Automating Twitter Growth & Content Strategy with AI | N8n x Open AI automation workflow

0:00let's just go okay so hey Vu how are you I'm good James how are you good you have a lot of screenshots on your desktop there looks like mine it's because of the nation workflow error handling I had to like take it to CH GPT and then ask if that was right or not so yeah that's the thing okay awesome so uh burning question on my mind is it Nation or or inen I think we can pronounce it whatever works inen or Nation I think it's Nation if I'm not wrong I want to I want to say Nation but I keep uh seeing people say inan so I'm like am I saying

0:45this wrong but uh seems like a branding opportunity for them to clarify it yeah looks like it yeah so okay like let's let's dig in here um this is our first sort of like you know Community Showcase if will uh you've been working on a automation for X um you know we have basically grown our whole business off of X um you know we try to publish like super consistently and get high quality stuff out there and I feel like we've learned a lot lately um about sort of what's resonating what's driving engagement things of that nature and had a bunch of DMS of people asking me hey

1:27what's your all all's workflow for for X how do you guys create this type of content so consistently and and stuff like that so you know we were chatting a few days ago and we were like hey let's uh try to start working with X data and see if we can analyze uh you know what our workflow kind of is behind the scenes and if we can sort of like crystallize it and figure out a way for us to get um input from what's resonating out there uh for our own content strategy so that we can grow our audience um you know a couple of the learnings

2:02that I think we've noticed that are playing into this uh one uh the things that really Drive success on X in my mind is like embracing new formats what do you think about that yeah I think the formats play a vital role even the analysis that I did uh sort of like if you know basically like took uh Greg's account to basically do the analysis and when you see there's like a lot of different formats that he strike and it's not just like you know super consistent and it's that new formats that brought out some kind of interesting performances also like attached to it per tweet so I think yeah

2:40when we look at it and also the hooks that you constantly say right how the hook has to be like uh structured and it's not just the hook the format itself right like the entire tweet how are we formatting it even if it's like one pager how are we going to format it uh information by information I think these are uh key criteria that we need to look at and that's very evident in the analysis that we've done as well yeah I think uh a tweet is kind of like the way I think about a tweet is it's kind of like a value delivery mechanism and uh there's certain formats

3:17that work at certain periods in order to like effectively communicate that value and uh the way that you you know speak and and talk about it and structure that format matters terms of like how you get attention build interest and desire and and things of that nature so um yeah I I don't think enough people think deeply about how to really do this and do it right uh there's a lot of work that kind of goes behind the scenes to to do it so that said uh why don't you open up your nation or in aen let's let say Nation uh workflow and um I know that you know

3:58there's a lot here so uh let's start from like a 10,000 foot view um what are people looking at um here to start like tell us about this area of the tool the editor screen and uh what's happening broadly here broadly it's picking up uh the tweets like right now we are analyzing Greg's set of tweets and you know um let's say that we are analyzing 100 or 200 recent tweets we know Twitter works the algorithm keeps changing the format keep changing so it's very important that we do batch by batch let's say like last 100 or 200 tweets and things like that so we take some of

4:39his tweets and we are basically like trying to analyze the engagement over here we're trying to analyze what is the kind of categories that he's spoken about in his tweet what does the format look like is it the image is he like often going for Threads uh is it video what is the format looking like and what are the hooks uh what is he like you know the psychology behind the hook is it like curiosity is it like you know the attention that's driving um and what does the Tweet structure look like and how frequently uh is he like tweeting in a day uh so all of this is basically

5:14like getting analyzed and I'm trying to use uh open AI as a mechanism to like take all the UN like you know the quantitative data or the unstructured data and push that into open AI so that open AI can give me a flush out data like hey this is the analysis these are the takeaway from each of the tweets these are the hook template this is the Tweet structure and things like that so this is the overview of uh the n10 um you know the canvas that you're looking at right now okay so pick a Creator in this case we're using Greg Eisenberg I think he's like probably one

5:52of the best folks on X period he knows the platform better than pretty much anyone I've ever talked to and uh taught me everything I know about utilizing X effectively um so we're trying to sort of capture that Essence and figure out like what is Greg doing uh by you know essentially scraping his tweets as they as they come in and then as you said uh we need to structure that data a little bit so that open AI can look at that data and then give us like actionable let's say insights on what's working what's not by pairing that up with the engagement data that you're able to get

