Leveraging AI

140 | How C-Suite Leaders Can Master AI-Enhanced Outreach: 5 Pillars for Scalable Growth with Garik Tate

• Isar Meitis, Garik Tate • Season 1 • Episode 140

Share episode

Are you ready to unlock the true potential of AI in scaling your client outreach?

In today's episode, host Isar MMeitis sits down with outreach expert Garik Tate to reveal a powerful framework for reaching and retaining clients. They dive into Garik’s five-pillar system for crafting outreach campaigns that scale—and scale well. From the essential (but often overlooked) basics of finding the right market fit to the fine art of managing inbound replies with AI, this conversation is packed with actionable insights.

Garik shares how his Philippines-based virtual assistant (VA) agency has leveraged AI to optimize every aspect of the outreach process, providing a step-by-step approach C-suites can use to elevate their strategies, improve efficiency, and increase revenue. He also reveals key tools and systems to help your team stay organized, responsive, and efficient—all while avoiding the dreaded spam folder.

In this session, you’ll discover:

  • The five essential pillars of AI-powered outreach and how to build on each for maximum impact.
  • How to leverage LinkedIn, Apollo, and other data sources to build high-quality outreach lists.
  • The difference between personalization and relevancy, and why only one really matters for B2B outreach.
  • Proven strategies for enhancing deliverability and avoiding spam filters, from warming up inboxes to effective list segmentation.
  • How to combine AI with VA support for efficient inbox management that never misses a lead.
  • Garik’s favorite tools for managing email flows, organizing contacts, and streamlining responses to build genuine connections.

Garik Tate is the founder of a VA agency specializing in outreach for scaling companies, with a keen focus on optimizing cold outreach through AI. Based in the Philippines, Garik has years of experience in data-driven lead generation and combines an innovative AI-enhanced approach with a deep understanding of human behavior in B2B sales.

 

About Leveraging AI

If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

Isar:

