AI Lead Capture Tool

The problem

At a trade-show booth, the bottleneck isn't getting visitors to hand over a visiting card — it's what happens after. Cards pile up, get typed into a spreadsheet days later by someone squinting at handwriting, and by the time a lead is actually followed up, the visitor barely remembers the conversation. Add cards printed in Gujarati or Hindi and the manual data-entry problem gets worse, not better.

What was built

A camera-first web app for staff to use at the booth: point a phone or tablet at a visiting card, and an AI vision model reads the name, designation, company, phone, email, city, and product interest straight off it — including Gujarati and Hindi text, transliterated to English automatically. Staff glance over the extracted fields, correct anything the model got wrong, and save. The card photo is stored alongside the lead, and everything syncs into a live Google Sheet and directly into the Production & Quoting Platform's (SYS-04) Expo Leads pipeline — no re-typing into a second system afterward.

The whole thing runs in a deliberately built mock mode first — realistic sample data instead of real AI calls, no external accounts required — so the full scan-to-save flow can be tried and trained on before any real credentials are wired in. Every real integration (AI extraction, Sheets sync, photo storage) degrades gracefully and can be switched on one at a time. The whole app sits behind a single login, since visiting-card data is real names, phone numbers, and emails.

Status: in development

Built ahead of Lukeron's stall at the Kutch Industrial Expo (11–13 Dec 2026), where it will run for real. It's a genuine, working tool today — the field test is still ahead of it, which is why it's marked in development rather than live.

Who this is for

Any business that collects leads at events, showrooms, or field visits and currently re-enters them by hand later has this same gap — and the gap compounds every hour a hot lead sits untouched in a card holder. A tool like this turns capture into a five-second action instead of an end-of-day chore, in whatever language the lead actually walks in speaking.