hey guys so in this video I'll show you how to build a no code LinkedIn agent that can save you hours in your content process now this is the first of two videos on LinkedIn Automation in this first one I'll show you how to automate your content process and in the second one I'm going to show you how to automate different kinds of LinkedIn Outreach now the reason I built this agent is because I recently started my own LinkedIn strategy and although it sounds easy just make a post every day uh anyone who actually does LinkedIn knows that it's not that straightforward LinkedIn can be extremely powerful for lead generation but to get results from LinkedIn and to consistently post quality content we need to research the top performing post inside of our Niche we need to constantly come up with new ideas we need to check what other creators in our Niche are doing we need to engage and comment on uh other posts and we of course need to write consistently value driven and engaging LinkedIn posts and this entire process can be extremely timec consuming now the agent I'm going to show you today they streamlines this entire process it has for example real-time access to the top performing Post in your Niche to help you come up with new ideas it can use rag to find similar post based on topics it can help you draft post based on ideas YouTube videos or other content and it can write posts in a tone of voice that actually works for LinkedIn by using F tuned language models and lastly it can also add the written post to your LinkedIn content calendar or directly post them onto LinkedIn so if you want to learn how to save hours on your LinkedIn strategy learn how you can let your agent chat the databases use Rag and find T models keep watching and I'll show you step by step how to set it up and of course the template as always is for free in my free community now if you don't know me yet I'm Ben I've been building AI agents and AI automations for businesses for over a year now and I also run a community with over 1500 members and if you're a business looking to adopt AI into your business you can also book me in for a free call in the description below so I'll first show you of my agent in action then I'll give you an overview of the setup for this system and then I'll give you a detailed breakdown of how you can set it up yourself so as always I build it in relevance Ai and the nice thing about this agent is of course it also helps you in this ideation process which arguably is actually uh the proc the part of the process where you spend the most time so for example uh what I could can do is check what are the top performing posts in my in my Niche find similar posts Etc so for example I could say something like uh check the top performing the top performing posts in my Niche for the last month right and what my agent's doing in the background right now is you can see here first getting the current date of course to check when it's actually the last month and then uh he has access to a database full of uh LinkedIn post scraped LinkedIn post from all the people I'm interested in so so he can query that LinkedIn database and here he can retrieve the most the top performing post right so you can see here this one has 582 likes and here I can actually check it out right so you can see here for example then she make users new AI features are coming to the platform right uh human in the loop step right very popular new feature I guess so that can serve as an idea for my next uh uh LinkedIn post right so I can check them out and added benefit for me too is I don't have to be on LinkedIn to do this research which usually ends up in me uh you know wasting a lot of my time uh in the LinkedIn Rabbit Hole so that's another added benefit but you can see we can check out the best performing post and base some new ideas on this and the nice thing is also this is real time and up to date because I'm scraping these new posts every day so can even say something like uh check check uh the posts uh for today of the people in my Niche right so I can check that you can see in the background it's caring my LinkedIn database again right so in this case only two uh and one of them is mine but you get the idea and then another nice thing uh so first of all is this database to to check basically a database full of all the LinkedIn posts of the people you want to follow or are in your n and second we can also based on an idea if you have an idea you can find similar posts that these people already uh uh posted in in the past about so for example um I could say something like find similar post on uh let's say uh per AI for personalized Outreach right and in this case uh he's actually using another tool which is the find similar post tool and this actually uses rag to Mo to find the most similar post so we're not only adding these post to our database in air table um but we are also adding it to a knowledge base inside a relevance AI which we can then query so we can hear here we have it here are s similar post related to AI for personalized Outreach so we can check out the post right how to use AI to book hundreds of calls right find your customer qualifying your customer find the contact details creating an offer personalizing Outreach right this again uhuh in our experience relevance is more important than personalization so we could take that and of course you know you can imagine this is useful when you're trying to come up with your next post so let's say we have an idea for next post so maybe I want to use that human in the loop which seems to be very popular um but I want to spin it make it make make it different for my Niche with I post a lot of content on AI agents so I could say something like write a post on on um