Transcription
Hey everybody, welcome back to the channel. Today I'm going to show you how to build a fully automated AI outbound agent across email, LinkedIn, and voice. Let's get into it.
All right, so in this video, we're going to be building a really sick system that is going to automatically cold message people across email and LinkedIn channels with personalized AI research messages. And then when somebody responds positively to one of those outbound messages, we're then going to immediately call them with our AI voice agent. We're going to use Vappy uh to qualify them and then book them to a meeting. So, at the end of this video, you're going to have a three channel fully automated AI outbound agent that can book you meetings on autopilot. All right, so without further ado, let's get into this thing.
Okay, so step number one is we're going to come into Victoria AI here. And if you've never used Victoria AI, this is actually an app that we've built over the last couple months at our company uh that can handle all of your automated LinkedIn and email AI outbound sequences. Um it's absolutely sick. I'll show you how to use it. Uh and then in subsequent steps, we're actually going to use Vappy um here to build our voice agent. Um and then actually make calls to people who are interested.
All right. So, what we're going to do is we are going to come into uh Victoria here. I've created a campaign called version 7 uh voice demo. All right. And you can see here we've got kind of our analytics, which we're going to skip for now. I've already attached um our email and LinkedIn accounts here. Okay. So, I'm just going to go enable these. So, we're going to be sending emails from this email here, uh, and then my LinkedIn account as well. So, I'm going to enable that. Okay. So, that's all enabled and we're ready to rock.
So, the next thing I'm going to do is I'm going to upload a lead list. I've already pulled a lead list from Apollo uh in this video. Um, and it looks something like this. Okay. So, I've got first name, last name, company name, uh their company website, their email, and uh their LinkedIn URLs here. Okay.
Okay. So, what we need to do now is just map our fields. All right. And so we're going to do first name to our first name column. Last name to our last name column. Email. Uh LinkedIn URL is going to be our LinkedIn field. All right. Company website is going to be our company domain. All right. And these two fields are very important. Um this LinkedIn URL and this company website field are very important for two reasons. One, if we don't have a LinkedIn URL for our lead, we can't message them on LinkedIn. Okay? So just keep that in mind. But then two, and this will make more sense in just a second in our next step, we're going to be setting up custom AI personalization fields um that essentially we're going to let our AI agent scrape every single lead's LinkedIn profile and every single lead's website for information. And then based off that custom information that we scrape from their LinkedIn and their website, we can then um actually create or curate custom messages based off that information. So, it's very important that we have uh LinkedIn URL and company domain uh mapped here in our lead file. Okay, we're going to map also our company name field just in case we ever want to use company name in our sequence. Uh and then remember what we're doing uh in this video is we're going to do uh email and LinkedIn outreach. And then if anybody responds positively, we're going to immediately give them a voice call from our voice agent. All right. So, we also need to map our phone number field um to a a a new variable. Okay. Um, and so I'm going to come down here to let's see, where are you? Where are you? Company phone number. All right. And I'm just going to save it as this variable. Okay. I'm going to hit upload CSV. Okay. So, our leads are being uploaded.
The next thing I'm going to do is I'm going to come to our AI personalization fields and we're actually going to set up our custom research fields that I was just talking about so that we can drop them into our sequence here in the next step. So, this will all make sense in a second. Okay.
So, what I'm going to do is I'm going to add a new field. Uh, I'm going to call it icebreaker. All right. Um, and I'm going to give it some custom instructions that says something like this. You know, research this lead in their company and write me a custom icebreaker I can use to kick off a cold outbound sequence. And then I give it some examples. You know, love what you're doing at XYZ. Um, you know, huge fan of what you're working on at ABC. You know, stuff like this. Okay. And the idea here is we're going to generate for every single lead, uh, a custom icebreaker that we're going to open with in our LinkedIn and our email sequence. Okay. Now, really cool here is I'm going to go check this read website and LinkedIn profile buttons. All right. And what that's going to allow me to do is actually scrape each lead's LinkedIn profile and website. Remember, it was very important to map those fields in the previous step um to build a custom researched knowledge base for every single lead. Uh, and then based off these instructions here, we're going to take all that knowledge that we were able to gather from their website and their LinkedIn profile and write that custom icebreaker. Okay? And we can feed this into our sequence in a second. All right? So, really, really powerful stuff here. And remember, you're doing this for every single lead. Okay?
So, I'm going to set up one personalized field called icebreaker. I'm going to set up one more. This one's going to be called painoint, right? I'm going to give it instructions that sound something like this. Okay? I want you to research this lead uh and tell me what specific issues this company has in building a client acquisition funnel. Okay. So, for this campaign, we're basically going to be selling our our SDR software. Uh so, we're going after people that are having issues getting clients. Okay. So, basically, I'm telling uh our agent here to go, you know, research their LinkedIn and their website, go scrape it, and then based off that information, I want you to give me a a customized pain point for this lead. Uh that tells me something about what trouble they're having in building their funnel. Okay? And I will hit add field here. Okay? And so now we've got two personalized fields that we can feed into our sequence here in the next step. Okay? So, really cool stuff there.
