How to 10x your conversion rate with AI

This AI agent is going to help you to boost your conversion rate on all of your inbound leads, and it's also going to help you sort through all of your leads so you can make sure that you're not wasting time on those leads that aren't really going to come of anything. If this is your first time seeing me, my name's Rory Ridgers, and I'm passionate about helping business owners integrate AI into their business to help them increase the amount of leads they get, close more sales, and automate boring and repetitive tasks. So if that's something you might be interested in, then do consider subscribing to my channel and following along my journey. In this video, I'm going to be making an AI agent that's going to take the information from your HubSpot contact forms or any other type of contact forms, and it's going to upload that data into your HubSpot CRM. But before it does that, it's going to be run through a Relevance AI agent, which is going to be able to do research on this lead to get things like their LinkedIn URL, their company LinkedIn URL, give you a summary of their company, give you their company website URL, and also even give you a lead score. So if you're a business owner that gets multiple leads a day, then this will be perfect for you because it will allow you to sort through those leads that, you know, aren't really going to come to anything and allow you to focus on those leads that are actually going to be worth your time. So yeah, let's get into it. Okay, so to begin with, we're going to be in Relevance AI, and I'm going to show you inside the Relevance AI agent so you can see sort of how it works and what's going to be happening. So this is the agent here, it's called Lead Manager. We're going to hit configure so you can see inside it. Okay, so first off, I've just given it a name, Lead Manager. Agent description: takes new leads and updates the CRM with that lead info. The name of the description doesn't really matter. At this point, but if you do come to use multiple agent systems, then the agent name and the agent description will matter because when your agent's going to be talking to each other, they're going to need to know what agent they're talking to. But for this video, don't worry about that, just put whatever you like in the name and description box. So then we're going to come over to tools. I'm going to show you the tools that we're using. Okay, so in the tool section, as you can see, we have Lead Researcher Tool, Simple Lead Scoring Tool, and Update CRM with Lead Info. Now at the moment, these tools are going to seem quite self-explanatory, but later we're going to go into a bit more detail on how these work. So the Lead Researcher tool, as you would think, just researches the lead as it comes in to get things like their LinkedIn, their company website, their company summary, all things like that. The Simple Lead Scoring tool is just going to score the lead and give you a couple other data points. And then the Update CRM with Lead Info is just going to take all of that information and then update it into our HubSpot CRM. So that's what we've got in the tool section. And if we come over to instructions, you can just see quite simple. I've just got role, task, and specific section. In the role, I've just said you're an expert lead manager and your task is to help users to get more information on their leads by using a selection of tools and updating the HubSpot with that information. Very simple. So then in task, I've just said your trigger is when a new lead comes in from the user's HubSpot and you will receive the lead's name, email address, and HubSpot contact ID. If If you're not sure what the HubSpot contact ID part is, then I'll explain that in a minute. And I've said follow this flow when you get a new lead come in. So first it's going to research the lead and company, score the lead, and then update the CRM with that new info. Simple. So that's just it, using the tools each step of the process. And in specifics, I've just said it's imperative that you always start with the lead researcher tool and always finish with the update CRM with lead info tool, just so it knows where it should start and where it should end. It just helps it a little bit. And then I've just said if you can't find any information, make sure that you just leave that area blank. But if you can find the information, then always do so. So that's the tools, the core instructions. And then if we just have a look in the Flow Builder— so in the Flow Builder here, you'll see that it's kind of what I put in the task section, but it's just with a little bit more information. So you've got the starting point, then the first instruction is when a new HubSpot contact is created, you'll be triggered via API and you'll use the tool lead researcher tool. And now in the flow builder, if you're writing this out yourself, if you want to reference a tool, all you're going to do is hit forward slash and then you'll get a list of your tools. So the second instruction is then generate a lead score using this tool, Simple Lead Scoring Tool, and then take all of those outputs and update the CRM with this tool, Update CRM with Lead Info. So very simple. What I'm going to do now is just show you a little bit more inside of those tools so you can see what's going on. Okay, so we're in the first tool. It's called Lead Researcher Tool, and as you can see, the user inputs to start off with are just lead name and lead email, and