WhatsApp engine
Why this matters for your business
WhatsApp is the most engaged channel in e-commerce. Open rates typically 80–95%, reply rates 6–15% vs. email's 0.4%, read time within 3 minutes of send for most messages. But it's also the channel with the most rules — Meta's 24-hour customer-initiated window, template approval requirements, opt-in language standards, and aggressive policy enforcement on spammy patterns. Get it wrong and your account gets paused; get it really wrong and Meta bans the entire WhatsApp Business account.
The WhatsApp engine handles all of that. You get four tiers (P1 through P4) covering the realistic scope of what a brand sends on WhatsApp — from order confirmations through marketing broadcasts — with built-in template management, 24h window tracking, opt-in handling, frequency caps tighter than other channels, and Meta-policy compliance baked in. You focus on the message; the engine handles the platform compliance.
The unlock is being able to use WhatsApp at scale without risking your business account. Most brands either avoid WhatsApp (leaving the engagement on the table) or use it recklessly (and get suspended). The engine makes the third path — disciplined, compliant, scalable — actually viable.
What this typically unlocks
| Outcome | Result |
|---|---|
| WA open rate | 85-92% typical |
| WA reply rate | 6-15% vs. email 0.4% |
| Cart-recovery via WA | 3× the conversion of email-only |
| WA Business account suspensions | near 0 with engine compliance |
| Time to set up + verify | 30-60 minutes |
| Per-customer WA cost | $0.01-$0.05 depending on country |
What you actually get — the four tiers
| Tier | Use case | Window required? | Approval required? |
|---|---|---|---|
| P1 — Transactional | Order confirmation, shipping, OTP | Anytime (utility template) | One-time template approval |
| P2 — Interactive | Cart recovery, replenishment with quick-replies | Anytime (utility/marketing template) | Per-template approval |
| P3 — AI Concierge | Real-time conversational support | Inside 24h window only | No (you write the response) |
| P4 — Marketing | Newsletters, flash sales, broadcasts | Anytime (marketing template) | Per-template approval |
Each tier has its own setup pattern, its own compliance rules, and its own sweet spot.
P1 — Transactional (start here)
Order confirmations, shipping updates, OTP verification. Lowest-risk, highest-trust use of WhatsApp. Customer expects these messages — that's why open rate hits 95%+.
| Setup | Effort |
|---|---|
| Approve 3-5 utility templates with Meta | ~1 week (Meta review) |
| Wire to Shopify webhooks (orders/create, fulfillment, etc.) | 30 minutes |
| Test in WA Business Manager sandbox | 30 minutes |
Typical first-week impact: 90% of orders confirmed via WhatsApp instead of (or in addition to) email. Customer support tickets drop 30-40% because customers see updates in real-time on their phone.
P2 — Interactive (highest revenue per send)
Cart recovery + replenishment with quick-reply buttons ("Continue checkout", "Pause subscription", "Yes, send tracker"). Customers tap; the engine handles the response logic.
Why this tier matters: quick-reply buttons convert 3-5× higher than plain text or email. The friction of typing is the friction of conversion; remove it and conversion soars.
| Setup | Effort |
|---|---|
| Approve interactive templates with Meta | ~1 week |
| Wire to journey workflow (cart-recovery, replenishment) | 1 hour |
| Configure quick-reply response logic | 1 hour |
P3 — AI Concierge (Pro+)
Real-time conversational answers. Inside the 24h window (customer initiated), the AI can respond with answers to "where is my order?", "do you have this in size M?", "can I return this?" — sourced from your store's data + product catalog + order history.
| Setup | Effort |
|---|---|
| Connect AI provider (defaults to platform's) | 5 minutes |
| Train on your FAQ + return policy (upload PDFs) | 30 minutes |
| Set escalation rules ("if confidence < 0.7, hand to human") | 30 minutes |
Operational impact: 60-75% of inbound WA queries handled without human; the rest escalate cleanly with full conversation context.
P4 — Marketing (use sparingly)
Newsletter-style or flash-sale broadcasts. Marketing templates must be approved by Meta and follow strict copy rules (no all-caps "FREE", no aggressive urgency, etc.). Per-customer fatigue cap is 2/week by default — much tighter than email.
Why tight cap: WhatsApp is sacred in many cultures. Meta's policy aligns: brands that over-send on marketing get flagged. The 2/week cap is the engineering floor; the cultural floor might be lower in some markets.
| Setup | Effort |
|---|---|
| Approve marketing templates per campaign | ~3 days each |
| Configure broadcast audience + segmentation | 30 minutes |
How the 24-hour window works
WhatsApp's central rule: conversations are customer-initiated. Once a customer messages your business, you have 24 hours of "open conversation" where you can send free-form messages. After 24 hours, you can only send approved templates.
The engine tracks the window per customer + auto-uses templates outside the window when available. You don't manage this manually.