Hello, and welcome to Leveraging AI, the podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. This is Isar Mehtis, your host, and we have an interesting show for you today. You heard me say this before, if you've been listening to this podcast, you can run your business. without HR, without finance, without marketing, without a sales team, without operations. It's not great, but you can do it without all of those. You cannot run a business without clients. Clients. If you don't have clients, you can run a nonprofit, but you can't run a business because then money's not coming in. And that's the whole point of a business. Now there's multiple Ways to get clients. One of the maybe least sexy ones that a lot of people avoid because they don't even know how to do it. And they feel it's weird is called outreach. The reality is called outreach is a very effective way to get new clients if you know what you're doing. So to know what you're doing is this. Fancy mix of art and science, right? And if you have a system in place on how to do this properly, you can get a lot more clients, which is something every business were once. But the trick is how do you do this? What's the process? And that's why I'm really excited about today's show because Gary Tate has a VA agency in the Philippines for years. And one of the services that he's providing for multiple companies is doing cold outreach on their behalf. Now. That means he has a solid system. And in the past year, they've obviously found ways to infuse AI into the five steps that they're using in order to do effective outreach. And they now, beyond the fact that AI allows them to do it better, it also allows them to scale them a lot more. And he's going to share with us exactly this, the five steps that they're using. For good cold outreach and how to use AI in order to scale and improve each and every one of those steps. Very exciting for me, hopefully very exciting for you as well. And hence, I'm really excited to welcome Garrick to the show. Garrick, welcome to Leveraging AI. Thank you. I'm excited to be here. I'm gonna look forward to this. Same here. So I, let's dive right in, right? I think people understand your background. You have an agency. You've been doing this for a while. It's not your first time doing cold outreach. And I really am curious about a, your process and B, how are you using AI in it? So I'll let you, follow your steps and I will try to. Make it as interesting as possible. I find this stuff endlessly interesting. So hopefully your audience does too. to make a cold outreach campaign that works, and when we're talking about cold outreach here, we're talking about, you can do it multi channel. So that could be an email plus LinkedIn plus text messages, plus, et cetera, et cetera, you can guess. Fancy as multi channel as you want, or, the core of it is still email. Ultimately, I don't think email is going anywhere. people have been saying for, last 10 years that this is the last year of email and, it's still around. And just saying sales is dead. no, it's still around. so that's what we're talking about, email primarily, but this can, this could apply really across multi channel. when you're building a cold outreach campaign, And I've built initially our first few companies on the backs of cold outreach. And then now we take clients and we add AI to it to supercharge it. you need five ingredients. as five pillars and your, Growth is really gonna be limited by whichever pillar is the weakest. so I think of it almost like an algorithm, A times B times C, whatever is the smallest number that's gonna be your constraint. big fan of the goal by Eli Goldratt, if you know that, the theory of constraints. But, the five pillars are gonna be number one is your product market fit. number two is your list, the list that you're contacting. Third is your deliverability. Are you actually getting, in people's inboxes? And are you doing that at scale? third, fourth is, the copy that's written. what people actually read when they first engage with you. And then fifth is, your inbox management, or really that's just your ongoing conversations. Basically, your sales, your back and forth in order to actually book the appointments. And then from there, you have a lead and you have a booked appointment. so those are the five pillars that, cold outreach done well is built on top of. No, I think it's amazing because I think a lot of people go to this thing. Oh, I'm going to send emails to people and see what happens. And if they respond, I will try to follow up as, as good as I can. and what you're describing is obviously much more in depth and goes way beyond just the sending the emails and responding to the emails, which is just, One aspect of everything that you define. So let's really go one by one and kind of see what companies need to do in order to perfect each and every one of those steps and how they can use AI in order to do that step better, faster, cheaper, maybe all of the above. Yeah. AI is a massive passion of mine. I'm building, Pretty much our entire portfolio of companies on the back of it. And, I had actually done a lot of cold outreach before catch BT, before, a lot of these tools came out of the market. And the problem was, I love the designing side of it, like the architecting the system, but I hated treating people like machines. I hated like having to, set like quotas and monitor people and make sure everyone was, hitting their numbers. I just. Hated that part, but once I came out, it's really possible to scale something quite effectively, quite ethically without having to treat people like machines. And so I'm huge fan of adding as much automation and AI into this as possible. So let's, just take the first one. Yeah. So product market fit. That's what you said. so product market fit. So this is, essentially the idea that, if you are selling, snow to an Eskimo, you're going to have a hard time. It doesn't matter if you are in front of them, doesn't matter if you're the best salesman in the world, maybe you can pull it off, but it's just going to be challenging, versus If you're selling, I think the famous example is, hot dogs at two o'clock in the morning just as the bar closes or, toilet paper during a lockdown or what have you, that's gonna be something very easy to sell. So there's really two parts to that. One is it you have a product that has a lot, gives a lot of value, and you have access to a market, or in this case, the market that value. Is given to, and, was actually talking with you about this, but before this call, I, in my mind, this is something that can go incredibly in depth on, but I hadn't actually thought about how to add AI into it so much, from, for me, it was a lot more about, talking to your customers, asking them questions, but you actually had, talked with me about, how you use AI a little more supercharged. You want to share that? Yeah. I love this. You're now interviewing me. that's fantastic. but yeah, it's, so one of the things that I do for myself and for my clients is build a content, a very well defined content calendar, if you want. It's not really scheduled, the topics we want to talk to you based on the specific. audience that we want to hit. And the way I actually start is I actually start by a list of links from social media. For me, my main channel is LinkedIn. So I take, 10 LinkedIn URLs off the profiles, off the people that I'm either already serving, that are my best clients, the one that stayed with me the longest are paying me the most amount of money that are doing the most amount of activities with me. My