Now, the next thing I'm going to do is I'm going to come to my sequence and we're going to start building our actual conditional sequence here. Okay. So, what's really cool about Victoria is um we can set up different flows based off um things that happen in our flow. So, for example, we're going to start off by sending a LinkedIn connection request. All right, we're not going to add a note to this, but you could if you wanted. So, we could say, "Hey, just reaching out. Uh want to share something cool with you or whatever." Uh but we're just going to leave our LinkedIn connections blank for now. Okay. We're going to wait 3 days. All right. And if at any point in those three days, our lead accepts our LinkedIn connection, well then we're going to send them down a conditional flow where we can send them LinkedIn messages. Okay? If they don't accept our LinkedIn connection after 3 days, well then we're just going to email them. Okay? We're just going to send them down an email path. Okay?
So let's build these out now. So if they connect with us on LinkedIn, well, we're going to send them a LinkedIn message. And that LinkedIn message is going to look something like this in this example. Now, this can be whatever you want, but this is what we're going to do today. All right. So, hi, first name. Drop in that icebreaker. Remember, that's a custom variable that we just um researched in our previous step. This is going to be unique for every single lead. Noticed company might be painoint. And again, that's that second AI personalization field that is going to be unique for every single lead. All right. Uh, basically we're going to frame their painoint against our solution. So, hey, we notice that you're dealing with this. Hey, we've built an AISDR that can help you solve that issue. Can I show you a quick demo? So, we have a CTA at the end. All right. So, that's going to be our first LinkedIn message. All right.
And then what's really cool is if after a day we don't hear back from them, well, then we're going to send them an email. And that email is going to be something like this. Okay? So, it's going to be like, "Hey, first name. Hey, I just sent you a note on LinkedIn." Which is really cool. So, we're going to start playing these sequences off of each other just to boost our reply rate. So, it's super sick. Um, but I I I figured I'd follow up here with you, too. Um, looks like you might be experiencing painoint. Um, by the way, our AI SDR makes outreach easy. Can I send you a demo link? Okay. I'll add a subject here, um, which is like, you know, quick demo or something. Okay. We're going to wait 2 days for a response on that. All right. And if we don't get a response on that after 2 days, then we're going to send another LinkedIn message. All right. And this one might say like something like, "Hey, just following up one last time here. Um, let me know if you're interested." All right. And for the sake of this demo, that's good enough. Now, you can make this sequence as long as you wanted. Um, but we'll just do one. We'll do two LinkedIn messages and one email, and that'll be good. All right.
Now, let's just go finish this up. So, if they don't connect with us on LinkedIn after 3 days, well, then we can't send them a LinkedIn message. We're just going to send them a couple emails. So, we'll send them two emails. All right. So, again, we'll do like quick demo. All right. And then we'll go grab like a similar email here and just clean it up a little bit. So, I'll grab that and I'll paste it in. All right. I'll zoom in here for you guys. I just sent you a note on LinkedIn, so I can get rid of that. Um, I'll probably just do uh probably just do icebreaker here. There we go. Okay, perfect. And then if they don't respond to that after 2 days, then we're going to send them one more email. Again, you can send as many emails here as you want, but you guys get the point. Okay, so you can build these really cool sequences here. We can also do AB testing if we wanted. Um, but it would look exactly the same. Um, just like this. And and the system will actually AB test. so you can see which one works best.
Okay, so last thing we do before we move on from our sequences here is we're just going to preview this and make sure that our personalization fields are working. All right, so I'm going to hit this little button here. All right, and I'm going to hit generate AI fields and let this kind of cook. And I just want to make sure that um our icebreaker and our painoint sound natural inside of our template. Okay, so this is kind of just like a last step little test we do uh before we turn this um this campaign on. Okay, cool. So those fields are generated. All right. So, hey Aaron, huge fan of what you're doing at Buildify, especially how you package turnkey systems. All right, cool. Sounds good. Notice Buildify might be struggling to keep a steady pipeline of qualified clients because outreach is manual, referral- driven, and unscalable. All right, cool. Sounds good. So, that looks good. We are good to go there.
All right. Now, what we're going to do next is since we have our leads uploaded, our personalization fields crafted, and our sequences built, we're very close to being able to turn on our email and our LinkedIn outbound. The last thing we need to do now is actually turn on our AI autoresponder and connect that to Vappy in our voice agent, which is going to be the second part of this video. All right, so what do we do?
