Has anyone here implemented AI or automation for an intermediary fiscal / fiscal agent / payment intermediary type business? By that I mean companies that sit between: • Government or institutions • And end workers or service providers and handle disbursements, approvals, tracking, compliance, and reporting. I’m curious about real implementations, for example: • Automating payment allocation or approval workflows • Reducing manual back office work (secretaries, data entry, reconciliation) • Compliance tracking, audit trails, reporting • AI agents for validation, exception handling, or follow ups If you’ve done something like this (even partially), I’d love to hear: • What problem you automated first • What stack you used • What actually moved the needle for the owner (time saved or cost reduced) Not looking for theory. Looking for real world use cases.Appreciate any insights 🙏

Documentation is one of the most important parts of building automation systems. It is also one of the most tedious. In this video, I show a system that turns a walkthrough video plus a workflow screenshot into a clean, structured Google Doc. No more writing SOPs from scratch. What the system produces: - A well formatted Google Doc - Clear sections for setup, nodes, APIs, and deployment notes - Screenshot placeholders so you know exactly where visuals should go - A production checklist you can use before shipping How it works, step by step: 1. Submit a form - Video URL (YouTube or an unlisted upload) - Workflow screenshot (file upload) 2. Extract the screenshot for later use - Convert the uploaded image into a base64 string - Store it so it can be referenced further down the workflow 3. Pull the video transcript with Apify - Run the YouTube Transcript Scraper - Return the transcript data 4. Clean the transcript with a Code node - Combine all transcript segments into a single readable paragraph 5. Convert the stored screenshot back into a file - Move base64 back into a file so it can be analyzed 6. Analyze the workflow screenshot with OpenAI - Generate a structured description of what the workflow is doing 7. Generate the documentation with Claude via OpenRouter - Feed it two inputs: - Transcript text - Workflow screenshot description - Output a JSON with: - Documentation title - Full markdown documentation 8. Parse the JSON output - Split out title and documentation fields cleanly 9. Create a Google Doc from Markdown - Use a community node that supports Markdown to Google Docs formatting - Auto-create the doc in the correct Drive folder Result: You record a 10 minute walkthrough, upload one screenshot, and the system drafts the entire doc for you. If you want to build it, the blueprint and prompts are linked in the description. Subscribe for more tutorials on n8n systems, AI agents, and agency-ready automations. https://youtu.be/a7O2-axMJEg?si=dRkeaTigxgVLyzLI