human in the loop steps are vital um to making AI agents uh more reliable right so we can do that and now in the background my my agent is actually using uh the fine tuned LinkedIn post writer tool so these this tool basically has access to my fine tuned LinkedIn uh language models which are trained on my past LinkedIn post and trained on the post of creators that I I like to sort of copy that tone of voice that that style of LinkedIn writing right so it's writing me four different variations so I can check uh see which one I like most make my my uh edits and that's it so you can see here we got variation one human Loop step steps are vited to making agents more reliable they help you verify accuracy ensure quality Prov oversight maintain trust right so you can see that it is sort of that that LinkedIn style of writing right and here we get another variation sometimes takes a human brain to make the right decision to make a more reli reliable always ask yourself when should a human step in to make the decision for example when building a sales agent consider agents draft a proposal sends proposal to a sales rep for review sales rep this ensures quality Etc right variation three agents are powerful but they're not perfect there might miss details that's where the human in the loop approach comes in by integrating human oversight into agent you can ensure accuracy improve decision making enhanced learning right so you can see let's say I like this variation three I can then add it to my LinkedIn content calendar but first I want to make some small adjustments so what I do is add variation three to my uh content calendar and here I buil in a human in the loop step to actually uh make some changes before I add it to my content calendar so here I can sort of read it make my adjustments right misinterpret dat I make decisions right so whatever here can here can make my changes I can also add an image right if I want I'll just put it for example purposes right and that's it then we can click approve and it will be added to my content calendar so right V variation has we successfully added to your LinkedIn content calendar I get check it out here human in the loop for AI so here we got it so you can see the power of this this agent right it's very much a co-pilot agent you're working together with it right but it it can really enhance sort of that that process of content creation and what I personally do is I just sit down two hours a week with my agent to produce five or six of these posts for the whole week right and it really streamlines the whole process a lot a lot more so and I think the real power of this is having that real time access to all of these posts right and there's lots of more use cases for that this agent can also repurpose content by the way it can uh for example repurpose a YouTube video for example this one right uh you see make a post based on this video right it then transcribes the YouTube video and um yeah writes these variations on them right so right Google's building an AI personal assistant I did too here's what it can do for me right manage my calendar hand on my emails if you've seen my previous video did a pretty good job the best part it's 100% voice controls via WhatsApp so uh and then another thing if you want you can also directly post from here to LinkedIn right so you can also here in this case I have it here right post the LinkedIn is waiting for approval right so you can you can choose the type of post right you can still change it here of course same thing you an approval step add in the images or the video right and approve and it was posted directly to LinkedIn now let let me show you the structure and the system overview for this agent because I think it's really interesting especially giving these agents access to this real time data um I think there are many more use cases possible so let me uh get you through the breakdown so I've set this system up of course in relevance Ai and I've used m.com uh to make some of the Integrations a bit easier if you're completely unfamiliar with both platforms I do have a full beginners tutorial also on relevant on my YouTube channel which I make sure to link up here uh both platforms are no code and completely uh free to start out with I'll make sure to link them in the description below too so the way this system is set up is here we have our LinkedIn LinkedIn agent of course which we interact with and our LinkedIn agent basically has uh access to five tools right I put these in two but this is actually one tool so the first one of course is ADD um the LinkedIn content or the written LinkedIn um content to my content calendar right on notion right that's tool one then tool two is post the LinkedIn tool where you can actually post on LinkedIn then the Third tool here is to write the uh LinkedIn post right and this Tool uh basically has access to my fine tunes LinkedIn language model which I trained on my past post and on post that I like from other creators to sort of get that tone of voice and that style of LinkedIn post right and then here's really what I think where the magic happens uh which is he has access to two databases right so we have a LinkedIn post air table database and we have a uh LinkedIn post uh knowledge base inside a relevant SI so I'll show you the air table first um here I have an air table so you can see we basically have here all the LinkedIn posts right and here we have the right creators LinkedIn post URL the amount of likes and comments and the date and if uh if if there's an image the image URL so uh basically we have access here to all of those posts of the people I want to follow right and we basically have the same uh for knowledge based inside of relevant CI which I can maybe