WhatsApp chatbot for e-commerce: when it pays off, and when it doesn’t
Most "WhatsApp for commerce" pitches you'll hear over-promise. Here's how to think about it honestly — what's worth automating, what's a trap, and what to ship in the first 30 days.
Start with the question, not the channel
The cheapest WhatsApp bot is the one you don't build. If your support team can answer 90% of your inbound DMs in under five minutes from a shared inbox, automation is mostly cosmetics. The lever is volume, not channel.
Three things almost always pay off
- Order status look-ups. Customers want "where's my package" answered in three seconds. A bot that pulls from your shipping API does that for ~₹0.30 per query, vs. an agent at ₹15–20.
- Abandoned-cart nudges. A 24-hour follow-up message recovers 8–12% of carts in our data. Cheap, high-yield, low risk.
- Returns triage. Bot collects order ID, reason, and photo — routes to the right team. Saves 4–6 minutes per ticket.
Three things to think hard about
- Recommendations. If your bot says "you might also like X," and X is wrong, you damage trust. Recommendation in chat needs the same rigor as on your PDP.
- Conversational checkout. Cool demo, terrible UX. Most users still prefer a checkout page.
- Promotional broadcasts. WhatsApp's policy is strict. One bad campaign can get your number flagged.
What to ship first
Order status, returns triage, and a single abandoned-cart message. Wire it to your existing systems. Don't custom-train a model yet — use intent classification with a handful of templates. Measure first-response time and human-handoff rate weekly. Iterate from there.
Working on this? Tell us about your stack and we'll point you at the right starting move.