AI-Powered Spreadsheet Automation: 10X Faster with VectorShift

--- type: transcript date: 2025-03-17 youtube_id: SSSmHUIUzmw --- # Transcript: AI-Powered Spreadsheet Automation: 10X Faster with VectorShift more and more people are interested in the idea of leveraging AI with spreadsheet so in this video what I wanted to show you is how you can leverage AI to automate repetitive tasks within a spreadsheet this could be a use case from lead generation to research to content creation and what I'm going to be showing you in this video is a noode way on how you can build out a workflow that leverages all of the latest and greatest state-of-the-art large language models out there and how you can generate different pieces of content for the different cells within your spreadsheet what we're going to do here is I'm going to create a new spreadsheet where I have net New pieces of content based on videos that I've already recorded based on the title and URL that I have from the first spreadsheet here is I want to take that information and I want to get a transcript I want to generate a Blog with some sources I want to generate also a script for an Instagram reel a YouTube short as well as a Tik Tok as a technical content creator one thing that I do find difficult is actually trying to create content across different Platforms in this video what we're going to be leveraging is Vector Shi which is a platform for creating AI apps as well as workflows you can go ahead and try out this platform for free if you haven't already and once you've made an account what we're going to do is we're going to create two different pipelines what we're going to do for our first pipeline is we're going to go over to Integrations and we're going to go to G she and we're going to select read from a sheet so you can go and authenticate if you haven't already and once you've done that we're going to pick the file and in this case our spreadsheet is going to be called developers digest we can click view folder and then it's going to show the two spreadsheets that we had with in here this is going to be the starting off point and what we're going to be doing within this pipeline is feeding these different pieces of information one by one to a child pipeline in this what I'm going to call this is the parent Pipeline and just for demonstration sake I'm going to just show you what this outputs if I just grab the Google Sheets node from the previous input and let's just say I want to see what all the titles look like what we're going to be doing within here is we're going to be passing in the titles as well as the URLs into our sub or child pipeline within here you'll see all of the different titles of our video and that's going to be the starting point of our next pipeline next we can go and we can create a child pipelines and within our child pipeline what we're going to do is we're going to have an input and you can rename your inputs and in this I'm going to call our input URL and then I'm going to have a second input and I'm going to call this title now once we have both of those items the one key piece that's going to do the heavy lifting of most of the logic is going to be the URL so the first thing that we're going to do within here is you can search for different nodes and integration so in this case there is a YouTube integration but if I just show you the Integrations quickly you you can use a ton of different things from Discord HubSpot linear wonderl notion the list really goes on this can be whatever your use case is and you can grab and create different workflows for whatever the task might be what I'm going to do here is I can connect the nodes just like this and that will allow me to leverage the text of the URL field within get transcript here once I have that what I could do is I could add in a new row to a spreadsheet and again we can pick our files I'll search for developers digest and in this case I'll select our after spreadsheet So within our title for instance what I could say is that's going to be the title that we had grabbed from the initial node there for our transcript what I can do here is I can go and I can grab this YouTube node here now that we have that before I continue on what I'm going to do is I'm going to go and deploy this and then within our parent pipeline all that we have to do to activate the child pipeline is we can grab this pipeline node here and I can select our child child pipeline here within our child pipeline we'll see that we have title as well as URL and what I can do within here is if I click this button here to enable list mode this is one of the key pieces that you have to enable is once I have that what I can do is I can select the information that I want to get for each of these inputs and in this case it's going to be the title as well as the URL what this is going to do now that this is within the list mode is we're going to go through that spreadsheet and one by one going to perform this sub pipeline that we had just created now before I run this what I want to do is just remove a handful of different items within here one thing that I'd encourage you to do is when you're setting these up especially when you're going to be incorporating in large language models and different things that will incur a cost with the invocations and the tokens that you pass one thing that I'm going to do is even though I have over 200 items within here is I'm just going to concatenate the list down to six different items now if I go back to our Pipeline and if I go and I click run what it's going to do is it's going to Loop through each row and for each row it's going to pass in the URL as well as the title into the child pipeline that we had just set up if you go over to your spreadsheet here you'll begin to see your information stream in to the spreadsheet in real time here here we see the title and then we have the transcript it's working through that pipeline one by one the great thing with this is as soon as I click run is I could have those 200 plus items as well as all of the predefined cells as well as the actions that I want to have them perform on each different cell you can begin to imagine the different use cases there could be potential research use cases or or imagine