show you quickly so here we have the same sort of setup inside of uh relevant AI now you might ask why right why don't don't I just have one because actually we don't always need uh rag or or uh these systems to retrieve data from for example if I want to uh retrieve the most popular post for a specific time frame or Etc just like I showed you in example we could just do a normal database query to retrieve the most liked post uh for a specific time frame from a normal database in this case air table or you can also Imagine just doing a SQL uh database query uh because that will be a lot more efficient than actually doing it with a rag now if I look at the example of finding similar posts that we cannot do of course with a normal database query so in that in that case of course a rag is amazing at sort of finding um similarities in posts and ideas Etc so we don't always need only rag I think combining all of these systems make our agents uh extremely powerful right so in this case right uh uh we have a normal database we have uh a rag system we have F Tube models and we all give it to one agent and of course the real power of this is that these databases are always updated and how do I make sure they are always updated basically we have another tool which is not part of my agent that's triggered every day to scrape all the LinkedIn posts of the people I want to follow and automatically updates it to both of these databases so that's how I make sure that these databases always have the most up-to-date information and my agent also has access to that that most up-to-date information and this is really where I think the future of these agents are heading is of course you can imagine the amount of use cases and the power of your agents having real time access to the most upto-date information inside of businesses this is just one use case but I think really this is where the future of Agents is heading now let me give you a a detailed breakdown of how this agent is set up and then I'll go tool by Tool and through these databases and this Tool uh to show you how you can set this up yourself so let me start with the agent setup right um now not as it's very much a co-pilot agent in this case we're very much instructing it what to do um it's not extremely complex prompt I'm not going to go through it in deta if you want to look at it in detail of course you can uh uh clone this template uh in my free community and you can check it out in detail I'll just go over it very quickly I use the framework I always use right we have the rooll you're an expert LinkedIn assistant for Beno sple right uh specializing in ideation and Linkedin postwriting right objective right you have basically three objective ideation uh content creation and Publishing either to my content calendar or to LinkedIn I give it a little bit of context about me right uh and about the task the bigger picture and an sop now the SOP not that important this in this specific prompt because we're very much instructing it every time what to do right there so there's not a lot of room for errors here so anyway I I've laid it out right if I want to do adiation right content creation by the way I used my AI agent prompt helper tool to help me write this prompt a lot faster it's always also available in my free community uh if you want to use it yourself it will help you write these agent prompts a lot faster and a little bit better so that's it and we have the instructions with some rules right um the tools right always good to give a very good description to your agent of what these tools can do right and how to use them and when to use them right so query LinkedIn database create a table database containing all LinkedIn post and statistics of people active in my space retrieve Bas post based on date range specific Creator stop performing posts and more when you retrieve posts you only report name blah blah blah right so just a description of all the tools he has uh by the way in my system overie I forgot one to which is to get current date right uh and then we have always in the example section of Agents right always give uh an example in this case not that important because again we're sort of very much instructing him what to do but in general good to give an example of uh an sop what he would have to do like a real world scenario so in this case hey I have an idea for LinkedIn post the impact of AI on content creation can you help me draft it right so you would basically tell tell your agent here in this example what would be the step by step in this scenario right so it gets a good idea of how these processes look and you make them a lot more reliable with these examples and in note section you double down I double down a little bit on the rules that's it pretty straightforward I haven't used the flow Builder right flow Builder is a lot more important when you work with autonomous agents right that have to follow a very specific SRP again this is a co-pilot where we instructing it what to do step by step not important to to include the flow Builder here uh I have not added in anything else language model GT40 always always recommended to use the best ones you can also use uh Cloud which I've had good experiences using Cloud for agents too um but yes pretty straightforward for the agent setup um now let me get you through the tools so I'll first show you um how to query your uh air table database or databases in general uh how you can let your agent query those then I'll show you how to uh set up this rag inside a relevant SII and how to give your agents access to that and then I'll show you the LinkedIn scraper that I trigger automatically every day to update these databases with new LinkedIn post and then lastly I'll show you uh