I built an AI system to automate my ENTIRE business while I sleep (Free templates)
Rory Ridgers walks through four AI automations that run in his agency: booking-to-CRM sync, meeting notes extraction, YouTube content planning with thumbnail generation, and personalized cold emails. Each uses free tools like Cal.com, HubSpot, Google Sheets, and LLMs to cut manual work.
Chapters
What's up guys, in this video I'm going to be going through 4 automations that I built that I actually use in my business almost every single day. These 4 automations are split into 2 different categories, sales and lead generation, and I'll walk you through all of them and you can see exactly where they fit into these categories and exactly what they do for me in my business. All of these automations are pretty universal as well, which means you can grab the free templates in the description and you can use these automations directly in your business. Just quickly, if you don't know who I am, my name is Rory Rogers and I run an agency called Autonomously AI where we find and eliminate operational bottlenecks in 14 days so business owners can focus working on the business rather than in the business. So if that sounds like something you might be interested in, I'll leave a link to my website as well in the description and go ahead and book up a free call. But let's not waste any more time, let's get into the video and I can show you around the first automation. So this here is the first automation, and this is actually one of the first automations I built for my own business, and it's a vital one that is so, so useful for me. Basically what it does is quite simple. When a new inquiry comes in through Cal.com— so on my website I use Cal.com to take bookings— a new inquiry comes in through Cal.com This system is basically just going to see if they're already in my HubSpot, because sometimes clients and things like that book up calls through my websites. And of course, we don't need to take them through this sequence if they're already a client or I've already talked to them before. So essentially what it does is it checks to see if they're in my CRM. If they aren't, then it's going to assume that it's a brand new inquiry, and it's going to send them an email that's just going to introduce myself and sort of tell them the next steps. And then it's going to create a new contact inside my HubSpot. Create a new deal for this project that we're going to be working on, then it creates a new company, and then it just associates all of them together so they're all bunched together in my CRM. So for all of these, what I'm going to do, I'm going to give you a high-level overview of how they work, then we'll look into it in a bit more depth, and then I'll show you an example of it actually working with a real-life scenario in my business. So like I said, the first step is just a Cal.com trigger, and it's simply just on booking created. And with Cal.com, to set up your credentials, all you do is you put in your API key from Cal.com. Then what we do is we search contacts in my HubSpot. So when I get a new booking from Cal.com, I'll get an email from them, and then we just run that email through my HubSpot to see if they're already in there. So we just return— well, we do contact, and then we do search, and then we're searching for— we're filtering for the email variable. And then we just drag in that email variable, um, from the Cal.com booking. So it's just going to return the contact that has that email. And then what we do is we have a filter node after it and we just say basically if the data from here exists, then it means that they're already in my CRM. So it stops here, which like I said, sometimes clients or other people book up calls from my website and I've already spoken to them before and they're not a new inquiry. So in that case it will stop the scenario. But if they are a new inquiry, so if there's no data that comes back from HubSpot when we search for our email, means they're not in my HubSpot, so it's a new inquiry. So it carries on going through. And then what we do is we just send them over an email. So the to variable is gonna be that same email variable that we use to search for them in HubSpot. So we're sending it to that email that they inquired with. The subject is just Autonomously AI Consultation, and then this is the email they received. It says, hey, and then their name, I've just received your inquiry and wanted to introduce myself. If you have any questions about our upcoming meeting, just reply to this email. I'm looking forward to speaking with you soon. And then just a little sign-off here because they've already booked up a call, so it's not like I actually really need to like do anything. I don't need to email them to, you know, find a time or anything like that. They've already booked up the call. I'm already going to be seeing them in a day or two, so it's kind of just an email to introduce myself. And like it says, if they have any questions about the call anything they want to run through before it, um, then they can just email me. And also that way they have my email. So if they're like not going to be able to turn up for the meeting or they're going to be late or something, they now have my email address so they can just email me back. And then what we do is we create a contact in my HubSpot. So we go to Contacts, Create or Update, and then the email