Collaborate on Invoice Bootstrap for it with Ease Using airSlate SignNow

Invoice bootstrap for IT

hey there it's prath so today I'm going to show you how to process invoices faster cheaper and quicker than any human could using AI automation invoice processing is a common pain point for many businesses and it consumes valuable time resources while also being prone to human mistakes AI automation can tackle this problem for you much better by more efficient ly extracting details from invoices and improving speed reducing costs and minimizing human mistakes using AI automation to process your invoices allows your business to reallocate your valuable human resources to task that better benefit from the human expertise and this can help reduce your business cost significantly and help you do far more with less so I'm going to show you how to take uploaded invoices as PDF files from a Google sheet and then extract key details from them using AI automation so let's have a look at this so here's an example invoice um with some details we've got the company name uh the address all the individual line items and these would uh need to be calculated and also extracted for us to have all these invoice details so here is a Google spreadsheet where I've actually uploaded the um invoice URL so so by the way this is actually been uploaded to Google Drive and so I've got these like four sample invoices and I sort of take a link from here and I've added the link to each of the invoices into this Google sheet and here is the invoice details that my AI autom has extracted for us which has got the total amount of all the invoices the company name uh the date and all the individual line items from the invoice as well as the address so we could easily adjust this Automation in various ways if we needed to so for example we could rather than extracting the PDFs from a Google sheet or Google Drive we could extract them directly from an email and we could also update these invoice details rather than into a Google sheet we could put them into a CRM or another kind of database and we could also add other modifications such as sending an email notification to let somebody know that a new invoice have been processed so there's lots of different ways we can take this core AI Automation and adjust it for your particular needs um and this could be incredible L useful for so many businesses that spend a lot of time basically processing invoices manually so let's get into how we would make this so the first thing we need uh is to have a make account so make is the platform I use for making AI automations and I would say it's one of the best ones and you can actually make a free account uh quite easily here so you'll need to make an account here another thing you'll need to make is an account with uh this company pdf. and this uh enables you to extract details from a PDF like to basically convert a PDF into text so that we can get the details out of it and again uh there's a kind of free uh limited trial offer that you can use to get uh sort of use it if you wanted to try it out and then the final thing that you'll need to set up is an open AI API account and this basically allows you to use uh tools like chat GPT inside an AI Automation and uh again the costs are quite low like in terms of pennies and cents to use this to extract uh details from a PDF so yeah you'll need to also set up an openai API account to use this so let's get into how we make the AI automation so this is the AI autom information I've made to extract from the Google sheet the invoices and take out the invoice details and then put them back into the same Google sheet which of course we could adjust in different ways if we needed to so let's go through each of the modules and how You' need to set each of those up so the first module is the Google Sheets module so the first part you need to connect your Google account and then we need to set up these uh different settings so we would choose select from Drive uh and I've selected uh my uh Google sheet that I'm using for this example so that's basically this sheet and that's called invoice processing example and then you select the sheet name um I set table headers as yes cuz it does have table headers and then so I'm searching these rows so what am I going to search the rose for well really what I want to do is I want to find the next invo that hasn't been processed so the way I'm going to do it is by scanning this totals field and I'm going to find the next column uh or the next cell that hasn't a total so basically this will be the next invoice uh link to process so the way I'm doing that here is I'm filtering and searching for the Google sheet rows and I'm saying uh find me uh a row that has a total where the value doesn't exist which basically empty so this is the search Google Sheets row module First Step so the next step is the text Baza module and in here what we're basically doing is we're extracting some information from this Google sheet link so uh and what we want to do is actually get there's an ID there's a number in that Google sheet uh link and that's the actual uh bit of information that we need to be able to actually download the PDF file which is part of what we want to do so this thing here is actually called a regex uh which stands for regular expression and that basically allows you to extract certain bits of information from a b up text and if you don't know about this don't worry you can use a pre- tool like chat EBT to find out what red X's are and actually you can even get chaty PT to make a reg X for you as well so that's really handy I actually use that uh to do this reject as well so this Rex will take that link uh