you have a list of emails and based on the email you want to research the company as well as the individual and do some sort of lead generation you can do a task like that very easily now that we have this is where it gets a bit more interesting let's say I want to leverage perplexity and for perplexity I want to use this to generate and enrich the transcript whatever I had within my YouTube video to create a blog post I'm going to say you are an expert blogger leverage the YouTube video title and transcript to create an enriched blog post in markdown with various sources and then what I can do within here is I can say title I can go and I can reference our title text and then for the transcript that we got above is I can go and grab the YouTube transcript grab the output and then finally for the model I'll use the sonar pro model so once we have that node what we finally have to do is just connect it to our Google sheet here so if I go over here and I grab that perplexity node and I grab the response what I can do here is once I go and publish this and I go back to our parent pip line here what we'll see happen is within our spreadsheet it will go through it's going to give us the title it's going to give us the transcript and then it's also going to give us that markdown blog post here is a summary of the key points within the video and here is our blog post here's the setup the capabilities so on and so forth and within here you can see that it's going through each of these different videos that I had within here and it's creating the blog post let's say I want a script for an Instagram reel a YouTube short as well as a Tik Tok what you can do within here and what's great with Vector shift is it's really easy to experiment let's say I want to use GPT 40 for generating a script for Instagram I'll say you are an expert script writer for Instagram res create a 30 second compelling piece of content based on the title and transcript and then within here a similar thing is I can grab the title node and leverage the text and then I can also grab the transcript here and I can reference that and then finally within here I can reference open Ai and the response and you can play around with these different models to see what works best let's just say I want to leverage Google as well as a model from anthropic to generate these scripts for different YouTube shorts instead of Instagram reals I'll say YouTube shorts and a similar thing here for Tik Tok I can reference Google and the response for Tik Tok similar thing is in the response right here once you have that just make sure that you deploy each time for all of the different changes here I just zoom out to what we created here we have the inputs we have the transcript and then for each of these different pieces of content that we're having generated within the spreadsheet we have all of these different nodes for all of these different providers here this is a relatively simple workflow but you could use this in a number of different use cases you could use this for researching or Googling for different pieces of information and filling out those different cells now if I go over to our spreadsheet what we'll see is for each of these items we're going to get the title that passes through we're going to get that transcript from that node from Vector shift we're going to have the blog we're going to have our Instagram real our YouTube short as well as our Tik Tok if we just take a look here is our blog post here is the Instagram reel that we have we have here is the transcript for the YouTube short and then finally we have the script for the Tik Tok each of these platforms have different algorithms and they perform differently and as you can see here for the outputs even though I asked for exactly the same thing effectively for each of these we have very different responses between the different models we use from Gemini Google as well as chat gbt so hopefully you found this video useful and this is just one example for how I could potentially use something like this for automating these different tasks CU if I was to do this and write out a blog post for instance for each video it would have taken a lot of time as you could imagine is having to create all of this different content would have taken an awful long time to create and for me personally by being able to have one original piece of content and being able to drive different pieces of content from that main piece of content it can potentially be something that could save a ton of time especially in my context where I have hundreds of pieces of content hopefully this gives you an idea on how you can leverage Vector shift as well as how you can automate different repetitive tasks within spreadsheets if I was to write out all of this different content I probably would have created maybe onethird or one quar of the pieces of content while I do creating content I am just limited to personal bandwidth and I've only really been able to focus on YouTube but by leveraging something like vector shift what I can all of a sudden do is be able to create all of these different pieces of content that I could potentially consider to use whether it's within blog post or creating these scripts for Tik Tok Instagram YouTube shorts or whatever it might be otherwise that's pretty much it for this video I encourage you to try this out I know this is something that's definitely going to continue to gain interest over the coming years in terms of people being interested in leveraging AI within their spreadsheets in the example that I showed you this would have taken me hours days potentially even weeks across hundreds of videos to create this amount of content and mind you the one thing that I do want to note is you can play around with these so I just gave a bunch of different examples but what you can do is you can begin to refine these over time as you begin to find and isolate things that work well and play around with these different models and see what works best for the particular task but otherwise that's pretty much it for this video check out Vector shift try this out on a spreadsheet that you already have if you found this video useful please comment share and subscribe otherwise until the next one