Draft keigo-checked Japanese contact form replies with OpenAI and Teams

Quick overview This workflow collects inquiries via an n8n Form, matches them to an FAQ stored in an n8n data table, drafts a Japanese keigo email reply with OpenAI, runs a second OpenAI pass to correct keigo, and posts the structured draft and review notes to Microsoft Teams.

How it works Receives a new inquiry submission from an n8n Form with the sender’s name, company, email, and message. Loads all rows from the n8n data table keigo_faq, aggregates them into one FAQ list, and adds your company name and email signature. Uses OpenAI to classify the inquiry into a category (pricing, demo request, product question, trouble report, other, or sales pitch/spam) and selects category-specific reply guidance. Matches up to three relevant FAQ entries by scoring keyword hits in the message and compiles them into a short FAQ text block. Uses OpenAI to generate a structured Japanese business email draft (subject, body, urgency, and staff check points) using only the matched FAQ answers as facts. Uses OpenAI again to correct keigo issues in the draft and list the specific fixes made. Sends an Adaptive Card to Microsoft Teams via an incoming webhook containing the original inquiry, the keigo-checked draft, the fix list, the FAQ topics used, and the check points for staff review.

Setup Create an n8n data table named keigo_faq with topic, keywords, and answer columns, and populate it (optionally by running the “Load Sample FAQ” branch once). Add an OpenAI credential and select your preferred model (configured as gpt-5-mini) for the classification, drafting, and keigo-check steps. Set your company name and signature text in the workflow’s company details step. Create a Microsoft Teams Workflows incoming webhook and paste its URL into the Teams HTTP request step. Publish the n8n Form and share its production URL where you want to receive contact form submissions.