that we saw in this file here in this uh spreadsheet and it will extract from this link uh this there's a little code in this link and that's what it's going to get basically which is the ID basically of the file on the Google Drive and so we can see then here as well uh this is the input to the text PA and it's basically taking this invoice URL of the road that we've got selected so the next module is the Google Drive download a file module and what that does again you need to set it so it's connected to your Google account and we set uh how we download a file is we're going to enter in manually a file ID so we set enter manually and then the file ID is what we've extracted from the text PA using that Reg in the previous step and that will uh enable us to download the actual PDF invoice document in this module so the next step is where we connect to the PDF Co account so again you would have needed to set up this PDF Co account and then once you've created that before this then you can connect to that free account here and you get certain number of credits that's quite enough for you to do a tests basically so what do we want to do well basically in the previous step we've just got uh and downloaded the uh invoice PDF from the previous step so the next thing we're going to do is basically to upload that PDF document to pdf. C which will extract the text of uh the invoice basically from that document and so the way we set that up in this module is we set input file to upload a file and then for file we set it to actually the Google Drive download a file module so that will basically pass along the document we've just extracted from the Google drive module previously so for conversion uh so pdf. can do lots of different conversions and um the one we want is just to convert the PDF uh into text basically um and yeah that's it oh sorry there is a final step where uh we want to set the output structure of what we're outputting from this step so we're taking in a PDF document sending it to pdf. co and then we're getting back text and we want to structure that in a certain way so uh this Json output so Json is just a certain kind of way of structuring the uh information or text data in this case and so we're asking it to structure it as adjacent output I'll show you what that looks like it's just a structured bit of data basic so that's the pdf. co module the next module is the AI part basically so that's the chat GPT part so let's have a look at that this is probably the most meaty and interesting AI part of this whole AI automation so again you would have needed to create an open AI API account uh as I showed you in the previous step and here you create um a connection to that account once you've got that there then select method so we wanton create a chat uh completion what that basically means is we're going to ask chat GPD a question and we want it to answer it that's all that means basically uh and really what we're going to do here is we're going to pass it along the text that we've just extracted from the PDF of the invoice and we're basically going to get chat GPT to figure out what are the different bits of information that we want and get those out for us so that's the clever part that we're going to use AI for in this AI automation so in terms of models we're going to use uh this one uh at time of recording uh GPT 40 is a really really good one to use for this task um and then to get this chat GPT to do uh what we want to extract these bits of information from the invoice we need to set a few different things so if you've taken some of my courses on my uh website the future AI um I have a course on how to use chat GPT and if you've taken that course uh you would have understood something called Ro prompting so if you haven't taken those course no don't worry uh so Ro prompting basically means asking chat GPD to pretend like it's a certain skill set or has a certain role so uh you could ask it to uh pretend it has this role and actually when you do that it actually does the task that you want much better than if you just asked it so the first bit so there basically a few different bits of instructions we're passing along to chapter p and the first one is this role so and we set the role here uh as um so this is a certain kind of message to chat GPT where saying here this is how I want you to behave when I interact with you and the instruction for the is you are a helpful intelligent data extraction assistant and so basically we're saying you're really really good at data extraction uh and that will actually help get better results so the next part of our instructions to chat gbt is the user instructions and that's basically where we're telling it what the task that we want it to do um and for that we set the role as user and here is basically the chat GPT prompt that we're going to use to actually do the job of getting the different bits of information from this text from the invoice that we've just extracted so here's the prompt so you extract data from PDF invoices you'll receive a text scrape of an invoice data and your job is to turn that invoice data into five structured Fields a total a company a date line items and from address address and we've also added some extra clarification for what prom address means and we what we're saying here is not the email but the actual address of the sender just so it doesn't get confused uh and also then we're next saying return the results in adjacent using this format so this is what this Json format looks like it has these um curly brackets and then it basically structures it so that it's got all the different bits of data labeled so for example we can see here we want it to Output