The $0.02 Workflow That Automates Data Extraction With n8n & AI

[@MarcusNeufeldt]

Can't get enough from n8n workflows, thanks for sharing, keep it up

[@hunterasneed]

Thank you so much, I really appreciate it and the support!

[@agentred8732]

From or Of?

[@hunterasneed]

?

[@inayansouto]

Hello Hunter! Congratulations on creating such an insightful and practical video. Your detailed explanation of how to automate document data extraction using n8n and OpenAI is incredibly valuable, especially for small and medium-sized businesses looking to optimize workflows. The breakdown of tools, such as structured outputs and seamless integration with Google Sheets, is clear and actionable. Kudos for making automation accessible and cost-effective for your audience. Keep up the amazing work, and I look forward to more innovative content from you!

[@hunterasneed]

Wow, thank you so much! This comment means the world to me and honestly made my day. I really appreciate it and am glad you're finding it useful :)

[@danielasalonen2572]

Great content!

[@mydogisbear1669]

Can you not just trim in the text extraction tool for structured pdfs and skip the extracted text step? Or is there something else that step does that I’m missing?

[@tubaguy0]

Very much appreciate the content. Please take some constructive feedback in the spirit of improving your future videos: please, please take some time to reduce (preferably remove) the filler phrases “go ahead and,” as well as “went ahead and,” from your vocabulary. I was distracted and frustrated by the repetition of those phrases in almost every sentence of the video. Wishing you well and looking forward to more videos!

[@hunterasneed]

Hey, I really appreciate the feedback. This is completely valid and something I definitely want to work on. I think when you’re used to the way you speak, it’s easy to not pick up on patterns like that so having others point it out (especially when it’s negatively impacting the presentation) is really helpful! I’m going to do my best to keep an eye on this and if you see more if it going forward please feel free to comment again as I’m sure you’re not the only one that noticed and was impacted. Ultimately I want to make the best automation content possible so feedback like this means the world. Hope you like future content and to see you around here! P.S. thank you for going about it in way you did. To your point, it was constructive (and polite) which is not always the case on here.

[@tudor-matei]

Thanks for sharing this Hunter. Have you tried using the Structured Output from the AI Agent node?

[@hunterasneed]

I did actually and I was having mixed results with it and also was having issues connecting the nodes given that I didn’t necessarily want this workflow to be an “agentic” one. That said I’ve had great results with the structured outputs functionality especially when used in a RAG capacity.

[@jessicaj351]

Hi Hunter great video! Do you have recommendations for how to extract from files with unstructured data (like paragraph form)? thanks in advanced. Also, just subscribed BTW

[@hunterasneed]

Yes absolutely! So this some approach should still work, you just need to specify what specifically you’d like to extract out of the paragraph in your schema. Ex. That could be something like list of speakers, main character, total, overview of paragraph, etc. basically as long as you in the schema tell OpenAi what to look for, it should be pretty successful with extracting it! Also thank you so much for the support! :)