dream clients. Or the people I want to have as clients, he's the perfect profile. Like I know he's the right size of company and so on. And I literally just copy and paste them into. into ChachiPT and I ask it to explain what is in common for them and build a detailed ICP, which is exactly what we're talking about, right? How do I define my ideal client persona? And it does an amazing job in it. And I give it the prompt is detailed, right? He talks about psychographics and geographics and age and sex and blah, blah, blah, blah, blah. But it finds ways how to aggregate. All of these things, and it creates, a page and a half of a very detailed definition of your client. And then what I do is I go ahead and add my stuff to it. So here's what I want to focus on. Here's the value that I bring. Here is the E things that I think are the pain points of these people. And then I ask you to do another cycle and I ask you to synthesize. What is probably the key messaging aligned with the key pain points and dreams and wishes of these needs of these people. And so this is how I help companies build a solid product market fit, because it helps them identify what's going on. Who the market is based on who their actual successful clients are. And some people, they don't even know, right? Some people are like, Oh God, I don't know. Like I can sell this to anyone, which is the worst answer you can give. but if you do, if you go through this process and you really think who are your current best clients, or if you're getting started, who do you think are your dream clients and why then AI can really help you together with actual links of actual people. Because. LinkedIn is a gold mine of data, where people are from, their seniority, how many roles they have. They know the size of the company they work for, how many people work with them, so many different things, education, right? All this stuff is available over there. So LinkedIn can build a very solid profile and then you can use it to build a messaging, which is, we can talk about in your next step. So yeah, that's how I do like product market fit with AI. I love that, and it sounds like what we're talking about here is using AI as your brainstorming buddy as yours, your second brain in a way, to reuse that old term to do the research, make sure you have a great product market fit. I'm somebody who thinks about everything in frameworks. I'm. and so just to dial in for your listeners, the way that I define product market fit is that you have a high delta between the perceived value, the, the perceived value, the, Price and the cost you have as high of a delta as possible between those three variables so you can imagine that is that the perceived value is as high as you can possibly make it and that's perceived value, of course, because sometimes you give amazing value, but it's hard for people to perceive it with where they're at in that moment. So you want as high of a delta as you want as big of a difference between that and your cost as possible. simultaneously, you want the cost to be as big of a difference from your cost to your, to your price, to your, yeah, I can reverse the terms, but I totally get it. if you're selling something that people think is 10, 000 and you're selling it for 1, 000 and it cost you 100 to do, then you're running a very solid business, right? Because people are going to pay you. 1, 000 every day to get something they perceive is worth 10, 000. And you're still making 90 percent marginal in each and in one of those things. So I totally get it. I think it's a great way to look at it. Exactly. and if your deltas are good across those three, then your conversion rates can be extremely high, which is that's just living life on easy mode. there's a lot of companies out there that succeed way better than you. And you wonder like, why am, why are they succeeding where I'm not? it's probably because their deltas are higher. The reason why I think the rich get richer. largely in business, is it their lifetime value per customer is higher because they figured out how to upsell across a longer funnel, so their delta between their price and the cost to, to get those customers is higher because they monetize them over a longer period of time. This is just, the advantage of being in business longer of adding more to your business. If you can hit that product market fit, it's one of the things that matters the most in your business. You can't and you can't shortcut it. but let's move on now to, to the other pillars. Yeah, no, by the way, I love this framework. I never heard it before and I think it's spot on. I love it. I stole a lot of those ideas from Alex Ramosi. I think he's an amazing guru. if, if your folks follow him, some of the ideas might be recognized. so the other pillars are your list, the deliverability, Your copy and inbox management. So the next one is the list. once you know your product market fit and you have the product that has a value, you have the market then that you want to advertise that to. The question is, where do you find them? And so for this is something that has definitely gotten easier as better and better tools have come out there on the market. So the primary three that I think most people can get started with, but there's really three plus other. So I'm gonna go through four, four examples. number one is LinkedIn. So there are great tools out there that allow you to, gather data from LinkedIn. if you have sales navigator. I highly recommend you can create a very detailed filter set to then list all your ideal client profiles. now that you've defined them by, the amazing methodology of, using AI and you define them, you apply that to the filters and LinkedIn, and you're going to get a great list. you can scrape that using tools like Phantom Buster. There's a lot of tools out there that actually let you scrape it. The challenge with most of is that they're going to bottleneck, using your account. So if you LinkedIn will notice if you scrape more than I believe right now, the best practices, this can change from time to time. I believe right now the best practice is less than 2, 500 scrapes per day. So if you want to do more than that, you're gonna need to look for some other tools. some tools that you can do is you can, rent third party. LinkedIn accounts through a store like Accountify, although I'm not affiliated with any of these, but these are some things that your listeners can look into, so you can rent, one and then just plug that in with Phantom Buster or another scraping tool, or the only other tool that is out there on the market, and which I would say is Bye. Bye. Frankly, at this point, it's not super well polished, but it does solve this problem really nicely. so your audience can contest with it. It's actually golden leads. I can't really say I recommend them because like I said, they're not super polished, but there's, somebody that lets you scrape a large data set from LinkedIn without having any bottlenecks, because they have their own farm of LinkedIn accounts that, that do some of the scraping. so that's. Probably going to be where most people should start on the other. I want to pause you for just one second, because I think the question people ask himself also on these tools, phantom buster is definitely the gold standard from my perspective. And I've used multiple of these tools in my past. it's the most expensive by a big spread, but it's also the most solid and with the most amount of capabilities. Could you scrape? I'm like, okay, I can scrape. Contacts from LinkedIn. How do you know which one you want to