it will get this from the HubSpot contact form. And if you're worried about how we're going to integrate the HubSpot contact form to be able to talk to the Relevance AI agent, then don't worry, I am going to show you how to do that in a minute. I'm just going to run through the tools first. So after the inputs, we've got an LLM step, and what it's going to do is it's going to extract the website from the lead email. So I've just put an example here. I've said, for example, Rory@autonomouslyai.com, the output should be www.autonomously-ai.com. So quite simple. It's going to save that variable as company website, and then we're going to do a Google search for the LinkedIn of that company website. So we're going to use the company website variable that we got here and just put in LinkedIn. It's going to do a Google search for that LinkedIn. So we've got another LLM step here, and what that's going to do is it's going to take all of the outputs from this Google search here, and it's going to try and identify the most likely result for that company's LinkedIn, because the Google search is just going to list a load of LinkedIn URLs that it could possibly be. So this LLM step, just going to find the one that's most likely the one that we're looking for. So if you like, you can pause the video and look at the prompt I've put. It's very simple: task and role specifications, and then just some examples. And now we're going to want to get the LinkedIn company's profile and posts. So we're just going to put down this block called get a LinkedIn profile slash company and post. And we're going to put the URL from here. And now that's just going to get the company profile. So it's going to get all of the information from that company's LinkedIn. And now we've got another LLM step here. And what that's going to do is it's going to take all of that LinkedIn data and it's just going to output the number of employees, the company summary, and the company industry. But something that we got to do here is we want to make sure that it's outputted as JSON. And I'll show you why, because after we have this step called convert string to JSON. And what it's going to do pretty much is when it exports it as JSON, it's going to be as JSON, so in code, and it's going to be one output. But we want it as three separate outputs. So what we do is we ask it to do it in JSON and then convert string to JSON. And then what that's going to do is going to output as three separate outputs rather than just one, if that makes sense. I'll do an example run now to show you what I mean. I'm going to say lead name is going to be Ben Francis, and then lead email, ben@gymshark. So it's just an example. This is a CEO of Gymshark. It's just the first thing that came to my mind. We're going to run it and see what happens. Okay, so down here you can see we've got the company variable output for the number of employees, 1,700. Then we've got the company summary, company industry, company LinkedIn profile, company website answer, and that's it. So yeah, those are all right. I'll show you what I meant a second ago about the converting string to JSON so you'll be able to see it a bit better. So as you can see, the LLM step has outputted the results: number of employees, company summary, and company industry. And now if we didn't get it to output that as JSON it would output it normally, but it would just be as one output, and we don't want that. So we ask it to output it as JSON, and then this convert string to JSON step, as you can see, it just separates those outputs into 3 separate variables. So I hope this makes sense. So we're just going to hit save on that, come out of it, and then for the next step, we've got the simple lead scoring tool. And now this is already a pre-built tool in Relevance AI, so you can just search up simple lead scoring tool and it will come up, but I'm just gonna walk you through what it does. The only user input it requires is the email address. This is just the placeholder one that was already in there. Then as you can see, it does another Google search for the prospect's LinkedIn, and then we've got another LLM step just to determine the most likely LinkedIn URL again. And we do that so that we can get the LinkedIn profile, company, and posts again in this tool. So as you can see, yeah, it runs that step again, get a LinkedIn profile slash company and posts, and then just another LLM step that's going to output just the email address, job title, company name, company industry, country, number of employees, and then a lead score. And as you can see, this is how it calculates the lead score. It does it based on the number of employees. And now you can change this, you can do it on other variables. For instance, you could say you could do it depending on the country they're in. So if they're in your country, then the lead score is going to be 1,000, but then the further they get away from your country, the lead score decreases. So it doesn't have to just be employees, but this is just what it's got in at the moment. and I think it's a pretty good way of measuring it. But yeah, you can change it depending on what you'd see as a better or worse lead. So that's going to output that as JSON as well. And then as we've seen before, construct— convert that to JSON, and it's going to do job title, company name, company industry, country, number of employees, email address, and lead score. Okay, so we're just going to run this tool with the same input as before, ben@gymshark.com, just