Shopify email triage means sorting every inbound message by urgency, intent, and revenue risk before you reply to any of them. Five categories cover roughly 90% of store email, and a handful of signals, such as order value and the word 'dispute', identify the few messages that can cost you a chargeback.
Key takeaways
- Five categories (WISMO, returns, damaged items, pre-sale, billing) cover about 90% of inbound email.
- Rank by four signals: order value, days since purchase, carrier status, and risk words.
- Use three or four tiers, not seven.
- Disputes and high-value refunds always go to a human first.
It is Monday, 8:10 a.m. You open Gmail and see 140 unread messages. Somewhere in there are three customers who will file chargebacks by Friday if nobody answers them today. The subject lines do not say which three. 'Re: my order' appears 22 times.
This guide gives you a ticket prioritization framework you can run by hand in Gmail, plus an honest look at where customer support triage with AI saves time.
What does triage mean for a small store?
Support ticket triage for a small business is a three-question check applied to each message: how urgent is it, what does the customer want, and how much money is at risk. A good pass takes under 10 seconds per message and ends with a label, a priority, and an owner.
Triage is not replying. It is deciding the order of replies. Without it, you answer in arrival order, which means a 'where is my parcel?' message from Sunday night beats a PayPal dispute warning that arrived at 7 a.m.
The three questions map to three outputs:
- Urgency sets the reply deadline.
- Intent sets the category and the playbook.
- Revenue risk sets who is allowed to answer.
For a store handling 20 to 40 emails a day, this is a 15-minute morning routine. Past roughly 100 a day, doing it by hand starts to eat the hours you saved by being small.
Which 5 categories cover roughly 90% of ecommerce support email?
Five categories account for roughly 90% of inbound ecommerce support email: WISMO ('where is my order'), returns and exchanges, damaged or wrong items, pre-sale questions, and billing issues. The remaining 10% is wholesale requests, press, spam, and oddities. Build your inbox around these five first.
- WISMO: tracking, delays, 'it says delivered but it is not here'. Usually the largest bucket by volume.
- Returns and exchanges: size, change of mind, refund timing.
- Damaged or wrong items: photo evidence, replacement or refund decisions.
- Pre-sale questions: shipping times, materials, stock, discount codes. These are lost sales if left for a day.
- Billing: double charges, failed payments, invoice requests, PayPal or card disputes.
Good ecommerce support inbox organization starts with one label per category and nothing more. Add sub-labels only after a category passes about 30 emails a week.
How do you rank by urgency?
Rank urgency with four signals: order value, days since purchase, carrier status, and risk words in the message. Score each signal as low, medium, or high, and any single high score on a risk word moves the ticket to the top tier. A 'dispute' mention outranks everything else.
- Order value: a €240 order deserves faster attention than a €12 one. Pick a threshold, for example €100, and treat anything above it as elevated.
- Days since purchase: a message on day 2 is routine. One on day 12 with no delivery is a chargeback in progress.
- Carrier status: 'delivered' with a missing parcel, or 'exception' and 'returned to sender', raises urgency.
- Risk words: 'dispute', 'chargeback', 'PayPal', 'bank', 'scam', 'lawyer', 'report'.
Citation capsule: how urgency scoring works. Chargebacks usually come from slow or missing answers rather than from the original problem. A practical urgency score combines four inputs a store owner can see in seconds. Order value above a chosen threshold, such as €100, adds risk because the loss is larger. Purchase age beyond the promised delivery window, for example 10 days against a 7-day estimate, signals an unresolved problem. A carrier status of 'delivered' paired with a customer saying the parcel is missing is the classic dispute setup. Finally, words like 'dispute', 'PayPal', or 'bank' show the customer is already considering a payment reversal. Any one high signal moves a ticket up a tier. Two or more means a human should read it within one hour.
What should a human always see first?
A human should always see four things first: any message mentioning a dispute or chargeback, refunds above your set threshold, damaged-item claims with photos, and angry messages from repeat customers. These carry money or reputation risk that a draft reply cannot safely settle on its own.
Add two more for judgment calls: legal or regulatory language, and anything where the customer reports a safety problem with a product.
The rule is simple. If a wrong answer costs more than the time to read the message, a person reads it. A 90-second human check on a €300 refund is cheap. A wrongly denied one that becomes a dispute costs the order, a fee, and the customer.
Which tier, response time, and owner fit each ticket type?
Three urgency tiers are enough for most small stores: Tier 1 within 1 hour, Tier 2 within 8 business hours, Tier 3 within 24 hours. The table below assigns each common ticket type a tier, a target first-response time, and an owner.
| Ticket type | Urgency tier | Target first response | Who handles it |
|---|---|---|---|
| Dispute or chargeback mention | Tier 1 | 1 hour | Human only |
| Refund above threshold (e.g. €100) | Tier 1 | 1 hour | Human only |
| Damaged or wrong item | Tier 1–2 | 4 hours | Human review of AI draft |
| WISMO with carrier exception | Tier 2 | 4 hours | Human review of AI draft |
| Standard WISMO (in transit) | Tier 3 | 24 hours | AI draft, auto-resolve if low risk |
| Return or exchange request | Tier 2 | 8 hours | AI draft, human review |
| Pre-sale question | Tier 2 | 2 hours | AI draft |
| Billing: invoice or receipt | Tier 3 | 24 hours | AI draft |
| Billing: double charge | Tier 1 | 2 hours | Human only |
| FAQ (shipping policy, sizing) | Tier 3 | 24 hours | AI draft, auto-resolve |
Pre-sale questions sit high for a reason. A shopper asking about sizing is deciding now, and an answer tomorrow often means a lost sale.
Where does Heylio fit, and can Gmail alone do this?
Heylio handles triage through intent detection and playbooks: it reads each message, identifies the category, and drafts a reply from your knowledge base. Low-risk categories such as WISMO and FAQ can resolve on their own, while refunds above a threshold you set go to a human.
It connects to Gmail, Outlook, Instagram DMs, WhatsApp, and a chat widget. For Shopify, it can look up order status, cancel orders, refund, apply discounts, change addresses, and issue gift cards. Handing a conversation to a human is free. You pay only per AI resolution, meaning a conversation the AI closes itself. Plans start at a Free tier with 25 lifetime resolutions, so you can test it on real mail.
It also tracks your corrections and suggests knowledge base and playbook updates, so the sorting improves.
Be honest about the alternative. A shared Gmail inbox with labels and filters does a lot at low volume. Filters on 'where is my order' and 'dispute' plus five labels and a starred-first habit work fine under about 30 emails a day. Tools earn their place when volume, or the cost of missing a dispute, outgrows that.
How do you set this up this week?
You can set up a working triage system in about 60 minutes: create five category labels, write three filters for risk words, set a refund threshold, and block 15 minutes each morning for a Tier 1 sweep. Review the results after seven days.
- Create labels: WISMO, Returns, Damaged, Pre-sale, Billing.
- Add a filter for 'dispute', 'chargeback', 'PayPal', and 'bank' that stars the message and applies a red Tier 1 label.
- Decide your refund threshold and write it down.
- Sweep Tier 1 first every morning, then pre-sale, then the rest.
- After a week, count which category was largest and automate or template that one first.
On Monday's 140 messages, this routine surfaces the three risky ones in the first five minutes instead of the last hour.
Start with the three tickets that cost money
Triage is a ranking exercise: five categories, four urgency signals, three tiers, and a short list of tickets a human always reads first. Set the Tier 1 filter and refund threshold today. Run it by hand for one week. If the same low-risk categories still consume your mornings, test automation on those, not on disputes.