Dude, I have been trying to create this for MONTHS. Almost since chatGPT was released. The missing ingredient was Zapier, and your explanation. Thank you so much for creating this video. I’ve been trying to do exactly this, but purely within ChatGPT with a carefully crafted prompt. My prompt engineering abilities have been stretched to the max 😂
Navigate to key moments👇 made via tubestamp.com 01:58 - Uploading test invoice and setting up extraction form. 03:04 - Creating automation for data extraction trigger. 04:32 - Converting invoice to editable document via Google Drive. 06:32 - Using Zapier's tool for text extraction from document. 07:49 - Converting underlying file for data extraction. 08:09 - Testing extraction of all text from invoice. 09:14 - Context for extracted information and Excel formatting. 17:15 - Live demonstration of data extraction and Excel formatting. 17:22 - Displaying underlying data record for invoice. 18:16 - Emphasizing automation process reliability for various invoices. Recap by TubeStamp ✏️
Hi there Corbin< Super helpfull your video's Thanx for that. As I now have found a work around stitching some of your video's but I can't find the solution for this. I NOg followed al your stops from Drive- javacoding- spreedsheet, extracting and uploading it again to Drive. But we scan all our docs on a scanner. The pdf files are random numbers. SO I wanted to change the title with OpenAI which I did, but the orriginal PDF gets lost thru the proces. ANy smart solution forthat? SO now It go's from PDF, spreedsheet, chatgbt and back tot Drive. Please help me out\
Hey Corbin, I am having trouble getting results back after submitting the invoice. What could be the problems? Also, when we submit the new invoice, are they supposed to be in the Google Drive because I don't see any new ones.
Great video 💯!! Thank you very much Corbin for this beautiful demonstration. Quick question for you: what would happen if we have several invoices with slightly different formats. Would we still be able to collect data with the same parameters initially established? Or should we adjust the settings?
Honestly i feel there are problem with this kind of automation, i already experience Chatgpt take wrong data or even the data before (which i dont know why its save in its memory), so it not reliable yet and then if chatgpt (assume gpt5) run perfectly then it supposes to do everything what you just do right now. i am not against this automation but i am not fully convince why i need to use it
If the data extraction part doesn't seem that effective to you yet, you can also use the logic here for other contexts, such as sentiment analysis, or pair it with other tutorials to get the user's email from an invoice and see if we already have data on the client, as shown here: th-cam.com/video/-lCA9di-nkM/w-d-xo.html&ab_channel=Corbinai
Yes, but the logic we set up here can go steps further past just data extraction. With this data, we can now identify further metrics in our Excel sheets that we may find pertinent, through AI.
Dude, I have been trying to create this for MONTHS. Almost since chatGPT was released. The missing ingredient was Zapier, and your explanation. Thank you so much for creating this video.
I’ve been trying to do exactly this, but purely within ChatGPT with a carefully crafted prompt. My prompt engineering abilities have been stretched to the max 😂
Glad it was helpful!
Bro Keep Going Bro !
You are the king of Zapier. We love you here from Dubai.
🤝
Navigate to key moments👇
made via tubestamp.com
01:58 - Uploading test invoice and setting up extraction form.
03:04 - Creating automation for data extraction trigger.
04:32 - Converting invoice to editable document via Google Drive.
06:32 - Using Zapier's tool for text extraction from document.
07:49 - Converting underlying file for data extraction.
08:09 - Testing extraction of all text from invoice.
09:14 - Context for extracted information and Excel formatting.
17:15 - Live demonstration of data extraction and Excel formatting.
17:22 - Displaying underlying data record for invoice.
18:16 - Emphasizing automation process reliability for various invoices.
Recap by TubeStamp ✏️
Awesome video mate, cheers! Quick question. How would you adapt it to handle multiple invoices at the same time?
Hi there Corbin<
Super helpfull your video's Thanx for that. As I now have found a work around stitching some of your video's but I can't find the solution for this.
I NOg followed al your stops from Drive- javacoding- spreedsheet, extracting and uploading it again to Drive. But we scan all our docs on a scanner. The pdf files are random numbers. SO I wanted to change the title with OpenAI which I did, but the orriginal PDF gets lost thru the proces. ANy smart solution forthat? SO now It go's from PDF, spreedsheet, chatgbt and back tot Drive. Please help me out\
Hey Corbin, I am having trouble getting results back after submitting the invoice. What could be the problems?
Also, when we submit the new invoice, are they supposed to be in the Google Drive because I don't see any new ones.
Great video 💯!! Thank you very much Corbin for this beautiful demonstration. Quick question for you: what would happen if we have several invoices with slightly different formats. Would we still be able to collect data with the same parameters initially established? Or should we adjust the settings?
In this context, it would be best to make your ChatGPT prompt very specific in what it's looking for so it can handle data coming from other contexts.
Love your videos, can you do the same with voiceflow ?
Thanks! What specifically are you looking for?
Could you give me that invoice I don't have one to do as an example haha
Honestly i feel there are problem with this kind of automation, i already experience Chatgpt take wrong data or even the data before (which i dont know why its save in its memory), so it not reliable yet and then if chatgpt (assume gpt5) run perfectly then it supposes to do everything what you just do right now. i am not against this automation but i am not fully convince why i need to use it
If the data extraction part doesn't seem that effective to you yet, you can also use the logic here for other contexts, such as sentiment analysis, or pair it with other tutorials to get the user's email from an invoice and see if we already have data on the client, as shown here:
th-cam.com/video/-lCA9di-nkM/w-d-xo.html&ab_channel=Corbinai
I would just use Excels import function
Yes, but the logic we set up here can go steps further past just data extraction. With this data, we can now identify further metrics in our Excel sheets that we may find pertinent, through AI.