AI Tool That Extracts Data From Emails/docs and Takes Action (Must Have Built Something Similar)

I'm building a tool that takes incoming emails (and their attachments — PDFs, Excel files, scanned documents, images) and runs them through a pipeline that looks like this: Email or document arrives AI reads and extracts the relevant structured data from it (this could be line items, prices, dates, customer info, specs — whatever is relevant to the use case) That data gets organized into a clean, structured format (a table, a database, a spreadsheet) so it can be reviewed or analyzed Based on that data, the system takes some kind of follow-up action — for example, if the email was a request for quote, it drafts a quote automatically using pricing rules; it could also be a confirmation email, a flagged exception for human review, or a sync to another tool This is not tied to one specific industry. I'm exploring which market this is most valuable in, so I want to hear about your experience across any industry — freight, logistics, insurance, real estate, procurement, healthcare, legal, e-commerce, accounting, whatever you've worked on. What I'm looking for: I want to talk to developers who have actually built something like this before — not people who are confident they could build it, but people who have already built a real version of this exact pattern (email/document in → AI extraction → structured data → automated action), ideally as an internal tool for a business, a client project, an agency build, or your own product. When you apply, please explain in detail: What the tool was built for (what industry, what was the business trying to solve) What triggered the automation (was it an email inbox, a file upload, both) How the extraction was done (what tools/models you used — OCR, LLM-based extraction, a specific API, custom parsing, etc.) What the output looked like and where the structured data ended up (a database, a spreadsheet, a dashboard, a CRM, etc.) What the "action" step was and how it worked (did it draft something, send something, push data somewhere, flag something for review) The tech stack you used end to end Any real results if you have them (time saved, accuracy, adoption, client feedback) Proof matters a lot here. Please include whatever you can share — screenshots, a demo video, a live link, a GitHub repo, case study writeups, or anything else that shows this is real work you've actually done, not just a description of what you could do. Applications without any proof of prior similar work will not be prioritized. A few other things to mention in your application: Your experience working with LLMs/AI APIs for document and email data extraction specifically (not just general AI experience) Whether you've built systems that needed to handle messy, inconsistent input formats (different senders, different layouts, different file types) rather than one fixed template Your general availability and rate