6:28to is that fair yes perfect awesome okay so uh yeah tell tell us what step one is how do we start here like what's the first thing that you had to do with this workflow yeah let me just uh zoom in a bit so the first thing is that like you said uh we need to scrape the uh the tweets right so basically like we we have two ways to go about it one is that you can choose a Creator uh or specifically you can go for keywords when we go with keywords the biggest problem that we have is that it becomes too broad uh you know

7:01randomly when it comes to tweets we see a lot of tweets being there some can be relevant some can be irrelevant so I think going with the Creator or the profile per se is ideal because we have the control of hey this person is in the niche or this is someone who has done really well on Twitter so I think twe scraping those tweets will be ideal that way so we've used appify uh like let me just open this and show so so we've used appify U particular uh API pack over here so that we will be able to like with just the URL of the profile we will

7:39be able to scrape the Tweet ID uh the bookmarks number of bookmarks when was it created how many of them have like liked it what does the text look like uh the language views code tweets and things like that replies retweets all of this and here is where you see the media URL like whether it is image or video and things like that okay okay so appify uh correct me if I'm wrong here just I want to make sure everybody has all the the details about what tools we're using to to get this output so um in Aiden uh you know you can build the workflows you

8:18can integrate tools like appify you can integrate custom code or or whatever but it doesn't provide a way to like go and scrape uh someone's tweets like out of the box so you know to get certain information into inen or Nation uh you have to use some other tools so appify is basically like a library of all these different sort of like scrapers and ways to get data um you know for your I don't know software product or your workflow or your AI model or or whatever and then in this case we uh we have this like X scraper uh one that we're kind of starting with yeah I think we can use

8:57multiple ways to do that one is apify the other is Twitter itself provides the uh the API X API which is default there but if you have a paid subscription you can go for it appify in always is much better because it's like like an app store where you can get all the apis like you want to scrip Twitter Youtube Instagram or anything you have the option to choose which one to go for and it's sort of um uh like you know cost effective and if someone wants to try this out for free they also have this go Google custom search API uh the biggest problem with that will be pagination

9:34because since it's free they'll have to add like five six more steps and for it to get get that right it's going to take a lot of time so appify by all means is like the best way to go about okay cool was it difficult to set up appify within an aen uh or was that pretty straightforward it was very uh straightforward it was like you have to just go to do I open it and show it how that works uh nah I think we can we can just give a overview for people probably or set a bullet points or whatnot but cool so you figured out how to do that and uh I the

10:12next thing I see is like 108 items like what's that mean and like uh once you get the data into the system uh what what are you doing next yeah so like I like I said I've taken 108 or 110 tweets ideally and it is very important one more Pro tip that I would like to give while we scrape tweets on appify is that make sure we are turning off uh replies and uh you know retweets and things like that is turned on because when we say 100 it can pick up all the replies that we did and there's a chance that just one tweet uh the API call will be like

10:48exhausted right so it is important we turn that off so what you see here is that the 108 tweets that have been like picked up recently uh the reason is because every time you give test step or the uh like you know the trigger is set up it'll take the recent set of tweets that the particular creator has posted now these are the number of tweets I've limited uh 100 tweets and you will be able to see like I said these are all like very you're seeing as a table format so this is the Json format so this is a very unstructured data if you directly put them in uh other format you

11:23might not get uh the exact data that you're looking for or if you've noticed there is no Twitter tweet URL and the image or video URL is all like in one particular row it is very difficult for us to like get those information in separate columns and things like that so what we've done is uh the first thing is get all that and put them in uh like tweet text views bookmarks like select specific set of uh data that we want to get added to the Excel we've added all that over here now we don't have set of data like URL we don't have the specific set of dat data

12:03like this particular tweet has whether it is an image URL or a video URL so that's what we are going to do next okay which is get the Medias right separately what are the photo URLs what are the video URLs like this tweet does not have a media but it does have something else so it basically picks up all of it and gives the data now the other thing is we need to get the URL so this has to be done in three different ways uh like one we got the all the data we got the URLs like media URL separately now we are getting the Tweet URL

