How E-commerce Brands Cut WISMO Tickets by 80% with AI

WISMO — "Where Is My Order?" — is the single largest ticket category for most e-commerce brands. Industry data suggests WISMO and shipping-status inquiries account for 25–40% of all support volume, spiking to 60%+ during peak seasons.

This case study walks through how a mid-size DTC brand (we'll call them Northline Goods) reduced WISMO tickets by 82% in 90 days using an AI agent with live order lookup — deployed on Mozo.

The starting point

Northline Goods sells home goods DTC with:

• ~12,000 orders/month
• ~3,200 support tickets/month (pre-AI)
• ~38% classified as WISMO or shipping-status
• 4 FTE support agents + seasonal temps during Q4
• Shopify + ShipBob fulfillment + Gorgias helpdesk

Pain points:

• Average WISMO handle time: 4.2 minutes (lookup order → check carrier → paste tracking link)
• After-hours WISMO emails sat 12+ hours unanswered
• Q4 2025: temp hiring cost ~$18K for WISMO alone
• CSAT on WISMO tickets: 72% (lowest of any category)

Leadership's goal: deflect 60%+ of WISMO volume without adding headcount.

Why doc-only chatbots failed first

Northline tried a generic FAQ chatbot in 2024. It answered "What's your shipping policy?" fine. It failed WISMO because:

1. No live order data — "Your order ships in 3–5 days" isn't helpful when the customer asks about order #48291
2. No carrier integration — tracking numbers live in ShipBob, not the help center
3. No escalation with context — when the bot failed, customers re-submitted via email with no conversation history

Resolution rate on WISMO: 11%. The bot was abandoned after Q4.

The Mozo approach: RAG + custom actions

Northline relaunched in January 2026 with a different architecture:

Layer 1: Policy knowledge (RAG)

Training sources:

• Shipping policy, international delivery, lost package procedures
• Return and exchange windows
• Holiday shipping cutoff dates (Q&A pairs updated seasonally)

This handles "Do you ship to Australia?" and "What's your return window?" — roughly 30% of the original WISMO bucket (customers asking policy before they have an order issue).

Layer 2: Order lookup action (API)

A custom action calls Northline's internal order API:

Customer: "Where is order NL-48291?"
Agent: [calls lookup_order action with order_id=NL-48291]
API returns: status, carrier, tracking_url, estimated_delivery
Agent: "Order NL-48291 shipped via UPS on Feb 3. Track it here: [link]. Estimated delivery: Feb 7."

The action handles:

• Order not found → ask for email to verify
• Unshipped → explain processing time from policy context
• Delivered → confirm delivery date, offer return link if needed
• Exception/delay → create Gorgias ticket via webhook with full context

Layer 3: Escalation with context

When the action returns an edge case (damaged in transit, wrong item, 5+ days late), the agent:

1. Collects confirmation details
2. Fires a webhook to Gorgias with conversation transcript + order data
3. Tells the customer: "I've flagged this with our team — you'll hear back within 4 hours"

No customer repeats their order number to a human.

Rollout timeline

Results after 90 days

ROI calculation:

• Mozo Business plan: $99/month
• Avoided Q4 temp cost: ~$18,000
• Recovered agent time: ~160 hours/month redirected to complex tickets
• Payback period: < 1 week

What made it work

1. Order lookup, not just doc answers

WISMO is a data question, not a doc question. The custom action was the inflection point — doc-only coverage plateaued at ~30% deflection.

2. Deployment on high-intent pages

Placing the widget on the order confirmation page and shipping help page captured WISMO at the moment of anxiety — before customers opened email to contact support.

3. Weekly failure review

Every Monday, the support lead reviewed the 10 WISMO conversations the agent escalated. Patterns became Q&A pairs or action logic improvements:

• "My tracking says delivered but I don't have it" → new escalation rule
• "Can I change my shipping address?" → new Q&A pair with cutoff times

4. Seasonal content updates

Holiday shipping deadlines were added as Q&A pairs two weeks before peak — preventing the accuracy dip most bots see in Q4.

Lessons for other e-commerce brands

You don't need perfect data on day one. Northline's order API returned basic status initially. They added carrier ETA and delivery confirmation in week 6.

Keep humans on exceptions. 82% deflection means 18% still need people — damaged goods, fraud concerns, VIP customers. The win is removing the repetitive 82%.

Measure WISMO separately. Blended deflection rates hide whether your agent actually solves the highest-volume problem.

Integrate with your helpdesk. Webhook escalation to Gorgias/Zendesk/Intercom means agents pick up with full context — no "please repeat your order number."

Could this work for your store?

If WISMO is >20% of your ticket volume, you're likely a strong fit. Requirements:

• Order data accessible via API (Shopify, WooCommerce, custom backend)
• Documented shipping/return policies for the RAG layer
• A helpdesk or inbox for escalations

Mozo connects to order systems via custom actions and to helpdesks via webhooks.

Get started

• Build an AI support agent — general setup guide
• Train on your knowledge base — content best practices
• Start free on Mozo — test with your shipping FAQ before connecting order APIs

Northline's results aren't unique — they're what happens when you match the right architecture (RAG + actions) to the right problem (WISMO). The tickets that used to eat your queue are the same ones AI resolves best.