Real merchant scenarios
Scenario A — DTC brand sees +280% reply rate vs. email
Setup. $4M brand. Adds WhatsApp P1 + P2 (transactional + interactive cart recovery) to existing email program.
60-day result:
| Metric | Email-only baseline | WhatsApp added |
|---|---|---|
| Cart recovery rate | 12% | 31% |
| Reply rate (any) | 0.4% | 14% |
| Customer-initiated questions | 8/week | 31/week |
Surprise: the 31 customer-initiated questions/week became a sales channel. ~40% of those questions converted to orders directly through chat — the customer was researching, the salesperson (now via WhatsApp) closed.
Scenario B — Subscription brand uses P2 for skip/pause
Setup. Subscription box. Pre-WhatsApp, churn was high because customers couldn't easily skip a month.
WhatsApp setup: P2 interactive template fires 5 days before billing: "Skip this month?" with quick-reply buttons [Skip] / [Keep] / [Pause for 2 months].
90-day result:
- Proactive skip rate: 14% of subs
- Of skippers, 88% retained next month (vs. ~30% of silent churners)
- Net subscription churn: −1.8 percentage points
The skip option is the retention play.
Scenario C — P3 AI Concierge replaces 60% of support volume
Setup. $20M brand, 4 support agents. WhatsApp inbound was ~200 messages/day; agents struggled to keep up.
P3 setup: Trained on shop's FAQ + return policy + product catalog. Confidence threshold 0.75 — below that, escalates to human with full context.
90-day result:
- AI handled 62% of WA queries autonomously
- Avg response time: 8 seconds (vs. 4 hours human)
- CSAT: 4.6/5 (vs. 4.5/5 human-only — slight improvement)
- 1.5 of the 4 support agents reallocated to other work
Scenario D — Marketing P4 broadcast for limited drop
Setup. Sneaker brand, drop launch. WhatsApp opted-in: 8K customers.
Approach: P4 marketing template approved 5 days in advance. Send fired at drop time. Daily fatigue cap respected.
Outcome:
- 8K WAs sent
- Read within 5 minutes: 82%
- Tap-to-shop: 31%
- Conversions: 8% (640 orders within 30 minutes of send)
- Revenue: $89K from one broadcast
The same audience via email would have driven ~$28K. WhatsApp generated 3.2× the revenue per send because of the immediacy and engagement.
Scenario E — Brand avoids account suspension via fatigue cap
Setup. Marketing manager wanted to send a 4th promotional WA in one week to boost a slow campaign. Engine refused (cap is 2/week per customer).
Manager appealed: "It's important — let me override."
Engineering's answer: "If we override, here's what happens: ~200 customers will exceed the WA fatigue cap; ~5% historically report as spam to Meta when over-fatigued; that's 10 reports; 3+ reports in 7 days triggers Meta review of your account; if flagged, the entire WhatsApp Business account can be paused."
Manager backed off. Sent the campaign via email instead.
This is the engine's value: the cap exists because it's the right risk-vs-reward calculation.
Best practices
✅ Start with P1 transactional only. Get to comfort with the engine before activating marketing.
✅ Approve templates ahead of time. Meta review takes 1-7 days; don't wait until you need a template.
✅ Use P2 interactive for journeys, not P4 marketing. Quick-reply buttons in journeys convert; marketing broadcasts fatigue.
✅ Train P3 AI Concierge on real customer questions — your support backlog is the best training data.
✅ Trust the 2/week cap. Pushing past it doesn't earn more; it earns suspension risk.
❌ Don't send marketing in the customer-initiated window. The customer messaged you about an order, not for a sale — respect that.
❌ Don't use all-caps "FREE", "URGENT", etc. Meta's auto-flagging will catch it.
❌ Don't skip the opt-in capture. WA-without-opt-in is a spam complaint waiting to happen.
❌ Don't disable the fatigue cap. Ever.
Plan tiers
| Capability | Free | Starter | Pro | Agency | Enterprise |
|---|---|---|---|---|---|
| WhatsApp Business connect | — | ✓ | ✓ | ✓ | ✓ |
| P1 transactional | — | ✓ | ✓ | ✓ | ✓ |
| P2 interactive | — | ✓ | ✓ | ✓ | ✓ |
| P3 AI Concierge | — | — | ✓ | ✓ | ✓ |
| P4 marketing broadcast | — | ✓ | ✓ | ✓ | ✓ |
| Multi-template management | — | ✓ | ✓ | ✓ | ✓ |
| 24h window tracking | — | ✓ | ✓ | ✓ | ✓ |
| Custom escalation rules | — | — | ✓ | ✓ | ✓ |
| Multi-WABA accounts | — | — | — | ✓ | ✓ |
See also
- Communications overview
- Email channel
- SMS channel
- Chatbots
- Sales engine journeys — uses WA inside multi-channel flows
- Sales engine GDPR & consent — opt-in audit