address is going to be that same email address variable that we receive. And then we're going to do the company name, the first name, and then the last name. The first name and last name will of course be from the Cal.com booking, and then the company name is also something that is on my Cal.com booking. So when people inquire, they put in their company name and that's going to be filled in here. And then what we do is we create a deal. So we simply just do it at the appointment scheduled, um, deal stage. So in my HubSpot, I have a bunch of different deal stages. Appointment scheduled is just going to be that first one. And then we just add a deal name here, which is just going to be the person's first name and the person's last name. So just their full name in there, just so I remember which deal is which. And then we just also associate that contact to that deal. And then we create a company as well. So it's just going to be Company Operations Create, and then the name is just going to be that company name that we receive from the Cal.com booking. And then what we do is we grab that contact that we just made. So we go to Contact Create or Update. The email is going to be again the email from the Cal.com booking. So this is going to be the contact that we just created. We're going to now grab that, we're going to update it, and we're going to go Associated Company Name or ID, and then we're going to put in the ID from that step before, and then company name is going to be from the Cal.com booking. So essentially we just create a contact, create a deal, and then create a company, and then associate this contact with this new company. So now everything is set up in HubSpot. So to give you an example of how this might look, this is when I got a new booking from Cal.com. It searched my contacts to see if the person's email was in there. It wasn't, so we kept going through here. As you can see, this variable is red because no items were returned from that step before. And then we just edit fields to get the person's first name, as you can see. And then what we do is we just send them over an email. So that email was sent to them and it just said, hey Daniel, and then the email body that I just showed you. Then we created a new contact. You can see all of these are now green because we have their email, their company name, their first name, and their last name. Perfect. And then we create a new deal. As you can see, we've created a new deal. Perfect. And then we've also associated that contact using the contact ID. Then we create a new company with just the company name and then we update that contact that we just made with that company that we just made as well. So now everything is associated. Okay, and now the second automation that I use all the time in my business is this meeting note taker. Essentially what this does is when I have calls with clients or new leads or anything like that, this system will take the meeting notes and it's just going to update my CRM with any relevant information that it might need. So for example, if the new lead or the client or something, I'm on a call with them and they say they actually want to move the project back a bit and they want to start in a week from now, it's going to update my CRM saying that they want to start the project now a week from now, for example. That's the kind of thing that it would be updating. So what we do here is we have a Google Drive trigger at the beginning because for all my meetings I use Google Meet and then I use the inbuilt Gemini note taker, which just means after the call Gemini is going to go through the transcription, it's going to take notes on the call, and then it will put it into a Google Drive folder. So we're just looking at that Google Drive folder when a new piece of data comes in. It obviously means I've just had a call on Google Meet. So what it does is it triggers when a new Google Doc goes into that file or that folder, sorry. And then we just use a get document Google Docs tool. So we just get the information from that document. So basically just the raw data from that document. And then what this agent does here is it does two things. The first thing it does is it finds the details of this person. So from the transcript that we get from the Google Doc, it won't say— it will say the name of the person, but it won't give me like their email address or anything like that. So I won't have something that I can update the CRM with. So I need to have like an exact thing, like an email address that I can go and reference the CRM for to be able to update data in there. So what this agent first of all does is it takes the name, so it'll say my name and then the name of the person I had to Google Meet with, and it's going to look through my email inbox and my Google Calendar to basically find the event or any emails from the person. So the person that I was on a call with it will find, let's say their name is Dave Smith, it will look through my inbox and it will look through my calendar to find anything to do with Dave Smith and it will just output their email address. So likely from the calendar it'll be able to find it every single time because if a meeting has just finished, then it can just check my calendar for the most recent event on there and it should be able to find that person's email because it has their name. So first things first, it's going to find their email