A device buyback business runs on information. Receipts from Apple Store purchases, trade-in records, gift card balances, shipment statuses. The flow is constant, and the margin on each transaction is thin. When that information lives in inboxes and spreadsheets, the business can only move as fast as the team can read and enter it.

The client is a US-headquartered company trading in used and broken iPhones, with offices also in Dubai had reached the point where that constraint was visible. Employees were manually tracking receipts, gift cards, and device inventory into spreadsheets. The same data arrived from multiple sources: PDFs, emails, carrier websites, and Apple Store records and none of which offered a clean, consistent API to pull from. Processing a batch of receipts took weeks. Errors compounded. The manual overhead was growing faster than the team could absorb.

‍

Technogise ran a scoping workshop to understand the full shape of the problem before writing any code. The instinct was to bring everything into one place and automate it but the right approach was to sequence it. Solving one problem at a time, with a quick iterative delivery model, allowed the team to show visible progress while building toward a complete solution.

Gift card validation came first. The organization held multiple gift cards received as refunds from Apple, and checking available balances across all of them was done card by card, manually. Technogise automated the entire balance-checking process. What had previously taken significant staff time could now be done in minutes and any problem with a card was caught immediately rather than discovered during reconciliation.

‍

The largest source of manual work was email. Every day, receipts for purchases, refunds, and trade-in devices arrived in the inbox in different formats like PDFs, HTML emails, scanned documents. Employees had to open each one, read it, classify it, and enter the relevant data into a spreadsheet.

Technogise built an email scraping layer that read incoming messages, matched keywords to classify receipt type, and extracted the relevant data automatically. A PDF parser handled attached documents. The result: purchases, refunds, and trade-ins were separated and recorded without any human involvement. A finance manager dashboard presented a clear breakdown of refunds and spending for reconciliation.

The time spent on receipt processing dropped from weeks to minutes. The team shifted from data entry to work that actually required their judgement.

Two further automations completed the platform. Shipment tracking aggregated status updates from multiple shipping carriers via a combination of APIs and scraping where APIs were unavailable into a single view. The team could plan sales based on actual inventory in transit rather than estimated arrival times.

Device purchasing was fully automated. An employee enters the device specifications and quantity needed; a Puppeteer-driven script completes the purchase through the appropriate channel. The entire acquisition workflow, previously managed manually, runs unattended.

‍

With manual processing removed from the back office, the operational constraint on growth was eliminated. The business had the foundation to scale and targeted a 20 × expansion.