scrape? Yeah. so for that, then you have to get, the best filters. and so the way that you start with sales navigator and yes. And then build the filters and then do it this way. So again, I want to add to that as well. Something that I'm doing and you, anybody can do as well. I scrape anybody that, that is in connection with me, right? So anybody who comments on my posts, anybody who likes my posts, anybody who reaches out to me, through direct messaging, anybody who, Comes to my live events that I do every Thursday. So I, these people immediately get on my scraping list. The cool thing is, so I, okay, I have a fair following and, a pretty large and a pretty engaged audience. So people are like, okay, that's not fair. What if I'm just getting started? if you're just getting started, you can do the same thing with your competitors, right? the people in your field, in your niche that are posting regularly and they get, hundreds or thousands of clicks and comments and so on every single post. You can scrape their audience and then filter it out. And I'm sure we're going to talk about this afterwards. I'm not going to ruin that, but there's ways to then take this raw list and then filter that out based on the audience that you want to, focus on. Yeah, no, that's, absolutely where people will get the best results upfront. from, we specialize and then going from that point, then like scaling to 10 X, 100 X. But if you're just getting started, that's absolutely the best place to get started. and how. Because each one of these pillars I can go in with a significantly more level of detail and a lot more tactical. is this a good pace we're going through? Do you want me to dig more into the list and some other tips and tricks for generating the list? let's dive into generating the list. Yeah, I think you will know how. Perfect. on the list, if this is, your first time doing this, the way you do it is you create the filter set and then you take the search URL and you plug that into the scraping tool, like phantom buster or what, or golden leads. The main challenge that most people have there is if they want to be, scaling their campaign is each search URL. You can only have about 2, 500 results. Most people have some silly ways of trying to break down a larger campaign. if you have 10, 000 results in your search, like you have a very large targetable, total addressable market, Tam, the way that you break those down, should primarily be geography. So you can, there are 50 States in the USA. So if you have a hundred thousand. From there, then you just do each state individually. If you have one state, that's too big. You just hit that state minus the major cities in the state, and then just hit each one of those cities separately. And this is super easy because you're just building a spreadsheet of 50 links for the 50 States. You then plug all that in Phantom Buster and you're off to the races without. really throwing away any data. It's a very easy way that I don't see enough people talking about. So that's LinkedIn. Honestly, Apollo is very similar. It's the same kind of thing. And if your audience doesn't have Apollo, the two primary ways that you can scrape Apollo is one, they do have API tokens where you can scrape data directly, but that can be pretty expensive because they limit you to only a thousand for their basic plan. If you want to hit 10, 000 or above, what you're going to be doing is you're going to be using a tool like Instant Data Scraper. That's probably the most popular one. It's a Chrome extension. Where you then, make your search, and this is something you should really get a VA to do or someone else, but you make your search similar as the, LinkedIn, divide by geography, make sure it's 2, 500, same principles, and then you just turn on Instant Data Scraper, first you have to save it to then get all the emails out of it, but then you run Instant Data Scraper on that saved And then that will scrape everything for you. You copy and paste and a spreadsheet. if you don't want to do that, you don't have a VA. There are other tools out there on the market. enrich IQ and export list are the two that I'm aware of. you can test which one works best for you. They just work differently. but you can. can hand them the Apollo link, and then they'll give you all the leads for really pennies on the dollar compared to a more expensive Apollo, subscription. So that covers Apollo for, Google Maps. To be honest, for most people's campaigns, at least, for our clients, that, this doesn't come up as much. But there are a lot of tools out there that, work very similarly. we create, searches and then plug them in tools and it scrapes through the results. Much better if you have a more. mobile and like phone number based campaign versus an email campaign, but it can still be really helpful for email campaigns. So I want to ask you an interesting question here. So you're talking about pretty large numbers here, right? So you're talking about getting thousands from LinkedIn every single day, thousands from Apollo every single day, thousands from other sources every single day. What is a good scale? Or, and I know that's a very broad question because it depends on the product, depends on the service, depends on the pride points, depends on. What is a good starting point? if I have 20 leads a day to do a cold outreach, is it even worth doing? Like, where does it become interesting enough to say, yes, I can, because you've done a lot of these, like what's the average conversion going from completely cold outreach to an actual sale, right? So is it one to a hundred? Is it one to a thousand? Is it one to five? And again, I know this is a very broad question and I know the variation is high, but I'm sure there's ballparks where this thing starts working and below that it's not worth doing. I love this question and I will never dodge a question like this. I, the answer is, it depends, here's how I think about it always. there's actually two, two, Ways I want to answer the first, I'll give is some principles that are worth keeping in mind. And the second is just real numbers. So a couple of principles to keep in mind is that generally speaking, this holds true pretty consistently is that the cheapest way to get leads is going to be earned media because there it doesn't cost a lot of money. So earned media is just, social media posts, content, YouTube videos, LinkedIn posts, stuff like that. yeah. That and essentially warm leads, which warm leads are essentially just networking. That's what we call networking. so those are the two cheapest ways, and earned media scales very effectively, but you have to give tremendous value for free for a long period of time before it starts to pay off. But if you're playing the long game, nothing beats it on a ROI basis, but it takes a lot of time, a lot of energy. So time is money, but that is the number one. The second cheapest is call outreach. The third, then the most expensive is paid ads. So for most people, what I highly recommend is that if you're gonna be in a space for a very long time, you should do earned media. If you're testing out a space, you should do earned media. I'm sorry, you should not do earned media. You should do cold outreach. You should send out, let's say 5, 000 emails. See what the responses are, see who's interested, tweak the products and add another 5, 000. and. To be honest, probably more than you need. We'll talk about numbers in a second, but that's a good way to test a market. If you're committed to a market, you want to earn media and maybe cold outreach. if this is a channel that, that