to keep it the same. We're going to run all and let's see what happens. Okay, so it's finished running and we've got job title as Chief Executive Officer, company name Gymshark, industry retail, country UK, number of employees— it's given us a range here— email address ben@gymshark, and then the lead score is giving us 25. Okay, so perfect, that's running fine. So make sure to hit save and come out of it. And now we have our third and final tool, the Update CRM with Lead Info tool. So as you can see, there is quite a lot of variables for this one. And what all of those are is pretty much just the outputs from the previous tools. So the ones that we've already seen, and as you can see, all of these are optional, and the only one that's required is the HubSpot contact ID. And that's because for this tool to run, it needs to have the HubSpot contact ID, otherwise it won't be able to run. But these other variables here, sometimes it can't collect them, so we've put them as optional because if it's required and it can't collect, for instance, the website URL for whatever reason, then we don't want the whole tool to just not run. So we've put them all as optional and only the HubSpot contact ID as required. And then all we have here is just, we've got the user inputs and then the only step is a HubSpot API call. And for this, the method is gonna be PATCH. You're gonna wanna connect your HubSpot account to this. So I've gone and done that. And then the path, this is always gonna be the path CRM/v3/objects/contacts, and then the HubSpot ID as a variable. And then this is what our body is going to look like. So this is the variables from our Relevance AI tools, and then the headers that are going to be in our HubSpot CRM, if that makes sense. So for job title, that's going to go into the job title header. Company website is going to go into the website header. Company LinkedIn is going to go on to the company LinkedIn header, so on and so forth. So the tools are quite simple setups. That's all of them. And now I'm going to show you how to integrate the HubSpot forms with this AI agent so that it actually works. Okay, so now we're back in our AI agent settings, and as you can see down here, there's the integrations bit, and there is already HubSpot integration here, but we're not going to do that just because I want to show you how to integrate this with other forms and other CRMs. So you'll be able to integrate different softwares with Relevance AI apart from just HubSpot. So you're going to click on API As you can see here, it gives you a bunch of information and I'm going to show you where to put all of that. So to connect the softwares, we're going to be using Make.com. And as you can see, it's a very simple setup. What we've got, all it is is HubSpot CRM node and a set on watch contacts and then a HTTP Make a Request node. So super simple. The watch contacts node, all you've got to do is just connect your HubSpot and then output properties, just go select all. And then limit is 1 because we only want to be doing 1 at a time. Hit OK. And then the HTTP Make a Request model, I'm going to show you what that looks like in there. Okay, so to begin with, the URL, that's going to be this URL, the endpoint URL. So you can just copy that and paste it into here. The method for Relevance AI is always going to be POST, so make sure to put that on POST. And for headers, you're going to want to make 2. So to make a header, you just click Add Header. So I already got my two in here, and what you want to do for the first header, come down to Sample Curl, and as you can see, content-type is going to go into the name, and the value is going to be application/json. So we put that in there, and then for the second header, it's going to be authorization, and as you can see, it says your API key here. So what we're going to do is we're going to type in authorization here And for the value, it needs to be our API key. And how we're going to generate that is by just coming up just underneath the endpoint URL. You see here, generate API key. Just going to hit that and then paste it into there. And then body type, we're going to go raw. Content type, JSON application/json. And then in the request content, what we need to do It's come back over to Relevance AI. You can see we get a request body here. You're just going to want to copy that, put it into here, and then you're going to need to add a few variables in. So I'm going to show you how to do that. So we have message, role, and content here. So the message and role we can keep the same— message, role, user, content. Then for the content, what we need is the HubSpot ID, the email, and the lead name, because this is what we need to send to the Relevance AI tool. So all you simply do is type in HubSpot ID and then a colon, and then you put the ID variable in from the HubSpot CRM. So yeah, it should look like this. And then it's going to do the exact same thing for email— get the email variable from the HubSpot CRM Watch Contacts node— and then the exact same thing with lead name. And then pass response, you're just going to hit yes. Okay, okay. So now your AI agent on Relevance AI should be connected to your HubSpot CRM. So what we can do is we can just quickly test it out. So what we can do is we can come over to HubSpot. And now I have just made an account for this video. I don't use HubSpot, so apologies if there's anything that might look different to yours on here. So as you can see, at the moment I've got no