So, I'm going to come here to responses. I'm going to turn on our AI responder. All right. And our goal for this campaign is to book meetings. All right. And so what we're going to do is every single time we get a response uh on LinkedIn or email, uh our AI autoresponder is going to handle those responses automatically for us. We we don't have to do any manual uh responding if we don't want to. All right. So, what we're going to optimize for here is booking a meeting. So, I'm going to paste in my calendarly link. So, anytime our somebody responds, our AI autoresponder is going to try to get them uh to book a meeting by sending our calendarly link, right? Then we're going to basically train our responder, give it some information about like what it is we're doing. So, we're trying to sell uh an AISDR product that helps B2B clients book meetings on autopilot. Okay. All right. So, we're basically just training our AI autoresponder. um giving it a little bit of information about what it's trying to sell. Um we can upload some supporting documents here um that can essentially serve as a knowledge base for our autoresponder. So if it gets asked questions from a lead, it can answer them. You know, if somebody says, "Hey, how how much does this cost? Where are you guys located?" Uh we can load all of that in here in a PDF. Okay. So I will do that here. I'll just upload a little PDF here. Um just like that. I'll say uh be friendly and kind. Okay, so I'll give it a little tonality. Um we'll do up to three replies per lead. Uh and then every single response I get, um I want to actually get notified. Okay. So I'm going to save this like that. Okay.
And so now our AI autoresponder is on. And so we are now in a point where our email and our LinkedIn outbound sequences are fully set up. So we could let this run. let our AI autoresponder handle all of those responses and just sit back, kick our feet up, and watch meetings get booked to our calendar. Or we can take this another step. Okay, and this is where we're actually going to add in our Vappy Voice agent. Okay. So, again, the idea here is when somebody responds uh to our outreach here, we're going to wait about 60 seconds to 2 minutes and then we're going to actually give them a live call to their phone number from our AI voice agent, uh, which is basically going to ask them a couple questions and then try to book them to a meeting. Okay?
All right, guys. I'm just going to slide in here really quickly and let you know that if you want to start building these really cool automated multi-channel outbound sequences to start booking meetings for you or your clients on autopilot, get down in the description below. Uh I will have Victoria AI linked down there for you where you can go sign up and start setting up these systems for yourself. Um and if you guys need any help building more complex systems, more lead genen systems, uh please also get in touch with us down below. We'd love to help you or your clients book more meetings on autopilot. That's what we're here to help you with. Anyways guys, thanks so much. Let's get back to the video.
All right, so what I'm going to do next is I'm going to come into Vappy. All right, and we're just going to set up a voice agent here. Um that is essentially going to uh call uh the lead that just responded positively um where we can book them to a meeting. Okay, now I'm not going to get into the weeds here of how we set up a Vappy voice agent. I have a bunch of videos on this that I'll have linked up above so you can go watch it. But essentially what I've done here is I've dropped in kind of a basic script um kind of telling it what to do. So hey, your name is Paige. You're calling someone who recently replied on LinkedIn or email expressing interest in our AI SDR platform. The goal is to thank them, confirm their phone number, and then send them a Cali link uh via text so they can book a meeting. Okay. Um and then I basically prompt it. Okay. So I say, "Hey, this is what you should say. Um hey, first name, thanks so much for applying and for your interest in AISDR. I just wanted to make it easy to set up a quick intro call with your team. I have your phone number listed as phone number and keep these two fields in mind. This is very important. We'll we'll talk about these in a second. Um we're essentially going to feed these variables into our voice agent for every single lead. All right. So, I have your number listed as blank. Is that still the best number to text uh you a link to book a time? Okay. If yes, perfect. I'll send you our colony link. If no, got it. What's the best number? Uh, and then we will basically text you to that number. Okay. Um, and that's it. Okay. So, they get a custom personalized call like 60 seconds after they respond. Uh, so it sounds and feels like this is like an unbelievably personalized process. Uh, the lead is super warm. Um, and we're going to book them to a meeting. Okay. So, that is that.
Now, how are we going to connect Victoria to Vappy? What we're going to do is in Victoria, we're going to come. We're in our um response tab here. We're going to come down to our response hook section here. Okay. And we're going to enable response web hook. All right. And so basically what this is going to do is every single time somebody responds positively, uh we're going to actually fire a web hook with all the details of that response to make.com. You could also do this in naden if you wanted or zap year. Uh we're going to use make today. Okay.