you're, want to engage with, but then earn media is really, The other way to go on top of, cold outreach and earned media. Just final note on this. I'm sure a lot of people can relate is we'll increase your conversion rate for whatever other marketing channel you do, because people will Google you and they're going to see your content. So earning is never a bad idea if you have the bandwidth and the skill for it. So that's some principles of how I think about this, but now to go into like just real numbers. Biggest caveat on the world, a hundred percent. It depends. It depends really on your product market fit, which is why that comes first, but. For most people, an industry average is a thousand emails for one fully engaged, qualified leads. So think of that as like a booked appointment. That's industry standard. Now, industry standard is brought low by the fact that most people suck. If you're really, so if you're quite good, you should be aiming for something closer to, let's say 300, a one, one qualified lead per 300. some people can do significantly better, but that's something that if you have a decent product market fit and you're on top of things, that's something to be aiming for. truly, 300, like 370 would, I've heard most experts speak in those ranges. We aim for 300 in our campaigns. so now I want to add something that is highly related to what you're saying. and then I want to continue like to the next steps of the list because I think they will connect very well. I think the biggest trick and maybe I'm wrong, you've done this more than I did. the biggest trick is making the cold outreach as warm as possible, meaning today with AI, like all these tools that we just talked about, collect information. That's all they do. The more information you collect, the more personalized you can make the outreach. The more related to that person, you can make it. The more that person feels that you actually are helping them versus annoying them, bugging them, trolling them, whatever you want to call it. Yeah. So if you can reach out to a person that is really your target audience, when you really understand the problem, where you know a lot of stuff about them, the company that work for their current role, the place they are in the world, this customer, they Service and what the issues they have with them, the more this feels less Hey, we've been helping 50 other companies, blah, blah, blah, blah, blah. Come and work with us. nothing about me. How do you even make that assumption that I need want to work with you? And so I think that's where AI plays a very big role, right? Because it knows how to take a lot of that data and turn it into a message. That sounds very personalized. And if you collected the right information, if you have the right product market fit, you can actually make this warmer. But yes, that's like my two cents, but let's go ahead. I want to hear the rest of your process. I love that. And let's actually skip to number four. so I typically put deliverability after the list, but let's talk about which is a copy and then we'll head back to deliverability. I want to share some high level thoughts on this first before we're going to tactical about you know what this is. so On a high level, I agree with what you're saying. 110%. I want to add in two, two nuances in this one is that there is a difference between personalization and relevancy. And if you. you'll settle for personalization, but you want relevancy. So personalization would be, Hey, I saw that you went to this college, go Lakers or, what have you, you went to Pennsylvania. Amazing. Buy my stuff. Like it's better than saying buy my stuff, but it's not as good as saying, Hey, I see that you're in this space. Did you know that people in this space benefit from X, Y, Z, that is significantly more relevant. So relevancy is showing. How relevant your product market fit is to them. It's showing, the product market fit part, how, what's the relevancy between those two? While personalization for personalization sake is better than nothing, but it's, it's not as good. And I think this speaks to a deep point, which is that, truly since the scaling of the internet and how the marginal cost of outreaching people has continued to shrink to zero, the, The societal norms are changing based on that. So before, if you spent a lot of time writing, a beautifully well crafted email, if the other person could recognize that work went into it, all things being equal, they were more likely to read it as a term of reciprocity. Oh, someone sent me a mail, a physical mail. golly, I got to read it. And we are hacking that we've already been hacking a pre AI with outsourcing and spin tax and a thousand other things. But we're hacking that so much harder with AI that's crazy, which means that rapidly culture is shifting and people's discernment for wasting their time is increasing as well. and so where the buck. The puck is versus the puck is going right now. The puck is still at a point where personalization still is better than nothing, but it's rapidly shifting and where it's going to be shifting to is really your A. I. Not personalizing just based on the data, but personalizing based on the data as it relates to your products. Let me give an example of that. so 22 services that we have in our, roster of campaigns. One is business brokerage. A lot of our clients are business brokers, thing like private equity and mergers and acquisitions space. so when contacting somebody, if they're interested in selling their business, one way to, increase relevancy is to say, we have all this IP, we know how to sell a business. You can, we'll look at their website, look at their content, and then give them. The types of buyers that would be most interested in buying their business at a fair price and with an ethical criteria set. so that's an example of, saying here are the buyers we will connect you with if you work with us, because we now have scanned your stuff. The other one is VAs. So these are personal assistants. Let's imagine that. And so telling their role, their company type, how they work with clients. Here are three tasks that VAs can help you with in your day over day life. So that's understanding them and using the AI to shortcut their thinking, which adds value because AI shortcuts the thinking. It's not just consuming data and regurgitating, which I know absolutely is not what you were saying before, but it's, I think, really valuable for people who haven't thought very deeply about this for your listeners to, to see where the puck is going with this. Oh, I love what you're saying. I'll add two cents and then we'll dive into the how you actually do this, as far as write good copy and do this. But, somebody said, and I don't remember it's like one of the big gurus, but I don't remember which one said that you cannot start a conversation in somebody's head. You can only join a conversation they're already having, meaning if you come and talk to them, I'm like, Hey, you should do this. And it's not a Monday morning problem. It's not something they've been thinking about already. Your chances of getting their attention is very low. What you are saying, which I agree with a hundred percent. I'm like, okay, I'm struggling with this thing. I'm struggling with this thing. I'm struggling with this thing. I'm struggling with this. That's what's happening in the back of their head. And you're coming in. Hey, I know how to solve this thing. People are like, okay, I want to know more. And so now sometimes you will hit that by luck. You'll spray and pray, right? So if you'll send