the find to LinkedIn post writer who and how I uh gave these fine tuned models inside of a tool uh and lastly I'll show you the poster LinkedIn and query LinkedIn uh sorry post the LinkedIn and the update LinkedIn content calendar tools so let me start with the query LinkedIn database and I'm going to show you um the tool inputs of the one we actually ran right so you can see here the query linked database right and here we can see what our agent actually did when I showed you my example right so um if you don't know yet right uh here we are on the tool in relevance AI right this is the tool name and this is the tool description now very important this tool description basically describes to your agent what this tool does right so you can really you in general you want to be very clear to your agent what these tools can do so very important to fill out these descriptions especially for a more complicated tool like this one where we LinkedIn database we want to be very clear to your agent what this tool does so for example I say this tool can retrieve records from an a table database database populated with LinkedIn post of top creators in my Niche The Columns of the database and I even specify which columns it has access to inside of the database right so and then here we have three inputs for uh our tool right and and in this case we have amount of post retrieve right which is of course it speaks for itself but the same here right this instruction fill out the number of post the user asked to retrieve this is also an instruction or a prompt to your agent on how to fill out this field right uh so in this case uh use five if user didn't specify so I didn't specify in my query when I said retrieve the most uh popular posts right so it used five correctly right now we store these in variables in number which we can then use in the next steps so the second input field is sort by and this sort by it can basically sort it by likes comments or don't sort you can sort the record records retrieved here if the user asked for specific type of post like top performing post use likes or top commented post use comments if not specified by the user then leave empty right again so I make very clear to my agent what he has to do here and basically this allows my database to sort right so if I want the top performing post I want to get the most liked one so that's why I'm sorting based on likes right so you can see I actually use a drop down here so you can either use likes can sort on likes on comments or it can leave it empty right because if I don't want to sort in for example I just want to retrieve the last three post of a specific Creator we don't really need storage we only need to filter based on Creator and date uh then it could just leave it empty so that's why we leave it on optional because if you put it on required you'll have have to fill something out and also when we don't want to sort right so again we save this in a variable right and last one you can see I'm actually really detailed in my prompt for this last one why because it's a little bit more difficult right this is the filter right so you can see this fils will filter the database when retrieving records um in an air table formula you will think step by step through the following process only use it to ask user ask to retrieve filter data from the database otherwise leave the field empty if the user needs a filter write the filter in an air table formula right example Show an example right instructions if the user didn't mention any specific filters leave this empty only input the formula nothing else no explanation no summary nothing but the formula that that's how I make sure that it doesn't put in anything else but the formula because if it does put in anything else it's going to go wrong right and I even again specify which Fields there are in the database which of course he needs to know when he makes these filters and the Creator names in the database right and then I have an example if the user as to retriev top performing post from last week for Ben a right we he puts in this filter right so you can see for my specific one in the example it put in a formula here which is is after date and this is the date of last month right because I asked to retrieve the most popular post for the over the last month right so you can see what it did here and then what do we do with this information right so we store all of these data points in these variables and then unfortunately in relevance AI we don't have a native integration with air table yet so what I'm doing is I'm sending this information to mate.com which does have an native integration with air table and the way I do that is with an API step here in relevance Ai and I select post and then I need a web hook where we can send that information to uh which I'm going to show you in in a second and then here we have the variables we're going to send over to this web hope right so we have the amount of post the sort and the filter right now if this all goes a little bit too fast for you if this is really new to you I do have a full uh tutorial also on how to set up these API steps and how to connect relev AI with make.com uh which I'll make sure to link up here so basically I'm sending this information over to make.com so I can find uh our tool quickly and I can show you the one we just used in example right so here you can see we have our web hook set up in make right which that custom web hook basically gives us web hook URL which we uh use in that tool right which we basically paste in here and that's it right that's how we send that information to make.com right now you can see that information was sent right amount of post sort by likes and the filter right so we store those