Qualify inbox leads and open CRM deals

Set up an n8n workflow that scores inbound email leads, stores their history, and creates CRM records when they meet your rules.

Done with n8n· Advanced· Jun 2026

What they did
They built a self-hosted n8n flow for a procurement client’s shared Gmail inbox. OpenAI tagged each incoming message, pulled out details such as budget, timing, project size, decision power, and urgency, and saved the lead state in Supabase so later emails could update the same thread. Fixed rules handled routing, follow-ups, and escalation, and once there was enough detail, the workflow created customer and deal entries through the CRM API.
How it went
In the first month, a business receiving about 200 inquiries a month saw roughly 3% turn into qualified leads. The author also said about 3% of those qualified inquiries became closed business. Human review was still used for quote drafts on hot leads and for exceptions or escalations.
Tips
The hardest part was making the stored-memory setup and lead-scoring dependable, so plan time for testing those pieces. Treat the orchestration layer as one part of a larger set of sub-workflows, not the whole system.

Try it yourself with n8n

Set up an n8n workflow for my shared [email inbox] that watches for new messages, sorts each one into [lead/support/other], and extracts [budget, timeline, project size, authority, readiness]. Keep each conversation tied to the same record in [database]. If the lead meets my rules, create the [customer/deal] in [CRM] and prepare a reply draft. Before any irreversible step such as sending an email, creating a deal, deleting a record, or submitting information, stop and ask me to approve it. If details are missing, ask the sender a concise follow-up question and save the response back to the same record.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.