a thousand emails and one of them will be relevant. but if you know your target audience and you know how to craft, and I love the way you framed it personal personalization plus relevancy, now you go to one, two, 300, that's perfect. Three times better than what you can do if you just spray and pray. And those other two thirds are wasted money, budgets, resources, people, licenses, tokens, like whatever, the entire ecosystem that are not going anywhere. So you basically just leaving money on the table if you don't do that. So I love this. Tell me, how do I actually do this? How do I have the right information and then use it? Preferably with AI to craft solid emails that will achieve relevancy in person personalization. Okay. so the. The tool that we use for this, I think is a tool that's come up on this podcast before, but we love it for this purpose is clay. com. Yeah. So the way that we automate this whole process is, we literally just plug in a LinkedIn link, that's a scraping on one side of the equation. Then it, percolates gets scraped and then automatically is added with APIs and loco tools, all the way to clay, which then we stack a series of, enrichments on. So with clay, it's, if your audience doesn't know it, it's a. Essentially, all it is, it's a spreadsheet where every single column is another API. So imagine that you're all your data that you've scraped on the far left and then every column you've triggered another API to do another action step and then another. And so you can do essentially chain prompting. So a lot of people. Early days, when chat GPT first came out, we're like, oh yeah, this is going to change copywriting forever. And it wrote terrible copy. it writes absolutely garbage copy because it's scraped the internet to learn copywriting. I know the internet is it's mostly bad SEO articles. So it writes like a bad SEO writer. That's it. That's what it writes like. because that's what it consumed. Like that's exactly what happened. I don't see enough people talking about that, but that's. What you know, you can trace it. It's a quirks to that. So instead, you have to tell it instructed very finely about how to customize a copy, how to write like the way you want it. And you want to break the problem down into a series of micro steps. And you can do this. And sometimes we do this ourselves. And we have a B split testing. Sometimes we just write a single master prompt that writes a single master email and says how we want to be written. But the best copy we write breaks the problem down into individual sentences and paragraphs and then as the AI Customize those before it all gets concatenated together So you can do something like that in clay chain prompting That through the problem clay is also incredible because it lets you scrape extra data on them So a lot of the day that you want to do To get, you can get with a clay, imagine enriching new email addresses, validating email addresses, scraping their website, scraping their LinkedIn. Those are the primary use cases that we have on top of the AI. a little again, just to quickly give, to give a. quick recap of what this can do. Think about collecting data from whatever source. So this could be what we said before, a scraper on LinkedIn, a scraper on Apollo, a list you bought from somebody doesn't matter. This is your starting point. And then you start doing two things. You filter them based on different criteria in clay. So every column could either filter the data or can add more data. What's the name of the company? What's the URL of the Go and get information from the Website. And that's an AI function, right? That actually goes there and quote unquote reads and summarizes what their website is about and what they're doing. And then what role are there in that's going to come from LinkedIn and how many people report to them that can come from LinkedIn as well, and what's their email address and so on and so forth, what products they sell and so on, and you can collect all that information. And then in the final column, you can send multiple data points from the previous columns to an AI like. Clay or Chachi PTO, it doesn't matter which, whichever one you choose and say, okay, I want you to take these pieces of information, the name of the company, what's written on the website, the name of the person, the role in the company, blah, blah, blah, and write an email that talks about the benefits of our product to what they're doing. And this is how you connect all these dots together, right? You take everything that Garrick told you so far of the data we've collected, and you create a personalized relevant message that then can be sent in whatever platform you run, right? This could be an email, this could be. Text message. This could be a LinkedIn message, like whatever it is that you want, you can now send that message, but creating the message based on all the information that you collected is what makes the magic happen. that's perfectly said. Yeah. Yeah. That's perfectly said. the only thing, I'll throw out at this point before, before moving on, cause this is a rabbit hole you can dig into for a long time is that. when you set up Clay, definitely as much as possible, replace, its native token usage with your own API keys. so it is one, literally 1 percent as expensive if you use your own, ChatsBPT, API token as opposed to using Clay's, native one. Clay has a very clear business model. I love how they're set up. This is how they make their money, but they expect you and they want you to use them efficiently. So you need to have your own API keys. And then there is one, one person. So I'm going to be putting out probably a lot more content on like clay tutorials. But I think that the best person, while those are being cooked up, if your audience wants to learn more about like how to use clay for copywriting is Eric Nowoslowski. If you don't know how to pronounce his last name, You can probably type in Eric and then Noah, and it'll pop up just to type in Clay Tutorials. he has good content on how to use it for a cold copy. Awesome. only do that though, after you finish listening this episode, then listen to this episode again, give it a thumbs up, share it, comment, you review the podcast, then go check out Eric's content only after you do all that. Okay, so now we have copy, right? Perfect. so what are the other steps? All right. and, Once again, every single one of these steps is constrained by the weakest link. So now that you have copy, you have to make sure that people can actually read it. You might have the best copy in the world, but now you need eyeballs on it. And you need as many of the eyeballs from your list as possible. this is deliverability. So with deliverability, the name of the game is Really twofold. One is setting up secondary accounts, secondary email accounts at scale that are well set up. and then second is monitoring them and making sure that your copy and your Email sent per inbox is at a reasonable rate. let's break this down. let's say you have a website like aico. reach. com. That will be your primary inbox. you don't want your primary inbox, your primary emails just, end up in spam. So instead, what you do is that you create a series of secondary domains, things like get AI code outreach or, team AI code outreach or my AI code outreach, whatever. And if you need help brainstorming, cause I struggle with this sometimes. So go to chat. gpt and say, give me 50 domains just like aico. reach. com plug those into, Namecheap beast mode and just see which ones are available, mass purchase and you're off to the races. now