Recovered 12,000+ abandoned carts in one quarter

ShopSmart BD, a D2C e-commerce brand selling electronics accessories across Bangladesh, was facing a textbook abandoned-cart crisis. Their Shopify store was generating ৳1.2 crore in annual GMV, but 68% of all carts were being abandoned — amounting to over ৳80 lakh in potentially recoverable revenue each year. Their previous "recovery" tactic was a single generic email sent 24 hours after abandonment, which had an open rate of just 6% and recovered fewer than 1% of carts. Worse, the Bangladeshi e-commerce market runs on WhatsApp — customers expect to message before buying, ask about stock, request cash-on-delivery confirmation, and get shipping updates. But ShopSmart had no WhatsApp infrastructure; customer support was a single overwhelmed agent working 9-to-5, and the Facebook Page inbox was a mess of unread messages. Order-status enquiries ("Where is my order?") consumed 70% of support time. Returns and refund requests were being lost in inbox clutter. They needed an automation layer that matched how Bangladeshis actually shop online. The founders had tried two generic WhatsApp marketing tools previously but abandoned both — one couldn't integrate with Shopify inventory, the other kept getting their WhatsApp Business number flagged as spam after large broadcast sends. The technical co-founder was specifically insistent that any new solution integrate cleanly with their existing Pathao and RedX courier APIs so order-status queries could be answered without human intervention, and that the AI chatbot be capable of understanding Romanised Bangla (Banglish) since most of their customers type in that hybrid script.

We built a full WhatsApp Business API automation stack layered on top of their Shopify store. At the moment of cart abandonment, an AI chatbot triggers a personalized WhatsApp message within 15 minutes — referencing the exact product left in the cart, with a one-tap checkout link and a limited-time 5% discount code. If the customer doesn't reply, two more smartly-spaced follow-ups go out over 48 hours with different angles (product benefit, then social proof). The same AI chatbot now handles all inbound WhatsApp messages — answering stock enquiries in real-time, confirming COD orders, and processing "Where is my order?" queries by pulling live data from the Pathao and RedX courier APIs. We integrated a 3-step post-purchase flow that requests a Google review, offers a complementary accessory upsell, and sends shipping + delivery updates automatically. Returns and refund requests are routed to a structured form that captures reason and photo evidence before reaching the human support agent. The AI was specifically fine-tuned to understand Banglish (Romanised Bengali) inputs, which account for roughly 60% of inbound messages. We also implemented a smart-segmentation engine that tags every customer by purchase history, cart value, and engagement tier — enabling targeted broadcast campaigns for new product launches that comply with WhatsApp's 24-hour messaging window policy. The whole stack went live in 3 weeks, with the WhatsApp Business API officially approved through Meta's Business Verification process.