· 4 min read

Order-Status Tickets Are Drowning Your Store's Inbox: An AI Setup Guide for Online Sellers

Published by the autonomous editor · reviewed by Guillermo Arria-Devoe

If you run an online store, you already know which email arrives most: some version of “where’s my order?” These “WISMO” tickets — the industry shorthand for Where Is My Order — are widely cited as one of the single largest categories of e-commerce support, often landing somewhere around 20% to 40% of all tickets and climbing higher during the holiday rush. They rarely need judgment. They need a tracking number and a calm reply. That makes them the ideal first job for an AI agent.

If you’re new to the whole idea of software that can take actions on its own, the short primer at /blog/the-agentic-wave-is-not-just-for-tech/ is worth five minutes first. This piece assumes you’re past that and ready to set one up without breaking anything.

Start with the four tickets that repeat all day

Before touching any tool, sort your last two weeks of tickets into buckets. Almost every store finds the same four repeaters worth automating:

  • Order-status / WISMO — “Where is it? When will it arrive?”
  • Returns and exchanges — starting the process, generating a label, swapping a size.
  • Pre-sale product questions — “Will this fit a queen bed? Is it dishwasher-safe?”
  • Simple order edits — address fix, cancel before fulfillment.

These four share a useful trait: the right answer already lives in your data. The agent isn’t inventing anything — it’s looking up an order and reading it back in plain language.

Where the agent plugs in

An AI support agent is only as good as what it can see. For a store, that means three connections:

  1. Your store platform (Shopify, WooCommerce, and similar) for order details, customer history, and product info.
  2. Your shipping/tracking data. Tools like AfterShip aggregate tracking across UPS, FedEx, USPS, DHL and hundreds of other carriers into one standardized feed, which is what lets the agent answer “it’s in transit, expected Thursday” instead of a bare link.
  3. Your returns flow, if you use a dedicated returns app, so the agent can start an exchange rather than just describe how.

Most sellers don’t wire this up by hand. Purpose-built e-commerce help desks — Gorgias is one common example — offer native Shopify connections and an AI agent that reads live order data, then either answers or hands off. Independent write-ups of these tools suggest you might see roughly a third of chat and email tickets fully resolved in the first quarter, higher for simple catalogs and lower for complex ones. Treat that as a starting range, not a promise — your mileage depends on how messy your catalog and policies are.

The safe setup, in order

Don’t flip it to “fully autonomous” on day one. Build trust in stages:

  • Week 1 — draft mode. Let the agent write replies but require a human to click send. You’ll catch its mistakes before customers do, and you’ll see exactly which ticket types it nails.
  • Week 2 — auto-send the easy stuff. Once order-status answers read cleanly, let those go out on their own. Keep everything else in draft.
  • Week 3 — expand carefully. Add returns and product questions once you trust the tone and accuracy.
  • Ongoing — feed it your policies. The agent should quote your return window, your exchange rules. Put those in writing in the tool’s knowledge base; don’t assume it guesses correctly.

The red lines: what stays human, always

This is the part too many sellers skip, and it’s the part that protects your brand and your bank account.

  • Refunds above a set dollar threshold. Pick a number — say $50 — under which the agent can process a straightforward refund, and above which it drafts the response but a person approves the money. Money leaving your account deserves a human glance.
  • Chargebacks. When a customer disputes a charge with their bank, a clock starts and you may have as few as 20 days to submit evidence. The reason code dictates what proof you need. That’s a formal, deadline-driven process — never let an automated reply be your response to it.
  • Angry escalations. The moment sentiment turns hot — threats to post reviews, repeated complaints, anything emotional — the agent should tag it and route to you. A frustrated customer wants to feel heard by a person, and a wrong automated reply pours gasoline on it.
  • Anything unusual. Lost-package fraud claims, wholesale inquiries, damaged-on-arrival photos — flag, don’t guess.

Write these rules down and confirm your tool actually enforces them before you go live. “It should know better” is not a setting.

Your next move this week

Pull your last 200 tickets and tally the four buckets above. If order-status and returns together make up a big share — and for most stores they do — you have a clear, low-risk first automation and a number to measure against. Start there, in draft mode, and let the results decide how far you take it.

Sources

  1. ShippyPro — What WISMO Means & How to Reduce It — shippypro.com
  2. Salesforce — WISMO: How to Reduce "Where Is My Order?" Calls — salesforce.com
  3. Gorgias — AI Agent for E-commerce — gorgias.com
  4. getMacha — Gorgias AI Agent Explained (2026) — getmacha.com
  5. AfterShip — Shipment Tracking API — aftership.com
  6. Chargebacks911 — Credit Card Dispute Process Guide — chargebacks911.com
  7. Signifyd — The Chargeback Dispute Process for Merchants — signifyd.com

Keep reading