you have all these domains you have to set up the email accounts and set up your infrastructure there properly. we ourselves have been doing this, once again, maybe just going back to what I was saying before about I hate treating People like machines. I want you machines like machines. it's just a slog to set the accounts. There's so many validate this and verify this. And sometimes it works. And so we're right now testing out some vendors. I can't yet recommend any, but there's plenty of vendors out there. if you post on group chats or forums, of people that can set accounts, For you, right now, the main deliverability is at the beginning of this year, actually, a lot of people were skilled, scared of using, of conducting Gmail accounts. And so people, also were having trouble setting up Gmail accounts as they were setting up a lot of, work on outlook, but now it seems like the pendulum is shifting and to, to your audience, the pendulum is just always shifting. This is something that if you want to make this your full time job, you can get really, anal about this, but if it's just one channel amongst many, then. Don't overstress about it. There's definitely an 80, 20 amount about how much you should stress about this stuff. instead focus on making sure that your copy is actually good because that comes the second part. After your infrastructure is set up, you basically just need to, do. a couple of different things, but you just need to be monitoring to make sure that people aren't marking you as spam. That's something that you can't technically track. But it comes down due to your copy. You want to make sure you're getting a high reply rate. generally speaking, a lot of people can expect around 4%. but you want to be aiming for higher. If you're under 2%, you're probably having deliverability issues. the third is you just want to be making sure that your emails aren't bouncing. So that's where it comes like. When you have a list of emails that you're going to be contacting, you want to validate that list using tools like Scrubby or, the biggest one right now that a lot of people are recommending are better contact rocks. No, sorry. That's to get the emails to validate them. people use million verifier. Sorry. If you aren't being marked as spam, so people aren't, being pissed off by your emails, if you're not being bounced because you validated your list and you're getting a good reply rate because you have a good product market fit and your copy represents that, then your deliverability is going to be totally fine. the last component is the total number of sends you have per email account. And for that, everyone has their own secret sauce. the way that you need to do this, though, in, in broad strokes, Is that you want to warm up your emails over time. So whatever sending tool you use, and for most people I recommend instantly, but if you're a hardcore user, you should definitely go a smart lead. the smartly community is awesome. But if you're just doing this for like you, you have your own company or you're just have a single campaign for your own marketing department, whatever instantly is probably a little more friendly for a straightforward campaign. Whichever one you choose, they have tools to warm up your emails. What that means is they have a network of email accounts, pretty much basically all their clients, that are all emailing each other all the time. And what that does is it simulates real behavior. And so the settings you want to do for this, you want to do a ramp up period over the course of about two weeks, ramping up between four to five emails per day, let's say four, until you hit about 50. And then when you hit 50, you want to modulate between 50 to about 25. So simulating a high number of sends, a very high number of sends. I don't send 25. Actually, maybe now I do, but, a lot of people don't. So you're simulating a high number, but you set the reply rate to, everyone has their own secret sauce. We said to around 40%, some people recommend lower, some people recommend higher, but, we try around 40%. So basically again, to explain people why you do this, all the filters in the world that try to stop spam are looking for, Oh, is this a new email account? That's now sending. 5, 000 emails in a day. It's a spam source. That's exactly right. If you grow this over time and you build an environment where emails are constantly being sent and that they don't start at 5, 000, they start at three and then they go to seven and then they go to 20 and then they go to 30 and then they keep on sending and replying to these emails randomly. Like it's bullshit. Like it's not real emails, but it's just showing the email universe that this is a real Email address with a real person behind it, even though it's not now, when you send people emails that are your cold outreach, it looks legit, like an actual email address that is sending emails to people. And that's how you get around a lot of the spam filters. So that's like explaining, why do you want to do this? Huh? so that's. Deliverability. Is there another? You said the last one was inbox management, right? Let me say one last thing about, deliverability and then jump over. Smartly has just now recently come up with a tool. You don't have to be a user sending. If you use instantly, you can also use this, but they've just come up with a tool for monitoring your deliverability that's, helps out. Remove a lot of the busy work on the guesswork with this, because when you're sitting on email, you do with no HTML tags, preferably. So that means also no links, no tracking links, no open tracking, all that hurts your deliverability immensely. If we had more time, I could dig into it. But because of that, you can't really track your deliverability of a lot of blind spots. So smart lead is something we're just trying out. And one thing I really like is on their office hours. I think every two weeks for the next little while, and depending on when this comes out, they're doing reports on what they're seeing industry wide. So you can get a lot of insights into what's working. all right, moving on to inbox management. so when it comes to inbox management, you can. essentially automate your inbox management with AI as well. And so I'll be talking a little bit if you want about how we would architect that. But for us, we just still keep that part human. So if somebody is replying to us, we want The person who decided to reply back to be a human. We want them to be cybernetically enhanced. So basically as many shortcuts and as much efficiency packed in that as humanly possible. But that's, how, how we're doing it. So let's talk a little bit about if you wanted to do this with AI. And I agree with you. The way I do these kinds of things when I write, emails with AI, they always go to my draft outbox. I don't actually send them. And so you can do the same thing with your team, right? You can then divide the different emails to the different people based on the region, based on the company, based on whatever it is you want to divide it. But then you're saving them the work of finding those responses and seeing what they were about and drafting the first thing. So when they come, they have an outbox of, one, three, seven, 20 emails ready for them to go that they can read. Massage, change, update again, add that human touch of understanding what this client may want and what my service actually does or my product and so on. But you don't start from scratch and just save a few steps. So what would you do to orchestrate something like this with AI tools? Beautifully said. the way that we