address so that I can go into my CRM and update the contact. And then what it's also gonna do is it's going to take the meeting notes and just give me a few sort of bullet point things that it thinks are relevant. It's just going to output those two things. So output the email address and the notes to update this contact. And then we're going to search for the contact. So we're going to use the step before where it outputs the email. We're going to then find the contact in my CRM. And then if it exists or if it doesn't exist— if it doesn't exist, then we're just going to stop here. But if it does exist, then basically it just goes up here and we just convert the request into JSON, and then we use a custom HTTP request to add the meeting notes into that contact. So for, for all my contacts in HubSpot, I just have a variable that's just notes, and in there this is where that data will go. So it's going to find their email address, find any sort of relevant information from those meetings, and then it's just going to update the notes for that contact. So if we have a look in here, it's looking in the meeting recordings folder, which is where all of those meeting recordings go, and then we just get an ID from the step before, and then we do a Google Docs get a document put in the JSON ID, and then we'll just get an output of what that data looks like. And then our note-taker agent here, we've basically just said to it, you're part of an intelligent system that can get key insights from our meeting. You'll be given the call notes from a meeting. Your job is to simply take notes slash action items from here that will be put into my CRM. You also have to find the email address of the person that the call was with. You can do this by using my calendar tool to get recent meetings and also my email tool. Using the names— Rory is my name— to try and find the email address and also the context from the notes. And I've just told it about the Google Calendar tool and the Gmail tool. And then the output, output in a way where the text can be placed into a JSON body and be fine. So then if we look at the tools it has here, it just has a get many messages in Gmail. And in here it can just do whatever it wants. So it can search for that person's name in my inbox, hopefully get some type of email back, and then it'll of course have that email address. It can also get events in my Google Calendar, so it can define the before or after date. So it can just do sort of like the most recent events in my calendar, and then again it will be able to get their email address from there. So from one of these two methods, it'll be able to get their email address. And then what we do in here, we've, we've flicked on require specific output format, and then we've just added a structured output parser where it just outputs the notes and the email as two separate variables. And then we go to search contacts and then we go into contact, operation is search, and we just filter it again like we did in that last automation. We filter the email header like this, and then we put in the email variable from that AI agent before. So we're just basically trying to find this contact in our CRM. And then if this exists, we just have a filter in here. So the JSON ID that we get from the step before, if it doesn't exist, it means they're not in my CRM, so we can't update anything. So it just stops. But if it does exist, it goes up this way, and this is just basically a JSONizer. So all this does is it takes the notes. So from this AI agent step, the notes that it's given us that we want to update our CRM with. We just give it that and we say, please ensure the provided piece of text will be able to slot into a JSON body and not error. Please output only the correct version. It will fit into a body like this. And then the reason we've done that is because we have to do a custom HTTP request to update this variable in HubSpot. So this is what it looks like. We have— this is the URL here. So it's just CRM v3 objects contacts, and then we put in the contact ID so we know which contact we're actually updating. And then for the parameters, well, for like the headers, we don't send any headers because we can just do the predefined credential types. So we just link up our HubSpot app token, which just means every single time it's going to be sent straight to our HubSpot account. So we don't have to put in any API keys or anything like that. We already have our credentials set up. So you just go to predefined credential type. By default, yours will be on none. You go to predefined credential type and then just type in HubSpot. In this case, of course, it's HubSpot. And then you just connect it up like that. And then this is a body that we send over. So properties and then meeting_notes. This is the name of the variable, the contact variable in my HubSpot. And then we simply just put in the data from that step before. So to give you an idea of how this looks, I obviously just finished a meeting here and then we got the information from the meeting. As you can see, this is just an output of the Google Doc. Perfect. And then this AI agent here, it used the Gmail tool and the Google Calendar tool and it found the email. So it found the email, and then it found— well, also outputted the notes on there as well, which is really, really nice. And then we searched the contacts for that email address, found the person