12:40separately so we have this particular code and this code is generated by chat GPT again all that we need to like say is that hey we have all this information we just need to like extract the Tweet URL separately how uh and it'll basically give us the Json code we can take that copy paste it and we might have to change the API key data over here other than that everything is like pretty uh same okay okay so you figured out how to extract the data structure the data um so what do you do after that we need to merge it right because we got all the data separately and we need to

13:18merge all of that together so now so that we'll have pext URL photo URL engagement data and things like that so we are merging all three together and we're using a merge node over here and now once this is merged so we all we have is like hey there's about five likes there's about 10 likes there's about 100 tweets and things like that but we don't know the Tweet frequency per day we don't know the overall engagement rate and we need to know total number of photos and total number of videos and things like that yeah now these are like calculations that basically you're doing with the data

13:56that you've scraped and structured uh here so you're adding these additional steps to kind of work with that information in the database and give us further outputs uh for whatever we do with this information down the line right exactly so in order for us to take it to AI open AI or gemini or whatever perplexity or whatever model we want to like plug in there to analyze that we just need to do one bit of calculation so that it doesn't just randomly take the data yeah okay so the frequency per day if you see it is taking by date like on 15 there's one on 14 14 14 14 14 so

14:33you basically like see the date this is going to be super helpful for uh chart GPT to analyze that number of tweets per day is it like on 14 there are about four on 15 there are about one so that calculation is done here again the same prompt we have to like put it across saying that calculate this I'm looking for tweet frequency per day uh so this will give you a custom code that you can add so this is the engagement rate the engagement rate is going to be more like what is the engagement uh score or uh rate looking like based on so there's a

15:09formula that I've added over here that I wanted to say that number of views uh divided by uh replies plus retweets plus favorites and quotes and things like that so that is going to be the engagement rate so it gives me the score over here so based on that I'm going to merge these two things and you'll be able to see the date the Tweet text the views bookmarks favorites uh this will be very important because it also gives the time uh stamp of the Tweet getting created y quod and whether it has a photo URL or video URL it's going to share that as well the Tweet

15:48URL you're going to get the engagement rate score and things like that yep okay okay great now we are supposed to add all of this we've merged it we are going to take it and add it to the Excel uh so this is going to be like you'll see all of this data getting added over here give me a yeah now the next thing that I made a mistake that many of us will actually like make mistake is directly connecting this Exel to this basic llm chain or open AI model or they basically like add AI agent over here uh what happens is that if there are about 100 uh tweets

16:36it's not just the 100 tweets it's going to like look into it's going to look at the engagement rate there's about multiple columns and multiple data that is getting added to the uh Google sheet now we need to Loop this so that it actually like process the information uh like let's say by batch so what I have done here is process it by 40 yep so so there about 100 so it is going to process it three times let's say for example so that we basically like get uh this to AI so the AI can analyze okay got it so just to review that part again uh can you can you tell

17:18me again what that why you you want to do this in batches like just yeah broadly why why would one want to do that so basically when we take all the 100 to to AP you know open AI to review it and analyze it it's a lot of data for it to like look into so a it takes a lot of tokens so open AI Works based on the tokens uh right so let's say in a day you can actually like if you in the paid plan you have like 100k tokens that you can use so when we and sometimes there's a chance that when we are pushing all

17:50the 100 tweets at a time along with all the data it'll take minimum of 30 minutes to 40 minutes for it to process and number two let's say you're using the fastest model which is gp4 Turbo instead of GPT 5 turbo right it is very slow because it's going to Shell out all the structured data for all of the tweets so you will exhaust your per day quota of tokens and there's also chance that it is going to like you know do all of this slow secondly um what happens is when you get the data it is not always structured sometimes so you'll have to like uh use a code to structure it so

18:31once again for it to structure you will have to rerun the whole thing which will take another 30 minutes of your time plus more tokens so I think that's one of the reasons it is always better we add a loop uh node in between so we can like the maximum that you can take the loop node is like 50 okay so so if I have taken 40 it still takes time it takes hardly like 5 to 10 minutes for me for overall uh loop to get over over but uh if we go without loop it's going to take like at least like 30 to 40 minutes

