay really gave me like the building blocks for that.

Thank you. I appreciate that.

Of course.

Hey, I had a quick question, um, about a workflow, uh, kind of piggybacking off of what you were saying, Matteo. Um, I got an email from Clay and it was about, um, you know, acting on news events, tracking career movements, creating custom signals, detecting when your ICP visits your web pages and monitoring brand and competitor mentions. I was wondering if we would be, uh, kind of building out that workflow during, uh, the Clay module or any module within the program.

So just, just to make sure I'm understanding correctly, you said the workflow is like the workflow that I just showed, or I missed some part of it, you cut out for me. Oh, can you hear me better now? I can hear you better now.

Yep. Awesome. Uh, I was saying that I, I got an email about a workflow from Clay. Um, and it mentioned like acting on news events via Vanta, tracking career movements via Rippling, uh, creating custom signals, detecting when ICPs visit your page and, uh, monitoring competitor mentions in tools like Cursor. So I was wondering if we would be building that out during any piece of the module. Okay.

So the specific answer I have, and then Matteo Tittarelli can correct me if I'm wrong here, is like, we won't build exactly every single workflow that you just mentioned in the course. What we will do is cover the principles and foundations that you would need to build any of those. And then if you were specifically interested in doing those, I think you could cover it in your capstone or in your own project. Or if you also went above and beyond to do it, we'd help you to emphasize.

But I'll just, and I'll just add on to that. We do have in the async homework exercises intent signals. And I think you, I think that's where you're gonna learn the basics and start to get your hands dirty on some of those. You're not gonna get through all of them, but you'll get through some of them.

Okay. Thank you.

Conscious of time, and we have a couple of raised hands. So, Jason first and then Nicholas, please. Hey, guys.

Thanks for doing this session. It's really informative. And like somebody said earlier, like, Yoshi, obviously, you know yourself, and you're kind of like going through it pretty quickly, but it's really amazing kind of see how this can kind of happen. So I'm— this is maybe like 2 questions. I'm not sure. It's kind of all rolled into just the concept of AI and doing cold outreach. So, you know, I've done cold outreach in the before times. I'll say like before, like, Gen AI and things like that. And, and I've noticed that it just has somebody who receives a ton of cold emails that like now that we're in the age of AI, I get lots more of these AI. And you can tell, like, I get more cold emails and I get a lot more of these where I can tell that they are like really like AI-generated emails. And so for me, it's just like, dude, that's like not even— I am like, that's not even— there's no effort there. I'm not even going to look at it, whatever. But I'm just wondering if you have any data. So I still believe that it's definitely something that can help, not minimally, with efficiency and like creating these campaigns and being more personalized for sure. I think a lot of it is, of course, in the execution of it, which I trust you guys know what you're doing there. But I'm just wondering if you have data that you can share that shows, you know what, like this is what happened with before we did these workflows, this is what happened after, any type of efficiency data, conversion data, anything like that.

Yeah, so I think there's, there's a couple of different answers to your questions that like you said, there's like 2 or 3 questions in there. Yeah. I would say, so I'll give you 3 answers. Answer number 1 is if you're thinking about like ROI data, impact data on like the Clay level, I will happily drop, right after I answer this question, like a link to all of their case studies and you can see that answer to that question varies depending on the use case. So like in this session, we just talked about doing an outbound flow end-to-end. You could also imagine using the same structure for CRM and making it signal the ABM workflows, DTM co-pilots. So it just depends on the use case there. I think answer number 2 to your question specifically for play-to-outbound and stats on ROI there, I do think it varies depending on the company. The person that I still think is a world-class expert on this is Eric Miloszajski. He runs a virtual agency called Growth Engine X. My funny— or maybe not funny. My claim to fame for Eric is that he has gone up against the SDR teams of some Fortune 500 that I cannot name for you. And use single creatives to beat their entire SDR team in terms of meetings booked, right? Because he's able to generate creative campaigns, hypotheses, and research at scale in a way that humans are not able to continue. On the flip side, what I can also tell you that Eric would love to do, find out through sharing, is that he does a free offer for many clients when he's not certain if outbound is going to be the right channel for them. And there are times where he tests a bunch of different campaigns Is it really worth the fee? Those tend to be more niche industries. Those tend to be industries that aren't as online, but it is not— there is not a silver outbound bullet for every business, which leads to answer number 3 for you, which is that I think a lot of it is just about experimentation, testing, and iteration. The big thing that I think about with outbound these days is that it's always going to be changing. Like the life cycle of your tactics are getting shorter and shorter. To your point, I agree, personalization, like that's In the before days when I was doing outbound, if I wrote a really nice personalized email to like a COO or a CLR, I would get like 5 to 10% response rates if I did a really good job. That no longer cuts it. They're getting like a dozen emails like that. So now a big thing I think about is timing is also really important in addition to relevancy. So if I can be timely and you just switched a job or your company just raised money and I can be personalized, that tends to stand out. But It's always kind of a crapshoot depending on your industry.

Thank you.

And Nicholas, I believe you had your hand raised. Digital hand. No? Okay. Anyone else has a question? Question for Yash, question about Clay, question about the program, question about GTM engineering in general.

I'll ask another question because I'm just curious about this go-to-market engineering. Honestly, I didn't know about this a few years ago, so I'm just curious to know more about Where do you see go-to-market engineering going in the years to come?

Yeah, that is, that is a heavy question. There's so many answers I could give to that. I think for the context of go-to-market engineering is at least somewhat inclusive of all. A few years ago, I'd say probably 2 and a half years ago now, and I would say the reasons why we invented the role are the same reasons for where I think it's going to be going in the future. The theme that I think you're seeing with people at our level, you think is that more and more jobs are becoming engineering task in nature, right? So go to market engineer does not mean that you have to write Python or Java in the same way that like content engineer, actually Matteo Tittarelli had a really good post about this the other day. So like you're seeing the word engineer get added to more and more job titles that never had engineering before. And the reason for that is because with the ability to, for AI to speed up the manual drudgery of what you were doing before and also for you to build systems to do it, Now I can still do all my GTM work, but it's going to be much more effective for me to build mini systems and tools to get through each of those parts much faster than I could before. And what that requires me to have as a new skillset is thinking in systems, thinking in first principles, and thinking mutually exclusive but collectively exhaustive. So I think you'll see those sort of core engineering principles start to bleed into professions or fields that didn't have much of that before.

Thank you. And a quick follow-up, would you say that an early stage startup, would they necessarily need a go-to-market engineer? Because I know a lot of the examples were focused on enterprise, but if I'm an early stage startup and maybe I'm going after mid-market or maybe I'm going after enterprise, well, it would make sense. But just curious to know more about the use case of go-to-market engineering specifically for the startup space.

Yeah, so I think a little bit of nuance. We're actually about to launch a course from Clay on like go-to-market engineering for startups because that's really our origins and roots, right? Like I named it like an enterprise right now, but we started in sort of a small part of the market. Um, I can't give you the full answer in 30 seconds or more, but basically