in there. So then it went through the switch, they existed, so it went up this way. This just JSONized it, which basically just means JSONized, um, because we're doing a custom HTTP request and the variable here is just slotting into this existing JSON body. Sometimes it can have sort of like illegal, um, characters and things like that. Like there's certain things like, for example, exclamation marks and things like that that can't really go in there there's certain formatting things that it needs to abide by. So we basically just say, this agent, don't worry about that, um, just worry about finding the email and outputting good notes and action items. And then this JSONizer here, just clean through that text and make sure it can be slotted into a JSON body, which is exactly what it does. As you can see, it might look kind of the same, but it just removes any illegal characters and illegal formatting. And then we send off this request, this custom HTTP request, and then you can see this is the response that we got. So it updated the meeting notes, with these meeting notes here. Perfect. Okay, now we're on to the third system, and we're on to the lead generation system. So for lead generation, I have two main sources of lead generation. That is cold outreach, so email outreach, and my main, main source of leads is going to be YouTube. So this YouTube channel that you're watching. So this system I actually did do in another video, so I'm going to link the video to this so you can watch a full in-depth breakdown of exactly how this works. But after building it in that video, I use it all the time now. Essentially what this does is it just generates content ideas for me, which sounds really, really simple, but when you're creating YouTube videos, coming up with ideas for videos and coming up with ideas for good videos as well is, uh, it can be pretty time-consuming because it's pretty easy to just come up with an idea for a YouTube video, but the difference between that should be a good YouTube video and a bad one is obviously quite large. So this system ensures that it's actually good YouTube video ideas that you're coming up with. So I'm just going to show you basically how it works from a high level, and like I said, I'm going to link the full video so you can get a full in-depth breakdown because this is a little bit of a bigger system. So first of all, we just have a form submission here, and in here I put the time range that I want to look for. So it's going to scrape past YouTube videos and give me an output of what I should be doing. And I just put the time range of here, so I could say like a week ago from now, that's the time range that I want to look at for videos that we're sort of scraping and basing our ideas off. And then also if there's anything I want to avoid, so if there's sort of a trend going around in the automation space that I just either can't do or don't want to do, I don't think it fit for my channel, I can put it into this avoid section here and it's just going to be passed on to the agents down the road so it knows to avoid them. And then what we do is we just scrape videos from YouTube here and then we do another HTTP request to YouTube to get all of the metrics. So we first scrape about 30 videos, you can change if you want, but we scrape 30 videos and then we get all of the metrics from them here. So things like views, likes, watch time, that kind of thing. So then we know which videos are actually doing well. And then we have this JavaScript here that basically just calculates actual metrics for all of that. So it's going to do the views per minute, the engagement per minute, the trends, everything like that, so we can determine which videos are actually performing well. And then this section here, all this does is it just loops over and analyzes all of the thumbnail images. So we just get analysis of all of the thumbnail images that are in these videos, and then we put all of that information into the Google Sheets. So like the titles, the authors, the thumbnail analysis, and also like the views per minute, the engagement per minute, everything like that. We put that into a Google Sheet, then that gets passed up to a reporting agent that's just going to go through all of that data and give us a big long report on all the kind of trends that are doing well, what's not doing so well, what kind of things we should focus on. That gets passed to another AI agent that actually gives us a solid video plan. So it's going to give us a title, a description, a prompt for a thumbnail, and also a justification on why we should make that kind of video. And then we're just going to generate an image, so like a thumbnail mockup. And to be fair, I wouldn't actually recommend using these for your thumbnail, but it's good to give like an idea of how the thumbnail could look based on what other people are doing. And it's also something you can use to like send over to your thumbnail editor. And then we just add all of that information into a Google Sheet. So this is a run that I did pretty recently, and I just want to break it down for you. So you can see here the time start was the 22nd of October, is the time range I wanted to do, and there was nothing in the avoid section here. And what we did is it got many videos, so it got 30 videos here, as you can see, as you can see, 30 items. And it get things like the title, the description, the