The problem
A 5-store jewellery chain in Rajouri Garden, Delhi (Punjabi wedding market, ticket size ₹30k–₹5L) had three expensive problems compounding each other:
- Manual gold-rate broadcasting: each morning at 10:30am, a staff member copied the opening MCX gold + silver rates, added the shop's making-charge markup, and posted them to a WhatsApp Broadcast list — manually, slowly, often wrong.
- Product query backlog: customers DM'd photos asking "what's this necklace worth in 22k?" or "do you have matching danglers?". Responses averaged 4 hours. Most buyers had moved on.
- Wati bill shock: ₹45,000/month quoted + per-message overage. For a 5-store chain pushing ~40k messages a month, the pricing math didn't work.
What we built
A full Akashvanni deployment with an industry-specific layer for jewellery:
- Automated morning broadcast — Akashvanni pulls live MCX gold + silver spot rates at 10:00am, applies the chain's making-charge formula (22k/18k, by category), renders the shop's standard rate card image with today's prices, and broadcasts to opted-in subscribers at 10:15am. Zero human steps.
- Price-dip alert trigger — if gold drops >1.5% intraday, a second broadcast fires ("Gold just dipped ₹X/g — Diwali booking window closing in 4 days"). This alone drove 40% of the walk-in uplift.
- AI chatbot for product queries — when customers send a photo + question, the bot (a) identifies category (necklace / bangle / ring / pendant), (b) estimates likely weight range and today's rate, (c) shares similar in-stock pieces from the shop's catalogue, (d) offers "visit a store" or "video call with jeweller" CTAs.
- Human handoff — the moment the bot detects purchase intent ("I want to come see this", "what's the discount for cash", "shagun date 15th), the thread routes to the on-duty store's agent inbox with the full conversation context.
- Agent inbox across 5 stores — each store sees only its own leads. Chain manager sees everything. Message templates pre-approved by Meta.
- CRM capture — every opted-in number tagged with first purchase, preferred metal, event (wedding/karwa chauth/diwali), and re-engaged 15 days before each anniversary.
The numbers (first 90 days)
| Metric | Before | After |
|---|---|---|
| Daily broadcast send time | 45 min manual work | 0 min (fully automated) |
| Opted-in broadcast list | 2,100 | 8,500 |
| Messages processed/month | ~15k (manual) | 42k (automated) |
| Avg response time to product query | ~4 hours | 90 seconds |
| WhatsApp platform cost | ₹45k (Wati quote) | ₹17k (Akashvanni) |
| Walk-in lift (vs same quarter prior) | baseline | +23% |
| Attributable to morning dip alerts | — | ~40% of lift |
| Bot-to-agent handoff conversion (visit booked) | — | 18% of handoffs |
Economics
Akashvanni runs ₹2 per sent message with no monthly subscription. At 42k messages that's ₹84k/month gross, but ~80% are broadcast messages priced at Meta's "marketing conversation" rate (₹0.88 each at India local), bringing the effective monthly cost to ~₹17k.
Wati's quoted ₹45k was subscription only — their per-conversation costs on top would have added another ₹10–15k at this volume. True monthly saving: ₹28k–38k.
Walk-in lift at avg ticket ₹85,000 with ~12% conversion on new walk-ins = ~₹3.6L/month in incremental sales attributable to the automation.
Tech stack
- WhatsApp layer: Akashvanni (official Meta BSP, direct connection, no reseller markup)
- Price feed: MCX API for live gold/silver spot, 60-second refresh
- Image rendering: on-the-fly rate-card PNG generation with each store's branding
- Chatbot: AI intent detection trained on jewellery vocabulary (tanmaniya, mangalsutra, haar, kada, polki, kundan, hallmark, 22k/18k BIS)
- Agent inbox: multi-user, per-store scoping, team metrics
- CRM integration: Zoho CRM webhook for every tagged lead
- Compliance: DPDP Act consent capture, Meta opt-in templates, 24-hour service window rules enforced automatically
Setup
Week 1 — Meta Business Manager verification, display name approval, WhatsApp number ported from the chain's existing business line.
Week 2 — MCX rate feed connected, chain's making-charge formula coded, rate-card template built per store brand. Opted-in list migrated from their existing broadcast list via consent re-opt-in flow.
Week 3 — AI chatbot trained on their catalogue (1,400 SKUs with photos + weights + making charges), agent inbox rolled out to 5 store managers, Zoho CRM wired.
Week 4 — soft launch at 1 store. Full chain live at day 30.
"Wati was going to cost us almost ₹50k a month to send messages we were already sending for free on personal WhatsApp. Akashvanni priced it per message, charged us a third, and threw in the AI bot for product queries — which actually turned into a sales channel we didn't have before."
What this means for your retail business
This pattern transfers directly to any retail category where (a) daily or event-driven pricing is a real thing (jewellery, metals, textiles, groceries, fuel, agri-commodities), (b) your customers are already on WhatsApp but you're reaching them manually or on personal numbers, and (c) you're paying any subscription-based WhatsApp platform. We've built variants of this for textile wholesalers, dry-fruit importers and pharmacy chains. Same architecture, different price feed.
Want Akashvanni for your retail chain?
₹2/WhatsApp message. No monthly fees. No per-seat charges. Direct Meta BSP, no reseller markup.
Book a 15-min scoping call →