in variables here in make and then we got to retrieve the records from the error table which is a native integration here so it makes it a lot easier so you can see it retrieved the aor table record so in this case it did five times so that's why we put in an array aggregator to put those five together and then we send it back to relevance AI with a web hook response here so I can show you maybe quickly in more detail you can see here in the aor table we have the formula right which we added in the variable we have a sort right so in this case I I put in uh a router right so in this in this router it's when when we want a sort based on likes or comments right so you can see this filter here is a condition right so if sort equals to likes or comments then this air table module is actually going to sort and this air table module is not going to sort right so that's how it works um so you can see here for example we can run this and you can see what happens and you see we got the five records back here now this might all seem a little bit complicated especially if you're new to this but highly encourage you to figure this out because if you can let your agents give access to these databases there's lots of very powerful use cases uh for this so this is it right now I'll show you quickly uh how the knowledge base inside of relevance AI is set up so here we have to find similar post actually I'm also going to show you through the agent right so that we can see what he did there um no not this one it's this one right so AI for personalized Outreach right so basically all we have here a lot simpler right again I have the tool description this tool does a vector search on LinkedIn post database to find most similar post for a given ID or posts Right add the LinkedIn post idea to find most similar post right pretty straightforward he puts in the idea AI for personalized Outreach we store that in a variable and then here we have the knowledge search step and the knowledge search step is basically rag built in relevance AI this is the nice thing in relevance AI it's already built in uh with most other no code Automation and AI agent Builders we actually have to work with these third party providers but uh in relevance it's built in now is it the best one out there no it's not but in my experience more than good enough for most use cases so here we add the knowledge Source right which is the table I showed you before I can show you uh maybe again right here we have our knowledge right and this is the table right we're doing uh the rag on right so it's the same table as on air table basically but we've only vectorized um the LinkedIn post column right so only this you will only search through these this column right which makes sense because we're looking for an idea based on the text of these posts right so in the query right you can see I've put in the variable of the idea of course and then the search type we in this case use Vector but you can also use keywords right if you have a specific keyword you're looking for in the knowledge uh Source then you could use keyword but in this case vector and then here we have some uh more advanced settings uh right we have the vector field right so uh the name of the field that contains the vector so in this case I put in the column name LinkedIn post right because we only want to look through those right uh then we have the model you can also choose the model uh for the rag right in this case I use this one I'm not going to go over that in detail right now but for example if you use use use another language you can use these multi lingual models uh things like that then we have here output all fields or not right in this case I want to Output all Fields so I can also get an idea of the amount of likes comments these these posts are getting and we can decide how many uh post to retrieve right so that's it we actually have uh another option in relevance CI which I sometimes use which is the advanced knowledge uh retrieval where we can do a few more things right again we have the query but we can do the search type we can do hybrid so instead of vector or keyword only we can do hybrid which basically combines them and for some of the use cases I've noticed hybrids actually working really good and really better than than either vector or keyword uh so in this particular use case I found it works better here with uh a vector search only but definitely worth experimenting with for other use cases right we have the number of results Etc and in this Advanced knowledge retriever we can also sort of transform that data right away before outputting so for example we can give it here a system prompt so in this case for example I could say something like uh you know give a summary of of the most uh of the the LinkedIn posts that are retrieved right so it will then retrieve the data give a summary of each one right away and then I'll put it back to me right not the best use case for this but maybe for you know answering question questions things like this this can be very useful you can see here we have also a few more options in the advanced knowledge retrieval which is temperature um embedding model right uh which we also had in the other one and a chunking strategy right so that's it for the knowledge retrial uh I can show you quickly how it works in practice so if we run this tool right you can see right it retrieves the treat most similar posts right how to use AI to book hundreds of calls right can see they're all sales related right based on the idea so how to actually scrape these LinkedIn profiles every day and make sure they're uploaded to this knowledge base and to the error table I'll show you that tool right now so basically I build a tool here as you can