19:05how did you find out about this like looping process this seems pretty technical to me you know like uh token utilization and you know efficiency and all that stuff uh how did you like even discover that you need to be doing that was it trial and error did you like prompt GPT and it said hey you should Loop this or something like how did you how did you come about this yeah I think when I went and said that this is what I'm looking for in a workflow CH GPT or Gemini whatever you're using it's going to give you a very straightforward workflow uh like you know this is what

19:38you need to do but it's like you have to keep prompting it over and over again I think after trial and error I I realized that I should have used Loop because uh the Koda got over for the day I I sort of had like you know uh processed it more than six or seven times I was waiting for each run I was waiting for more than yeah 40 minutes per run and things like that was that when we use up all of our open AI tokens in like one day yeah that's that's how I actually like analyzed that realized that you know it's because of this yeah and later

20:12U when I went and asked chpd that why it's taking so much time why this is uh then it gave me two things one is I need to use Loop uh mode node and the two is I was using AI agent so there's a node called AI agent uh so it said it is not necessary for for you to use AI agent node at all at this particular only one function it has to do so it asked me to like go for basic llm chain so that you know the the processing is faster as well as Loop and this will actually like reduce the time of getting the result

20:44out okay got it all right so we figured out the the looping uh mechanism there and now this is kind of like the cool part or the fun part for me is like when we get AI involved in the workflow so tell me what we have here so so uh basically we need to give as much detailed prompt as possible Right like we have all that we need so how what is that we are going to like get in return uh from the AI is very important so I think the prompt plays a vital role over here so now I have a prompt that you're a expert Twitter growth uh and

21:18engagement analysis or something based on the following tweets analyze what works best in terms of the content format timing Etc and it gave me a lot of interesting things like for example I need to in order for it for a open AI to again work faster the other best pro tip is that um when you say that you want to look at the frequency or date and etc etc you just need to like go to this table format and drag and drop this over here so that it becomes easier it get gets context like which data it has to refer to which data set it has to refer

21:56to to get the analysis now you see the Tweet sets over here this is the overall Insight it gave me hey what is the common category so Greg tweets often fall under categories like startup growth person personal productivity entrepreneurship insights related to marketing and technology that perform well it gives me gave me that what is the engagement level so this was like top performing category this is like overall Insight right like you also need to know tweet level uh Insight you also need to know overall Insight which is very important right so and you know the insight's really where the value is too I mean like as a as a marketer you know

22:34you can have all kinds of dashboards and data and and stuff like that but it takes time to comb through that and and find the patterns and find the insights so what we're doing is giving a bunch of this data to an llm and and it's finding those insights and patterns for us so yeah this is kind of that qualitative piece where it's actually looking at it and looking at the structures and reading the tweets and the content and trying to let us know like hey what's going on here and why is it working exactly okay so cool other things that we see over here the coolest part is

23:07this right like all of this are basically based on the tweets that it has it gave me what's exactly working but it gave some actionable recommendations that what we can focus on like if you're someone who wants to be like Greg or who wants to like create tweets uh around the same Niche and things like that so here here are some things that you need to know in terms of content strategy what do you need to like look at uh what are the Tweet formats what is the consistency uh what you need to like focus on all of that is like the actionable uh you know recommendation

23:40that it gave so focus on like educational content like share actionable tips use Frameworks so if you see over here it will say that for particular set of tweets for Greg Frameworks has worked really well for him like you know infographic has worked really well so these things will be like super helpful for someone when they want to start something that these are something that they need to consider for a format for Twitter yeah you know what's interesting here uh you know this tells us what Greg does and then there's still this like qualitative Gap or this like Insight step of okay how do I make

24:16that relevant to me or my Niche or my audience you know so I think a really interesting next step on this workflow uh down the line if if people want uh would be being ble to like merge in our own tweet data and show me the gaps like how can I improve like you know this specific tweet or what's a hook or what's a format or whatever I should be utilizing um it could almost like repurpose or restructure uh your tweets automatically based on whatever it's learning from the Creator that you're using for inspiration that's a that's a huge payoff for a workflow in my mind I swear

