Automate Your RFP Process with OpenAI Assistants

This n8n workflow demonstrates how to automate oftern time-consuming form filling tasks in the early stages of the tendering process; the Request for Proposal document or "RFP".

It does this by utilising a company's knowledgebase to generating question-and-answer pairs using Large Language Models.

How it works

• A buyer's RFP is submitted to the workflow as a digital document that can be parsed.
• Our first AI agent scans and extracts all questions from the document into list form.
• The supplier sets up an OpenAI assistant prior loaded with company brand, marketing and technical documents.
• The workflow loops through each of the buyer's questions and poses these to the OpenAI assistant.
• The assistant's answers are captured until all questions are satisified and are then exported into a new document for review.
• A sales team member is then able to use this document to respond quickly to the RFP before their competitors.

Example Webhook Request

curl --location 'https://<n8n_webhook_url>' \
--form 'id="RFP001"' \
--form 'title="BlueChip Travel and StarBus Web Services"' \
--form 'reply_to="jim@example.com"' \
--form 'data=@"k9pnbALxX/RFP Questionnaire.pdf"'

Requirements

• An OpenAI account to use AI services.

Customising the workflow

OpenAI assistants is only one approach to hosting a company knowledgebase for AI to use. Exploring different solutions such as building your own RAG-powered database can sometimes yield better results in terms of control of how the data is managed and cost.