see LinkedIn scraper uh now this tool basically you can see all it has is one input field which is the LinkedIn URLs right and I'll just give you a very quick breakdown there's a little bit of code involved here but don't worry about not understanding I'll give you access to this template of this tool and you don't literally don't have to change anything uh it can work for you like this but basically what it does is it goes over each of the LinkedIn profiles scrapes the LinkedIn posts right and then basically in this code we filter right so we filter to check uh for the new ones every day right and we also filter we sort of transform the data uh into a a data structure right an array of objects that we then use to insert to the uh the knowledge base of relevance AI so you can see insert data to knowledge right and we input the variable here and that basically allows us to get that data inside of that relevant AI knowledge base which we then can use of course to find similar posts and then I do the same thing JavaScript code but then I transform the data into a specific data structure for an API called to aor table right to up upload these new posts to my table too so you can see here I have my air table API step you can do this to in make.com if you want but uh I just did it here in the API step um so yeah that's it uh so this this basically this tool uploads it to both of these databases right now how do I make sure that it gets triggered every day to do this again that's basically what I do inmate.com so I can show you here um let's say scr LinkedIn here we go right so basically what I have is a Google Sheets right with all the creators that I want to scrape post from every day right and this you can I will also make available to you all you have to do is change the types of creators you want to follow and scrap post from every day right and this will be automatically uploaded to your air table and to your knowledge base in relevance AI so that's what it does I can run it once so you can get an idea right you can see it will run it on each on each of the creators and upload new ones so we can actually check if there's any you can see just a new one was added right and I'm going to show you the Google Sheets so just open up quickly the Excel right it's basically just a list of LinkedIn URLs right and you can add as much as you want and it will basically go over each of the ones and scraped them as you can see here scraped eight different ones so that's how it works and that's how we give our agent access to this sort of real time data right now let me show you uh how to actually give that tool where it writes the LinkedIn post based on a fine tuned model right so if we go back to our agent right we can check this step right so in this tool we have a few options right so we have the find find tun LinkedIn post writer tool writes four variations of LinkedIn post for either YouTube repurposing content based on an idea or content based on an uploaded video right so that's sort of my type of content of course YouTube videos short form videos or or ideas right so basically I I tell it what type of post it is right the type of LinkedIn post to write use idea if the post based on an idea use YouTube If it based on a YouTube video link and use video if it is based on a video URL right which is when I directly put a video give a video to my agent right so it is this case he chose idea which is correct of course and add the exact idea post the user suggested right so in this case human in the loop steps are vital to making agents more reliable right again we store this in a variable in this case YouTube link right optional right so it didn't fill it out right video link also optional didn't fill it out and a free template uh CTA link also optional didn't fill it out so that's it and then unfortunately uh you might have seen one of my previous videos where I actually put the fine t model inside a relevance a but it works a lot better actually if we also use make.com right because we can use it there a lot simpler so what what did I do there here I basically put in the same thing right I put in a API step just like I showed before uh but in this case I'm setting it to another flow which I'm going to show you in a second where I connect These Fine two models to now I also built in a filter here which is for the type of post because each of the different type of post has a little bit of a different flow in side of make.com because for example with the YouTube video we first have to transcribe the YouTube video and then put it into theine two models and for the video uh you can see here right so we have type here for idea right we send over the idea to the web hook right we have another API step which is has a condition on type equals YouTube so only when it's a YouTube repurposing one and here we also we will also send a link right over to so we can actually scrape that you YouTube video and then here if it's a video that I uploaded directly to my agent we actually have a convert audio or video to text right step which we do when type equals video and then we do the same we send over the transcript actually here right you can see the transcript of that video to our make the com animation I'm going to show you that one right now so we got here F tune right you can see we have three different ones here I can show you the normal one first so this one is is based on an idea right and as I have four want four variations I basically put in four steps so really simple here we could just use work with a normal open AI API model now the difference is here we can just instead of just choosing it from the drop down because if we choose the model from the drop down here in the open AI that we make we don't