24:54I I think it's just one node away right like all that I need to like add is like another node uh over here yeah like from the same this thing from the same output that I get I add another node and then like say that hey these are the tweets analysis or if you want directly from here say that I want to like these are the set of tweets that I have can you like identify gaps or something it's pretty easy for us to like uh get the gaps like the next step and you know one of the things that we've talked a lot about is like uh and you know we've

25:26talked about this on X and and stuff stuff like that but um you know you have to have the the right Logic for your your workflows you know so like thinking through the logic here is such an important step I think it's missed by a lot of people but like once you figure out the logic here for for X I mean this is extensible to other platforms and other channels and other use cases right like let's say we wanted to learn hey how the heck is Greg growing so fast on YouTube right now um you know and we just started a YouTube channel so what can we learn and how can we package up

25:59our content uh in that manner like the packaging uh is so key for like content success and audience growth so being able to identify those gaps not just on X but across channels too would be super super valuable I think it is very much possible in the same workflow all that they have to like do is like what we scraped Twitter as like the first step the next thing from the trigger is they can scrape from YouTube of Greg they can scrape from LinkedIn of Greg uh they can scrape from multiple platforms and like build a whole package the steps are pretty much the same right like Twitter

26:35has different engagement data this has different engagement data all that you need to like do is the steps are like pretty much the logic is pretty much the same yeah the end is even this uh like you know the how we are going to merge it how we are going to get the AI to do what is the prompt that we going to do will be the key part over here yeah yeah cool okay so you know we're adding uh the AI model and it's able to give us you know these insights uh what's after that the other thing is that we also need to do tweet by tweet analysis right

27:07like for instance yeah that's an overall Insight that's great but what patterns can we identify from each of the Tweet like for example what is a takeaway from this particular tweet like what are the top three takeaways that I can take and what does the hook look like what is the hook template look like what does a tweet structure look like and uh what is the average engagement percentage what is the category of each of the tweet and from there you can still like uh build a analysis that whether startup growth is like uh number of tweets are like high or in startup growth if the number of

27:40tweets is high what is the average engagement uh looks like over there so that it also gives some user who's starting new or even for us that uh what C content pillars that we need to focus on I think that's the most crucial step before we even start writing a tweet right so that is the next step that it is going to do so I've also like created another Branch from the loop apart from this because I'm calling this as like a qualitative overall Insight AI Branch because you can only have one prompt if you're going to add another set of prompt problem will be like that is an

28:17overall insight and this is tweet level so you would not get matching data that you're looking for okay okay I see uh can you show yeah show us exactly what you mean here with a with an example yep so now I have all the data from the previous node the Excel sheet right like now all I I'm looking for is what is the category of each of the tweet I need to know the engagement percentage per tweet that I need to know I have a formula set up here um what is the performance rating So based on the engagement rate do you call it like high

28:50performing or average performing or low performing and things like that yeah um so what are the takeaways if it is for a new user so what takeaway do I have to like give from each of the Tweet like share three insights what worked what did not work from here what is hook you're enriching every single tweet with metadata uh essentially that uh you know once you have that metadata attached to each and every tweet you can then do a lot with that in terms of visualization and you know pulling insights that you can quickly uh gather and and stuff too right like hey group The tweets together

29:25that talk about startup growth and have an engaging rate above 5% uh you know things like that once that data is enriched it then extends your ability to to get useful insights exactly and in fact the same uh workflow itself you can add another AI node and then say that uh can you like create a tweet like this like you know like this is a hook template this is the category that I want to play with can you like create a tweet like this and give that what you want to cover in this and you know you can add additionally and it's not just the analysis you will also be able to

29:59like flush out some tweets from here cool okay awesome what's next so the next thing is again when it comes to the AI node we will have unstructured data it's a Json data that it's going to give you out sometimes it'll have a par J Json data so you'll have to basically like structure it again so I have added like a code node here again Char is going to like help you out to remove all the inconsistency so you will be able to see the takeaway the hook hook template uh the Tweet structure and when it was created the day of the week time of the

