Built an automated business card capture system that extracts contact details from photos via AI and creates structured records in a Notion CRM - turning a stack of cards from a networking event into searchable, actionable contacts in seconds.

The Problem

After networking events I'd end up with a stack of business cards and no efficient way to process them. Manually typing details into a contacts database was tedious, easy to put off, and meant useful connections went cold while cards sat in a pile. I needed a way to go from card-in-hand to contact-in-CRM with as little friction as possible.

What I Built

A photo-to-CRM pipeline triggered from Slack, powered by AI vision:

• Slack channel trigger — I take a photo of a business card and drop it into a dedicated Slack channel. That's the only manual step.
• Gemini AI extraction — n8n picks up the image and sends it to the Gemini node for vision-based data extraction. It pulls out name, email, phone, company, job title, and any other details visible on the card.
• Notion CRM write — the extracted data is written to a structured record in my Notion CRM database, ready to search, filter, and follow up on.
• Zero data entry — from photo to contact record with no typing, no forms, no manual processing.

<aside>

Tech Stack & Features

Platform: n8n

AI: Gemini (vision-based extraction)

API: Slack API, Notion REST API

Key modules: Slack trigger, Gemini AI node (vision), Code node (JSON parsing), Notion database module

Patterns: Event-driven intake, AI vision extraction, structured data write, zero-friction capture

</aside>

<aside>

Documentation

Technical Documentation

SOP - Business Card Capture

</aside>

n8n Workflow