have access to our f t models that's why we go on map and there we can add in our own model right now if you have no idea about find tuning yet or never done it before I have a full uh tutorial also on my YouTube channel which I will link up here too um on how to F tune uh these these models based on your tone of voice or other people's tone of voice it's not it sounds very complicated it's not that difficult you can do it in a in an hour or two and it really makes a huge impact in terms of the tone of voice uh for these for these language model so highly recommend uh you do it if you have some time uh anyway I'm not going to explain in detail how to find two models in this video cuz that goes beyond the scope of this video but the name of the your your F two model you can find in the open a platform right at in your dashboard at the fine tune section when you have trained your own model you'll get your model name and you can then input it here to have access to it inside of make right and then I give it a long prompt here on how to also write good good LinkedIn post right and I basically repeated that four times right then I put them all together into one that's why I have a Json here and then I send it back to the web hook response right uh back to relevant say so my agent has access to it so I can show you quickly how this works in practice so we can run for example I'll turn this on so you can see it live right so we can run it here and here you can see inm make.com right the idea was send over that's put into these language models right he fine to LinkedIn mods it goes through all four and there we go now it's send it back so if we go back you can see we got the four different right posts you can see the LinkedIn style right so that's it that's it for the find two Link in post right there if you really want to know how to find tune make sure to check out my other video because I I I highly recommend commend it if you're doing content creation uh more often now let me show you the last two tools which are uh updating my content calendar on notion right so let me go back to my agent and I'm going to show you content calendar uh here we go right again unfortunately relevant doesn't have native integration with notion so I'm sending this data again to uh to uh make. to update my notion right so again title right three word summary of the post human in the loop right you've seen this already me using it right text for the link uh LinkedIn post this is just the entire LinkedIn Post in this case I didn't put in a description here you can see Flags it because there's a human approval step right so still probably good to just tell my agent input the the link written LinkedIn post right so here we have the LinkedIn post type of posts right and image very straightforward we send that all over to the API which we send to make.com right and in make.com we connect it to notion so uh LinkedIn content calendar as you can see we have two types of post here that's why I have a router right no one with a file and one without a file right and what I do here is uh updated database right so if you don't know how to do this you could find your ID of a database right by view database and then here in the URL this first part before the question mark this is your database ID right you copy that and then once you set up the create a database item in notion you fill out that and then these these uh input Fields will pop up automatically here I put in the variables right the title the type and the post and the image URL in this case and in this this one's on uh without uh image URL as you can see so that's it pretty straight forward I think um and then lastly we have of course the post the LinkedIn tool which actually post to LinkedIn directly right now I usually prefer to have it on my content calendar check it every morning put a new post out from there uh instead of doing it directly but you can also do that if you want to so how is this one set up again we have the type of post because we can actually post a video post an image post or a text based post right so you can choose that here your agent right you have the text field here we have the video here which if you upload a video to relevant SII it will automatically make a URL out of it which you can then input there or you can upload an image yourself there because I would always recommend to use the human approval step where you can then add these images if you want again we do the API step to make.com and there we actually posted uh to LinkedIn because we have the LinkedIn modules there um so you can see it's actually turned off but yes you can see we we send it here right we get a file in in the case so this this is actually the one that's being called when it's an image right so it downloads the image from a URL and then posts it to LinkedIn right uh that's it uh these these modules are are built into relevance AI so it's pretty easy to automate the LinkedIn posting if you're interested now that's all for this video uh might be a little bit complicated if you're new to this um but remember the whole the entire template of everything I've shown is for free in my free community so you can check it out in lot with lots more time and in more detail if you want to I highly recommend you do if you're interested in building these AI agents and AI automations for businesses or for yourself because uh knowing how to use rag how to let your agents chat the databases how to uh use fine two models really uh you know allow you to make really powerful AI systems uh so that's it for this video thank you so much for watching if you're still watching and uh uh please like And subscribe if you're still watching and thank you so much hope to see you in the next one