Zoha Liaqat's Work | ContraWork by Zoha LiaqatZoha Liaqat
AI Automation Developer
Most online stores lose a large share of their shoppers at checkout. They add items to their cart, get distracted, and never come back. This workflow wins some of those sales back automatically, with no manual follow-up.
What it does
When a customer abandons a cart, the workflow logs it and waits. If the customer hasn’t bought after 1 hour, it sends a friendly reminder email. If they still haven’t bought 24 hours later, it sends a stronger “last chance” email with a discount code. Customers who buy at any point never get another email.
How it works
A webhook receives the abandoned-cart event (customer name, email, cart items, Cart ID).
The cart is logged to Google Sheets with the status pending.
The workflow waits 1 hour, then re-reads that cart’s row so it sees the latest status.
If the status isn’t purchased, a personalised reminder goes out through Gmail and the status becomes reminder 1 sent.
After a further 24 hours, the status is checked again. If the customer still hasn’t bought, a final email with a discount code goes out and the status becomes reminder 2 sent.
A separate purchase workflow updates the sheet to purchased as soon as an order completes, which stops the sequence.
Why it’s reliable
Every cart is tracked by its own Cart ID, so simultaneous customers never overwrite each other.
The status is re-checked after every wait, so nobody gets a reminder after they’ve already bought.
The Google Sheet works as a live dashboard of where every cart stands.
Tools used: n8n, Webhooks, Google Sheets, Gmail, Wait and IF logic. This project tackles a real small-business pain point: inventory silently running out because no one's tracking stock in real time.
Order processing (webhook-triggered): Incoming orders (supporting multiple items per order) are received via webhook, split into individual line items, and logged — each order simply appends a new entry to an order log rather than directly overwriting stock numbers.
Live stock calculation: Current stock for each item is computed on the fly, using a formula that subtracts all logged order quantities from the starting count.
Daily stock monitoring (schedule-triggered): A separate scheduled trigger runs once a day, checks every item's live stock against its own custom reorder threshold, and sends a Slack alert for anything running low — so restocking happens before customers ever notice a shortage. I built an AI support triage workflow in n8n that handles incoming emails automatically. When a new email arrives in Gmail, an AI agent first checks whether it's from a genuine customer, then decides whether it needs the owner's attention. The team is alerted on Discord, and urgent emails are forwarded to the owner. A second AI agent then writes a reply, which is sent to the customer automatically. Every email and reply is logged to Google Sheets.
Result: spam and irrelevant emails are filtered out, customers get a fast response, and the owner only sees the emails that need them.
Tools: n8n, AI agents (Mistral), Gmail, Discord, Google Sheets. I built an AI-powered lead qualification workflow in n8n that turns every form submission into a fast, personalised follow-up. An AI agent qualifies each lead as hot, warm, or cold. Hot leads get a custom email written by a second AI agent, and the owner is notified right away. Warm leads get a professional template email, and the owner is notified too. Cold leads get a friendly, low-pressure reply. Every lead is logged to Google Sheets for easy tracking.
Result: every lead gets an instant reply, and the business owner only spends time on the leads that matter.
Tools: n8n, AI agents, Gmail, Google Sheets.