thumbnail URLs, and that's kind of about it. And also the publish time. And then we do another request here just to YouTube again, um, but this time you get all of that data back. So we get 30 items, but then we also get things like the views and the engagement. So if we scroll down, you can see the views favorites, the comments, and the likes, which is perfect. And then we have this JavaScript here that's going to go through all of them. It's going to give us some pretty cool kind of analysis of it. So if we look through here, you can see it gives us the, again, the metrics here. Then it gives us how old the video is in minutes, and then also the views per minute and the engagement per minute. And then it scores all of them. So the score is 10 out of 10, the views per minute score is 7, engagement score is 3, and then the freshness bonus is just 0. And then it tells us that this video is in the 90th percentile, which really, really cool. As you can see, this is really, really helpful and valuable information. So then we loop through each of these, we just extract the variables that we actually want, which are these variables here, and then we analyze the image. So we go through and we look at all of the different thumbnails we just put in that thumbnail variable, and we get an analysis that looks like this. The image is a thumbnail design for a YouTube video. It features the title Vibe Automation in a bold white font with a purple background. Below the title, there are two sections labeled prompt and workflow. So yeah, as you can see, it gives us an output of how all of these thumbnails look. And then we just add all of that information into a Google Sheet. And this is how the Google Sheet looks. So we have a video title, an author, description, thumbnail, thumbnail description, and then the views, views per minute score, comments, likes, engagement score, and then overall score. So we put all that data in. This is how it looks. We get the video title in, the author, the whole description, the thumbnail, the thumbnail description, and then the views, views per minute score, the comments, the likes, the engagement score, and then the overall score. Cool. And then what we do is we shoot that back up to here where we just get that information back from the Google Sheet, and then we just filter out all of the ones that we probably aren't going to actually take any notice of. So we look at the overall scores and we say if it's greater than 6, then we go through and we look at it. If it's not, then we're going to ignore it because it's probably not something they actually want to base our videos off if the video isn't performing very well. That's why it's so vital to have that JavaScript step in there that's going to give us like the views per minute score, the engagement score, so we can actually filter out the ones that are actually doing well. We can analyze them, just look at the numbers and see if they're doing well. It's not like subjective or anything like that. We can objectively say which ones are doing better than others. So from here you can see 15 items were discarded and we kept 15. So we just split it in half and then we gave that to a reporting agent. And in here the prompt is just saying, you're an AI agent that's going to help me create YouTube videos. And then we've said you'll be given a video title, description, thumbnail description, their views per minute, comments, likes, engagement score, and then an overall score Your job is to look at all the ones with high scores and then look at the video descriptions, video title, thumbnail descriptions to output exactly why the video is doing well based on all the information you have. And then we just give it all of that information from the step before. So we aggregate it and then we pass all of that information into here and it gives us a report that looks like this. As you can see, it's pretty long. I'm not gonna read through it, but it's just exactly what it said in the prompt. It's just an overview of why these videos are doing well, what kind of things they're looking at. And why they're doing better than the other videos. And then we pass that to this agent here. We pass this full report and we basically just say, ah, you've given this full report, I want you to output a full video plan based on what is doing well at the moment. As you can see here, you can read through the prompt when you download this full template, but I don't want to bore you by just sitting here and reading out prompts. Um, this is what it gives us. We do require specific output format, by the way, add a structured output parser like we did on the other automation with just title, description, thumbnail prompt, and then justification. And this is what it gives us. So the title: Automate Your YouTube Channel Content Creation with NAN+ AI Video Tools Step-by-Step Tutorial with Templates. This is the description for it. This was the thumbnail prompt. And then here, the justification: This video idea is directly inspired by top-performing videos that excel due to their laser-focused, outcome-oriented content. And then of course it gives us the full justification. And then what we do is we just do OpenAI generate image, and we just give it the thumbnail prompt from that step before. And you can see this is the kind of thing that it gives us. So like I said, it's not something you probably actually use on your