30:38day and things like that's morning afternoon mid evening and things like that so now this will be getting added to Google sheet but before that one more thing that we need to add so that the entire workflow is complete is this if node uh so we are connecting the end of this Excel sheet the qualitative data to if node I Al the end of this and looping the true back to this uh Excel sheet the the Excel where we got all the data of Twitter tweet right so uh this is very key because for the loop to get completed uh if we are actually like adding this uh if node it is going to be

31:19helpful for the loop to get completed so that's why we are adding the if node as well as the loop node to this entire workflow that's uh that's going to help us Okay cool so now I'll show the the final Excel right like the review part we like you can see aesome um so this has back go back to like column a um okay so this is all the tweets um the URL or the link to the tweet it looks like that's not in there but that's okay yeah the uh so the next column if you go up so we can see the menu on column C this would be the

31:56category the engage Eng agement uh rate uh performance rating tell me what that means so like I said based on the engagement percent for the Tweet it's going to say is high performing and when I I also prompted saying that when you say high performing you need to say that why it is high performing so it gives it says that like you know the the Tweet effectively challenges a common belief and presents a compelling argument uh with strong insights into the industry which likely resonated with many users and one of the the reason is that obvious it has the data of retweets replies and things like that into the

32:31calculation So when you say 30% is the engagement percent and this tweet has in this particular category why do you feel that it is high performing is the output that we are kind of getting okay cool here yeah please please go yeah I was just going to say that that's really helpful if you can kind of take us through each of these uh in in this sheet would be great yeah so if you see the takeaway Also let's say the Tweet is uh like longer or shorter whatever it is now we need to know what are the takeaways that this particular tweet has because this is also going to be helpful

33:05when you write a tweet that what should the takeaway be for the end user otherwise the Tweet does not make sense so it is challenging common beliefs there's a focus on valuable insights rather than just information engage with current trends at Technologies to make your content relevant so this is going to be very generic because I've not asked something like hey in this particular thing what exactly so again the specificity of the prompt will give us a better data even when it comes to takea away or a hook right right like if you know if you wanted to expand on that takeaway and make it like relevant to

33:39your content or your product or the value proposition that you're focused on uh you know that would help hey what's you know what what's the takeaway or the payoff that the end uh user is going to get like I I can Envision where like let's say you're merging in like the boring marketer account data and you're like uh you know digging into the takeaway a little bit more to you know am I saving time am I helping people save money am I helping them be more efficient how do I categorize those levers and then like which takeaways are actually like providing uh a high engagement rate and

34:14stuff so that that's kind of interesting yep okay the other thing that was so cool for me is this hook part the hook template and tweet structure tweet structure template will be pretty much the same based on like you know uh what sort of things that they are adding but here if you see who creates curiosity by contradicting a widely held perception which is deep seek has just proved the worthless GPD rapper startups are actually the ones with real modes so it takes the first line of the tweet and it says why this hook why this is a great hook or something if you see here again uh

34:51interesting historical perspective generating curiosity about how this relates to the present so iide the first line and says what this is all about so it gives an idea that I want to implement something like this so you don't have to like about the hook yeah if people could just write better hooks on X their growth and success there would like 5x overnight you know like I I think that that's like the one of the biggest learnings we've had is like you know having that super strong hook on a on a post is so clutch and I think that extends into other channels and things like that too you think about ad

35:29creative you know you need that strong hook like giving people a reason why they need to pay attention you need to grab their their attention so that they can engage with the content um so just that alone to me is like you know worth this entire workflow is identifying good hooks um you know that in and of itself is super valuable yeah the other thing also is that thread I mean like whether it's a thread whether it's a single tweet whether it it so that's the format right like let's say structure allows for deep dive into complex topic keeping readers engaged through multiple points so it

36:05says that and it says there's a hook there's a problem statement there's a personal experience there's a resolution so this is how entire tweet is structured someone wants to write a tweet like write a thread they will know what Greg has done and they can actually like simply duplicate it and then like go about it yeah you know what's crazy about like how Greg does this too is that he he has just kind of internalized this you know I he he just knows how to do this like so instinctively well after how many reps I think he's has over 50,000 tweets on his account or something like that uh when I

36:48when I looked the other day I don't quote me on that but like just the the Reps of going through this so many times he's internalized and learned it it's just second nature and instinct to him when he writes or creates something you know uh which I think that's so cool um uh that he's been able to do that you know like he's he's got a problem statement a personal experience and a resolution these are things we know as marketers that we need to have in our content or in our ads or whatever but Greg just does this which I I think is uh is pretty incredible and like he's

