3 ms·

Thanks for the response here -- 0 constraints. Lots of PDFs in and a final PDF out. Questions vary per RFP but we have 90% of cases covered in a google doc answ

Thanks for the response here -- 0 constraints. Lots of PDFs in and a final PDF out. Questions vary per RFP but we have 90% of cases covered in a google doc answer bank. I'd be curious how this could be within a template -- so maybe Copilot would be the right tool here. Raw output of text to then format would still be 20% of the work without that.

sebg 2y ago
Got it—no constraints, mostly PDFs in → PDFs out, with 90% of answers covered in a Google Doc answer bank. Given that, here’s a streamlined approach: 1. Extract RFP Content Azure Form Recognizer, AWS Textract, or Adobe Acrobat API can extract text and structure from PDFs. If RFPs follow a standard layout, LlamaIndex or LangChain can help chunk and categorize sections. 2. Retrieve and Insert Answers Use Google Docs API to pull relevant responses based on keyword or semantic search. OpenAI Assistants API or Claude 3 can rephrase and tailor the responses dynamically. 3. Automate Formatting with a Template Microsoft Copilot (Word/Excel + SharePoint): Copilot can fetch Google Doc answers and insert them into a predefined template. If you standardize the output format in Word, you can generate a PDF-ready final draft. Alternative: Python + Pandoc + Jinja2 templates if you want a more controlled workflow. 4. Output Final PDF Convert the formatted document to PDF using Word’s built-in save-to-PDF, Pandoc, or LibreOffice headless mode. ** Low-Code Implementation If you're looking for minimal coding: Microsoft Power Automate + Copilot → Automatically extract, populate, and format RFP responses. Zapier + OpenAI + Google Docs → Retrieve answers, format them, and push to a doc/PDF. Would Copilot’s Word integration work within your workflow, or do you need more automation for final formatting? If you want a completely hands-off approach, a small workflow automation script could eliminate the remaining 20% of manual work.

sebg 2y ago

Got it—no constraints, mostly PDFs in → PDFs out, with 90% of answers covered in a Google Doc answer bank. Given that, here’s a streamlined approach: 1. Extract RFP Content Azure Form Recognizer, AWS Textract, or Adobe Acrobat API can extract text and structure from PDFs. If RFPs follow a standard layout, LlamaIndex or LangChain can help chunk and categorize sections. 2. Retrieve and Insert Answers Use Google Docs API to pull relevant responses based on keyword or semantic search. OpenAI Assistants API or Claude 3 can rephrase and tailor the responses dynamically. 3. Automate Formatting with a Template Microsoft Copilot (Word/Excel + SharePoint): Copilot can fetch Google Doc answers and insert them into a predefined template. If you standardize the output format in Word, you can generate a PDF-ready final draft. Alternative: Python + Pandoc + Jinja2 templates if you want a more controlled workflow. 4. Output Final PDF Convert the formatted document to PDF using Word’s built-in save-to-PDF, Pandoc, or LibreOffice headless mode. ** Low-Code Implementation If you're looking for minimal coding: Microsoft Power Automate + Copilot → Automatically extract, populate, and format RFP responses. Zapier + OpenAI + Google Docs → Retrieve answers, format them, and push to a doc/PDF. Would Copilot’s Word integration work within your workflow, or do you need more automation for final formatting? If you want a completely hands-off approach, a small workflow automation script could eliminate the remaining 20% of manual work.