YouTube channel, but it gives you a kind of a good idea of what it might look like. And like I said, you can always like send this over to your thumbnail guy, um, for inspiration for them. And then we get that image, we simply upload that to Google Drive so that we're able to get a web view link for it here. As you can see, we get a web view link which when we open this, it's just going to actually show us the image. And then we just add that information to another Google Sheet. So this is inside the same Google Sheet. We have videos and then new vid ideas like that. So videos, new vid idea. And then what we do is we just delete the information from this sheet. That's probably why— probably wondering why it's completely blank even though it ran through. It's because at the end it just deletes the information. That way when it starts up again, it's not going to get confused by having like loads of different videos in there, like the older videos that have already been analyzed. So at the end it just deletes them all. So like I said, if you want a full breakdown of how this works, I'll leave the link to the full video where I break it down and you can see it. And also the free template is going to be available in the description. And then last but not least, we have a very simple classic cold outreach system. So what this system does, it's not a full end-to-end cold outreach system. It just does what in my opinion is the most important and the most time-consuming part, which is personalizing all of your emails. So in this system, this is assuming that we already have some kind of list of leads which you can easily scrape from Apify using Apollo or LinkedIn or anything like that. It's pretty easy to get a long list of leads for fairly cheap, and this system is just going to create a personalized compliment for each of them. Them based on their website content, which like I said, I think this is a massively crucial part of cold outreach. I think if you're doing cold outreach and you're not personalizing them, it's completely useless. And with a system like this, you can do it all pretty much instantly, whereas if you're doing it manually, if you know, like 1,000 leads, it would take you forever. It'll take you weeks to do that, but this does it in a few minutes. So what we have here is just a simple click to execute. So I just run this when I want. It's not on a timer or linked up to a form or anything like that. And we just get the information from the Google Sheet, which is just a long list of leads and their information. We simply loop over each one, we scrape their website, so we get all of their website content back, convert that into Markdown, and then we have a compliment here. So we just pass that information to an AI agent that's going to generate a compliment, and then we humanize it, which is a really nice thing, a nice touch that I like to do. So this just generates the raw compliments, so kind of just the thing to compliment more than the actual compliment. and then this step here will just humanize it and make that actually sound human. And then we simply just put that back into the Google Sheets. So it seems, seemed really, really simple, but this system is massively useful. And I'd say a lot of the, a lot of the heavy lifting here is in these prompts because these output really, really nice compliments. So what I'm going to do is I'm going to show you a past execution of this so you can see. So this ran in 8 minutes and 45 seconds, and it was scraping, uh, 100 leads here. So it's fairly quick. It's definitely way quicker than doing it yourself. So just to show you how it starts off, I'm going to show you this Google Sheet, the information getting here. So we've got first name, last name, email, LinkedIn, title, company, company website, company LinkedIn, industry, employee count, country, tech stack, key information, and then the compliment, because I've already ran through all of these. So we get all of that information back here. So pretty simple. Then we have a loop over items, set the batch size to 1 so it loops over each of them once. And then what we do is we have an edit field step. So we grab the company and the company website URL. And the reason we do that is because with the loop over items node— I have been over this in other videos, but it's something that's probably going to catch you up because it used to catch me up all of the time. Um, after the Loop Over Items node, you can only use the variables from here in the step directly after. So if this step wasn't in here— so let's remove this, let's connect this up— I can only use the variables from here in this node here. So if I wanted to use the company name in here, I wouldn't be able to, right? Which is really, really annoying. So what you do is you just add an Edit Fields afterwards, and that just means the variables are going to stay constant throughout the whole workflow. So it's a really, really simple fix. I do it all the time. So we grab the company name and the company URL, and then we just scrape that for a HTTP request. So we put in just the company URL, as you can see, and we get back this raw HTML. And something to note here as well is I've gone to settings, and on error I've done continue using error output. As you can see, that's why there's two prongs here. That's just because sometimes when you're doing a HTTP request on a website, sometimes it just won't be