37:22he's the best one of the best in the world at this so um you know just being able to see and kind of validate that these are the things he's doing is super cool yeah and in fact you remember the challenge that I faced when I started out Twitter right um I think many of the challenges were answered basically because of certain tweets that Greg has shared and when when when someone like me who I I wouldn't call myself like I know Twitter so much but to some extent if I have understood how this Twitter works I think I have to like credit it to Greg's tweets and one of the reasons

37:57why you know we took his tweets is because it's easy for us to like break it down and it's easy for us to like duplicate it and take this thing forward yeah I think that's also a good sort of like um you know topic or or transition is like um how how does this workflow or how will this workflow help you you know in in your day-to-day role for example so um just to give people some context uh you know Vu is basically like a growth manager at boring marketing so she helped me with all kinds of stuff like getting our community launch working on our emails like thinking

38:32through our growth strategy um helping uh with content research and and things of that nature so um you know that's a very manual kind of like process uh you have to spend a lot of time digging through data to understand the patterns you know you have like lists and bookmarks that you're going through and everything so how do you envision something like this like workflows helping you in your day dayto day and uh what kind of unlock is this for you and and how do you think about it in terms of like you know the leverage it gives you in your in your work and stuff so

39:06just as a fellow marketer how does how does this help you I think for me uh like I said I'm going to be adding another node for the Tweet creation process at this point of time because like when we have to create some tweet uh the biggest challenge that I have at least is like what you said Hook and the how am I going to structure the entire tweet it's not the cont pillar because we have it all defined in things we also have a calendar in place now I need to like convert all of this into a tweet in itself right so I think I'm going to use this particular uh

39:39workflow for that reason and uh let's say that I'm going to add another notes saying that help me with this is the hook this is something that I'm looking at what will be the like I want a tweet generated from here I think that's something that I'm going to immediately like uh take use of uh cook template and the Tweet structure template uh based on this tweet yeah that's something that would be super useful for me yeah and I think we'll we'll get to this now but uh you also can pipe this data basically into looker studio and visualize it and and see it so I know this is kind of an

40:17early version like of of this but if you if you go to the looker dashboard I think this is pretty cool just to give people an idea of uh one like how can I I get you know if you look at a workflow in in Aiden and even a a spreadsheet or a database there's a lot there you know so let's say that um you know you wanted to communicate something to a coworker or a client or a customer or something along those lines being able to sort of package up the insights and the data is uh is really cool and uh that's what you've done here with with looker

40:54Studio yep so ideally uh what exactly is like working imagine like imagine if it is not Greg or someone who's like in the same space right uh and especially like whoever is going to set up the workflow the biggest unlock over here is that what category is like highly being like talked by that person I think that actually adds insane value in itself because do I need to like talk about the AI Trends do I need to like talk about productivity what are the first five categories I'm going to pick up for my Twitter to basically get going right I think that is going to be

41:31like super helpful secondly the biggest unlock is this engagement perc right like I have the category now what particular category is sort of having higher engagement which means it's it's another great additional Insight that I can have that I'm going to Cate I'm going to like take a tweet of this category and also like make sure that see which particular engage like categories and getting more engagement let's say AI Trends and cryp cryptocurrency how am I going to like Club these both so that the engagement is also insane and I'm also like able to like build a tweet around it and other thing is frequency because like you

42:08always say James I think it's Twitter is a very fast platform right you just can't like just post it and expect everything to work on its way on you have to like go reply you need to identify people who are doing like really good in the particular Niche and things like that how often are they like re re sh retweeting quoting uh identifying something interesting and sharing it out there I think a very small uh tweet can actually like do a great uh job in terms of impression and engagement right so how many if if it's like today how many like tweets do I need to like go and like create uh does

42:43it change does it like you know do I need to like do twice or Thrice or something of that sort so this is going to like give me that idea so that I can plan my calendar or I can plan my Twitter accordingly yep yep cool the last thing is this because I have the hook template all of that like set up here I can actually like go and like you know click and then like take use of it or something but I would prefer it like you know with a chat node and take it forward from there yeah yeah looking at this uh like kind of early example we've