contacts in here. What we're going to do is marketing, go to forms, and I've just made a really simple form here. And what I'm going to do is click actions, share and then this is the share link. You can also embed the code on your website by just hitting embed code and then copy that code. But for now, I'm just going to click share link, copy that, paste that into my web browser, and then this is a contact form here, so I'm going to fill it out quickly. Okay, so for this example, I've used someone called Nathan Coe, which is the head guy at AutoTrader. And if you don't know what AutoTrader is, it's basically just an app where you can sell used cars. It's just the first thing that came to my mind. and I've managed to get his email address. Was to test it out, you need to have a valid email address, and I need to have a valid email address from a company. So I just decided to use this guy, Nathan Coe. So we're going to hit submit on that. Thanks for submitting the form. I've gone back over to our Make scenario. I'm not sure what's happened to these nodes here, but usually this would be on, and I've turned it off just for the sake of this video so we can run it manually. So I'm going to kick run. So as you can see, it's recognized that there is a new contact in our HubSpot CRM, and it's obviously seen that, recognized it, and sent it to our AI agent on Relevance AI. So if we come over to our AI agent on Relevance AI, we can see that it is actually doing the task now. We can see it working. Okay, so it's done it instantly. So I'm just going to run you through what it did in that process. You can see 3 steps performed in the background. If we click on there, Yeah, so it used the Lead Researcher tool, used the Simple Lead Scoring tool, and then used the Update CRM with Lead Info tool. Perfect. So we can click on these actually to see what it outputted. So this is the input that it got: the ID, the email, and the name. And then as you can see, this is what it's come back with: the LinkedIn, the industry, the company summary, the number of employees, and the website URL. Then we can have a look in this Lead Scoring tool. It's again identified the industry, the name, the country email address, job title, lead score, and again the number of employees. And then if we go into the update CRM tool, as you can see, this is kind of just a lump of code, so it's kind of hard to make sense of, but it has worked because we've got the status code 200. That just means that it's worked. And then if you look here, this is the outputs that is actually outputted into the CRM. We've got the name, email, job title, company name, company website, LinkedIn, size, industry, company summary, lead summary, and the lead score. Okay, perfect. So let's now have a look in our HubSpot CRM and see if it's actually worked. Okay, so as you can see, we're in our HubSpot CRM now, and this is what the contact looks like. We've got the name and the email, which we got from the contact form, and it's giving us a lead score of 50, the job title, LinkedIn account, website URL, company summary, the company LinkedIn URL, and the industry. So imagine how valuable this would be if you're getting a bunch of leads for a day, and you could just scroll through and see the lead score and see which ones are best to use your time on, if that makes sense. And then if you aren't too sure, you can just go through, check out their website, check out their LinkedIn account, check out their company's LinkedIn account. If you haven't got time for any of that, you can just check out the company summary on there. So this is going to save you so much time when it comes to trying to qualify your leads and make sales, and your conversion rate should hopefully shoot through the roof. So there is more that you can do with this AI agent and this automation. You can send emails to these people that have a score over a certain amount and booking calls or booking a demo or something like that. Or you could also create a plan of action for this lead and update your CRM with that. So you can upload the knowledge base of your company, how you do things, what works well with these type of leads. And then the AI agent can generate a custom plan of action for each of these leads, which is gonna save you even more time. But in this video, I didn't do that. I went with the simpler version. But if you are following along with me and you have what I now have, you will be able to add in these tools as you please. Okay, so now if you've made it to the end of the video, you should have a powerful AI agent that's gonna save you wasting hundreds of hours on those leads that don't really go anywhere and let you focus more time on the leads that are actually gonna go somewhere and allow you to close sales. So if you were interested in working with me so that we could integrate AI into your business somehow, so like this, it could save you a bunch of time or it could generate you more leads or it could just automate some of your boring tasks that you have to do every day, then just send me an email. We can book in a free 30-minute consultation. Where we can go over some of the ways that AI could help you and your business. Or you can always check out my website and fill out the contact form on there, whatever you'd prefer. And to everyone else that made it this far in the video, do consider subscribing to my channel. And if you've got any video suggestions that you'd like to see, just put them in the comments and I'll try my best to do them.