able to scrape, right? Because I firewalls or anti-bot protection or something like that, or the website might not even exist, and you don't want that to break your entire system. Because if you're looping over things and let's say you're on lead number 47 or something and their website doesn't exist anymore or they've changed their URL and it breaks, that means it's just going to stop the whole system and it's going to stop looping through them, which is a nightmare. So what you want to do is just add this error output here and just say on error, just go back up to this one and it's just going to skip over that lead because we can't even do anything with them because we can't scrape their website, so we can't make them complement. So it's just going to skip that lead and just go to the next one. As you can see, uh, we tried to scrape 102, 10 we couldn't scrape, so it only ended up doing 92. So it's about 10% of them from my batch of leads at least where we couldn't scrape their website. So anyway, for majority of them we get the raw HTML and then we simply convert that to Markdown. So we do HTML to Markdown, put in the data from here, and this just makes it way easier for our LLM to read by just chopping out a lot of the HTML that doesn't need to be in there. And then we have this agent here that's going to generate a compliment. So we've said, you are writing a short personalized compliment for a cold outreach email between founders. You'll be given text from the company's website. Your goal is to reference one specific detail, phrase, or initiative from the text, compliment it naturally as if you personally noticed it while browsing. Keep it under 35 words, avoid generic phrases like I love your site, great mission, etc. I found when you're doing cold outreach, LLMs can just output like the most generic prompts, just like I really like the colors on your website or something like that, which is just terrible there's no point even having that in there. Um, sound genuine and conversational as if you're genuinely impressed, not selling. And then here we've given it some examples of good compliments, which is really, really good for us. Um, examples of good compliments and bad compliments, it really helps with writing things like this. So something like, I like the line on your homepage about turning chaos into clarity, that's exactly how we think about operations too, is something— an example of a good one. Example of bad ones is like, love what you're doing, your site looks amazing, you seem passionate about your work, things like that. And then extra tips to make it sound more human: remove the 'I' from the beginning of the sentences. So rather than 'I really like,' just say 'really liked.' It sounds a little bit more human. And then here you can see we've given it the company name and then the scraped website data, and then it outputs things like this. So if you have a look, I've just picked a random example here. It says something like, 'Really like how you pinpoint broken promises as the biggest sales stoppers. That focus on trust feels so critical and often overlooked in B2B growth strategies.' So you can see it's picked out a specific thing. Obviously on their website it says something about broken promises, so it's found that and it's output to that, which is really nice when you're getting these compliments. I get emails through sometimes with compliments where it's just like, oh, I really like your approach to this, and it's just like so bland and generic. If you actually pick a specific thing from their website, like a tagline or a catchphrase, something they say— so something like they say broken promises— is really going to resonate with them a lot more. And then we pass that to a humanize step, and the prompt in here— I'm not going to read you out the entire prompt again, but we've basically just told to make it sound way more human by giving it some examples. And of course, this template is going to be completely free, so if you want to grab it, read through this prompt, change it, you can go for it. Of course, you can pause the video right now and have a look, but we've given it some examples like before and some rules to just make it sound more human. So I've just picked the exact same output as the step before, and it changed it to, really like how you call out broken promises as the main sales killer. That focus on trash actually feels pretty critical and kind of overlooked in B2B stuff. So you can That alone just sounds way more human than the step before. The one before sounded all right, but this one just sounds way more human, as if a human has just gone onto their website, seen that, and then typed it out. And then we just simply add all that information into a Google Sheet. We go to append or update a row, and then the column to match on— so we need to have a column to match on— is going to be the company. So we select column to match on, do it as company, and then we put in the company name variable. This just means the compliments are going to be going into the right row in Google Sheets. Then we just put the compliment variable into this header here. And then it'll end up looking something like this with all of these compliments in here, um, attached to your leads, ready to import straight into Instalead or something like that.
More from Rory
Dynamic Workflows are crazy. This is how to actually use them (beginner's guide)