would do it is, we have a, something we do already have is we have a large database of common, replies. we have a notion page that we organize, with examples. We call our FAQ, of common replies and how we respond to them and every week when we're sitting down with our clients, we go through the edge cases or the things that happened that week that we're maybe unsure of. In the beginning, we had to reply much slower. Now we want to reply every time you. Within 24 hours or less, but before sometimes, you have to talk to the client about the tone about what's okay, what's not okay. So we grew this FAQ. And the way that we would handle this is that every time a lead replies, we set a webhook, and this is something we actually do for tracking, a webhook to receive the reply. And then we would send it to a, thing like a custom GPT, I think on the AI, on the API side, it's called a. GPT assistant, that has access to a knowledge base. So that knowledge base is your list of FAQs and you would go through, a set of chain prompts. So first of all, we categorize all emails into one of, I think like six categories, To simplify the one that's most important, that ultimately actually makes you the most money is just people who say, yep, cool. set up a meeting. So bam, meeting request. that's the thing that really matters the most. Yeah. Yeah. Reality is that's gonna be 70%, 90% of your revenue anyways. All your extra effort is gonna be, trying to get the rest of the 20%. so 80, 20, a hundred percent. So meeting requests, and those are the easiest to reply to you say, cool, maybe add any humanity, answering their questions, but you reply to that. Then there's also info requests where they want more data, referral requests when they send you to someone else, future requests, they say, hey, let's talk about this in the future. then there's also not interested in, and there's a few others. But those are the basics. So we would chain prompt it where it first says, what type of email is this with some examples. And then depending on which category, it would then look at the, database, the database, the knowledge base. To then say OK, craft an email where the intention is still to book a call. Here's five examples of how we book calls. Now look at the knowledge base to draw relevant questions, answers, and then it would as. The next step, send that into smart lead and add it as a variable attached to that client where then the inbox manager when they, bring up that email, then they can just check that variable and say what I had written for them. So I want to pause you just for one second, because there's a few really important points here. one is people like, I don't have this database. the reality is you do, because if you look at all the emails that you received and that your sales team or customer service team or whoever team does it, BDRs have responded, you have all those emails. So you can take all of that, put it in an AI that has a really large context window. If you want the biggest one right now, it's Gemini 1. 5 pro from Google. So you can take a huge amount of emails and throw it in there and it will help you create the FAQ document, like your actual database. okay, these are the most common questions we're getting. Here are the three most common emails that actually got us. Beyond that step, once we got the first response. So you can have all of this be built with AI. So that builds your knowledge base of how you should respond to specific categories of questions. The second step is how do you categorize the incoming emails, right? So that's also AI, right? So an email comes in and say, okay, these are the categories that I have. Here are examples of. The emails in each category, and then AI is actually really good at categorizing. So I've done this on hundreds of thousands of rows, and it still does a very good job in categorizing things into the categories. And then the next step is, okay, what are the responses? And it's the same thing. Here are the three best responses that we have. Here are the considerations where to pick one, two, or three. Go combine this with the actual specific questions that they asked and write a respond, put it in the outbox or in some kind of a management platform that people have access to that. Then again, when they come to it, they don't have to go read the email to the thing, so they, it does. Not 100 percent of work, but 50 to 70 percent of the work, which means with the same amount of people, you can do significantly more responses and they're going to be well crafted. That's beautiful. and if your audience is intimidated by that, but they've taken the other steps of the process, just get a VA to draft up all your documents. The things ahead of time. Go over everything that happened that day. Just go over all the emails with them and just train them like it takes two weeks. maybe more depending on complexity of them to get your tone. Make sure they record their goddamn FAQ. That's super important so they can refer to that later. you can make a diagram of your decision tree and if this if this, and then, just elbow grease it but this is a. is definitely ripe in our scopes for automating and using AI. yeah, I will add one more thing. If you want to go a level beyond that, you can use tools like make or Zapier or anything that allows you to then chain a few of these together, right? So with more considerations, with more external tools that can really do more stuff, say, okay, in these use cases, I want to use this thing in clay. In those use cases, I want to use this. process in chat GPT or a custom GPT or whatever the case may be. And you can do and get access in specific use cases to specific different knowledge bases. there's a lot of ways to make that even more, automated and more specific to specific use cases. Derek, this was, Fantastic. I think there's drinking from a fire hose, right? For people who never done this, you touched on a lot of points and gave a lot of information for people to think about on how to do this properly. If people want to learn from you, work with you, follow you, and so on, what are the best ways to do that? Yeah, the best way is to connect with me on LinkedIn, and I have a YouTube channel as well. Most of the content we put out so far has been, on AI generically, but moving forward, we're gonna be specializing a lot more just to be talking about these topics, cold outreach. So if you want to follow that, then definitely can. And, If, if your audience is interested, something we touched on a little while ago is, if you just want to send out a thousand emails, you're best suited to be sending to people actually engaged with you who know you. And so we're going to be, offering, listeners to show if they reach out to me on LinkedIn, we don't yet have like a profile. we don't yet have a form to fill out or anything like that, but just contact me on LinkedIn. we're going to be, offering to select, to select prospects. A free scrape of the people who have been engaging with their LinkedIn content, giving them their email addresses, giving them their contact info and what contact they had engaged with. So if you aren't sure who's been engaging with you, who's been commenting and liking, etc, but would like to and like to have a thousand emails of people who actually have engaged with your content, then hit me up on LinkedIn and we can have a conversation. Fantastic. Thank you so much. This was really great, very in depth, and I really appreciate you taking the time and share with us. Incredible. Thank you so much.

People on this episode

Head Shot

Isar Meitis

Host