43:15done here gives me a few different ideas I think eventually you could you know have sort of a menu where you're able to kind of go between various creators I think you could roll roll up all that information into an overall kind of dashboard and then I think like um having a a bigger table for the hook template tweet structures or whatever that has like the engagement data merged there where I can sort the tweets by key metrics and then I can quickly identify the the best hook template the best category things like that and have the data right there um I think that would be really cool and then yeah I

43:56think like the final pay off is kind of like how do I take all this information and like get something very useful out of it quickly so it's like hey like rewrite my tweets or repurpose my tweets and then that's an a table for you or it's a chat-based uh ability to to say hey here's a topic I want to write about uh you know I want to write about um marketing for startups and uh you know it needs to be based around Ai and that it knows between all the cre ators I'm analyzing and collecting data on how to structure that in my voice and then it

44:32can pump up pump out tons of drafts or you know potential things to uh interact with or look at on a calendar so yeah like having that calendar view at the end of the day being able to look between multiple creators and kind of like have a a big picture of you would be super super powerful and um you know a lot of people Greg included you know a lot of them talk about like the future of of SAS and what does it mean when someone can like spin up a tool with cursor or something like that as as a developer well you think about marketers

45:09and you know we have all these subscriptions and and all these tools that we use there's a bunch of ones for x uh for YouTube you know stuff like that but um you know what this type of work enables us to do is to really customize everything for our use case uh to iterate on it to arrive at what we're really trying to get to um so you know if you're using a product or whatever that has 10,000 customers or whatnot like the ability to customize that uh for your use case is difficult um but now we can basically spin up apps for you know use cases uh that we have uh

45:48the day of and get like an immediate valuable payoff from it that we can then go and Implement in our marketing right away that's the power of all this yeah I think I really like uh the calendar view part that you mentioned right like if they can add multiple like that's something that I would also like love to dig into which is adding multiple creators a identify the like overall pattern across all of the the creators including categories including the hooks and stuff like that have like a a resource library of sorts like you know in place and ideally like build a calendar from there like take same

46:31AI oh I think uh veru froze here we'll give her a second so yeah having like a calendar view that enables you to you know build out like a complete content strategy for like your brand or your product or your value proposition and all that's populated I could see that being really useful um I could see uh the AI being able to like score those Concepts or something along those lines so that they could predict performance or help you prioritize which ones uh you should be uh you know talking about on on x given the different Trends and and things of that nature um okay Vu dropped

47:17off there but she is back now hey we lost you there for a second uh I was just kind of riffing on the content calendar piece and uh just saying that you know it'd be cool to be able to like kind of score the different you know Concepts and rank those that I can prioritize them know where to start know what to do yeah I think the the possibilities are like a lot over here Gap analysis calendar right tweet uh so many things we can do from here cool so how can people like in our community spin this up for themselves how can how can they utilize this what can we

47:55provide them uh if they want to play around with it or experiment yeah I uh there is a workflow document uh like a detailed workflow document like what is the node one node two node three all of that should be and I'll also share the Json uh file of it so that you know it's very easy they don't have to like worry too much about the code but they might have to change here and there because there are variables variable metric that will be attached to it but the Crux will be pretty much the same they don't have to like do it from the scratch uh Jason F

48:28workflow document and obviously this uh the video is definitely going to help them I think it's going to save a lot of their time to be honest because I spent a lot of time because of that Loop node uh if node and uh like before I was using the Google custom search API instead of appify so it took a lot of my time so yeah I think they're going to save a lot of time awesome cool well uh thanks for walking us through this workflow I know you've been like working hard on it and all that and it's coming together really well um I think we

49:02should uh you know share this with the community get their thoughts get their feedback and we can keep kind of iterating on this and uh yeah I could see this coming together uh looking at different creators different channels uh customizing it to your brand giving you a calendar to work off of that you can share with your team that also includes those big picture perspectives or insights um so yeah let's keep chipping away and and see if we can get there and I think uh it'll have a great payoff for people for sure looking forward to it James awesome cool thanks Vu I'll see you inside slack see you bye bye