the total amount of the invoice uh we going this label total and this is where the total value would go and here's the company which for label for the company and that's where the company value would go and so we're doing the same for the date the line items and the address and they would all be output in this Json format which has this curly bracket and structured labeled format when that's really handy as well because it makes it easier to extract this data later and our final instruction is to respond only in Jason uh and the first character of your response should be a curly bracket do not output Jason or other kind of formatting well so we're being very very strict and and telling it um exactly how we want it to Output the data and that's really important when using AI tools like chat GPT this prompt needs to be a very very clear precise nice set of instructions so that chat gbt knows what we want it to do and do the job for us well so the final bit of the chat GPT prompt is basically where we pass along the information that we wanted to do the work on and here is actually where we pass along so it says body here that's actually what the PDF Co um output uh from as a result of uh looking at that PDF invoice so you can start to see how the bits of information passed along have this color code so for example we can see this is orange and that tells us this kind of comes from that previous module so that's basically the AI part that will take the textual information output by PDF Co and it will actually extract all the different bits of our invoice that we really want to extract so in the next module this Json module all this does is it basically turned it into a proper sort of Json uh data structure basically so it's taking the result as we can see the color coding again from this green open AI module and it's going to convert that into a proper Json structured uh data object then this tools part what that's going to do is it's basically going to set uh a value and what that's going to do is it's going to take the total that it's calculating uh from the previous step and actually convert that into uh a proper name and number uh cuz it doesn't quite do that otherwise without this step now the final step is basically where we update the Google Sheets that we got the link from so in this module again we connect to our Google account and then we again uh find our Google sheet that we've just extracted that invoice link from so we're doing that here my drive search byp pass spreadsheet ID invoice processing example that's that uh invoice select sheet one so to update this Google spreadsheet we first need to choose the right row number but we've already got that because when we found the the right row which had the next invoice to process in this previous step it actually also saved the row number so this is what this row number is again we can see the green color coding and do you see that flashing on the left it's basically telling you oh I got this row number from this module which is basically the search step of the Google sheet so that's really cool so and then um it's just saying well does it contain headers and it does and this is the final bit where we're basically writing back the extracted invoice information to this Google sheet so here is where it will write the the total to the total column the company to the company column the date to the date column the line items the line items column and the from address to the from address and that's how we end up with these different parts of the invoice all extracted and written back to the Google sheet so let's run a test to process another one live so what we're going to do is we're going to process this second invoice here so let's just have a look at what this looks like so so this second invoice is from a company it's just these are all just test demo invoices so this one's from a company called demo sliced invoices and this invoice is for web design for a total of $93.50 and it was done on the uh in January basically so we're going to extract all these details from the invoice using our AI automation so what we should end up with is in the spreadsheet once we run this AI automation we should find hopefully all those extracted details put into this row okay so ready let's give it a go so let's uh get ready to run this AI automation by pressing the Run once button here and so so here we go so it's searching for the Google sheet it's processing the link it's downloading the document it's uploading it to pdf. co it's return the text it's run chat GPT to extract the bits of the invoice and it's actually done it um it's written it back to the Google Sheets so let's go and have a look see if it's in there so great here it is so here we have our extracted invoice total the company name that is taken out the date uh the line items that's all right and it's got the address as well so perfect it's done a great job so you could run this AI automation all the time and it could be processing hundreds or even thousands of invoices for you uh saving you huge amounts of valuable time that you can direct towards you know far more interesting things that require your valuable human input while your AI autom saves you considerable time and money so I hope you found this really useful and I hope it's helped you better understand what AI automations can do to help improve your business and to help reduce your costs and help you do more in far less time and if you'd like to better understand how AI automation could help process your invoices or help with another area of your business book a free AI automation consultation with with me in the link in the description and I'll see you in the next one