So I'm going to come into make. Um, I have a new scenario here called Victoria AI lead genen. And what I'm going to do is I'm going to drop in a web hooks um module. All right. I'm going to add a hook and it's just going to be called uh voice agent trigger. I'll hit save. I'm going to grab this web hook URL and I'm going to come back to Victoria. I'm going to drop this in right here. Okay. I'm going to save this. Okay. So, I want to test this now. And so, what I'm going to do is I'm going to come to make. I'm just going to run our web hook here. I'm going to come back to Victoria. I'm just going to test this web hook here. You can see that was successful. Then if I come back into make, you can see that we're passing all this data into make here about our lead that responded. Okay, so they just pretend they just responded on LinkedIn. Uh we can see who responded. Uh we can see their email, all their information, and then these custom fields. Okay, now these custom fields here is where a phone number is going to show up. Okay, because this is a test web hook, um you can't see that, but whenever somebody responds, they're going to be responding as a custom field um with that company phone number variable. Okay, so that is looking good.
What do we do from here? Okay, so I think what we do next is we are going to put in a little delay. Okay, so we're going to sleep for 60 seconds. All right, so somebody responds on LinkedIn or email. um we immediately send a web hook that they're interested. Uh and then we're going to sleep for 60 seconds. So we're going to wait 1 minute. Okay. And the next thing we're going to do is we're going to drop in a Vappy module. Okay. And we're going to make an API call. Okay. We're going to post. All right. Our relative URL is going to be call. All right. Content type application JSON. And then body we're going to make our API call. Okay. And I will zoom in here so you can see this. Right? And this template is going to be linked down below in my free school community. So you don't need to worry about any of this. Um I'll have this all linked where you can just download it and upload it into your account. All right. But what we need to do is essentially give um Vappy the assistant ID that we're going to actually be making this call with. Okay. So that is going to be this ID here. So I'm going to copy this ID. I'm going to drop that right here into my agent. Okay. Then we got to give it a phone number to call from. Okay, so if you ever used VP before, you can come to phone numbers. Um, and we can just go copy a phone number here. Okay, so I've got a couple couple dummy ones here. So we're going to paste that in. So that's just saying, hey, we're going to make the call from this phone number and we're going to actually use this voice agent to, you know, be our speaker. Okay. Okay.
The next thing we do is we need to drop in the actual phone number of the lead that we're calling. Okay. Okay. So, if you remember back in Victoria, uh, we actually mapped our company phone number field here. All right. And it's called company phone number. Okay. And so what we need to do here is we need to map this field. And if you remember when we did our test web hook, um, we got kind of some dummy custom fields here. But, uh, that company phone number field isn't coming out of our test web hook. It'll come in when somebody actually responds. And so what we can do though is we can kind of like copy this like this. We can come to like you know a a Chrome um URL here up at the top and then we can just change this to our the name of our variable. And so whenever it comes through actually um this will actually trigger. Okay. And so you can see now I've got that that mapped um just like that. So whenever somebody actually responds and we send that actual company phone number variable to our custom field, we'll be pulling that in here just like this. Okay. So that's perfect. All right. So now we've given it the ID of the assistant that we're calling from. We've given it the phone number that we're calling from. And we've given it the number that we're calling. Perfect.
Now we have a couple variables here that we need to populate as well. So, if you remember back in our Vappy prompt, I will go pull up our assistant. Remember, we have these two variables here that we want to pull in dynamically uh for every single person that we're calling. Okay, so we have first name. All right, and we have phone number. Okay, so what we're going to do now is in our API call is we're going to actually uh feed those in here to our prompt. So, every single time this assistant calls a lead, uh it knows their name and it knows their phone number. Okay. So, we're going to come back to make into our API call. And this is going to be uh first name. Again, you want it to tie exactly to this. All right. So, that's good. All right. And that first name, we're just going to map from our web hook. First name. All right. And then here, we want to pull in a variable named phone number. Okay? Cuz remember, it's phone number. All right. And again, what we're going to do is we're going to use the exact same field that we just did up here. Okay. So, we're going to call that number and then we're going to ask them on the call, hey, um, this is this is a good number to text you, right? Okay. And that's it, guys. Okay. So, now we're going to hit save. And we're going to hit save on this. And our scenario is set up.
And so, that is it, guys. It is that simple to set up a fully automated, personalized, scalable AI outbound system that books meetings for you on autopilot across email, LinkedIn, and voice. Normally, building something like this in make.com or nen, it's not really possible. Um, it's incredibly difficult to do. But because of the power of Victoria here, uh, as well as this ability to send web hooks out now um, to make.com or nadn, you can start to layer a lot of these systems together um, to really generate uh, really powerful outreach sequences that book you meetings on autopilot. It's absolutely insane.
So anyways guys, thank you so so much for being here. If you guys have any questions or comments, get down to the comments and let me know. I'll get back to you as soon as I can. If you need help setting something like this up uh for your own company or for your client's companies, uh please get in touch with us. Uh I'll have our website linked down below as well where you can get in touch with us. Um if you guys want to see more videos like this in the future, consider subscribing. That'd be awesome. If you like this video, go down and smash that button for me because that would also help us out a lot. Anyways guys, thanks again for being here and I'll see you all in the next video. Peace.