Canned replies

Canned replies (called Canned responses in the dashboard) are reusable text snippets your human agents can drop into a conversation by typing a slash-shortcut. They save typing on the questions you answer over and over - shipping times, return steps, sizing guidance. Find them under Settings → Behavior → Canned Replies, or go to /app/<storeId>/settings/canned.

How they expand

In the inbox composer, type /shortcut and the matching canned response’s body expands into the message box. They’re a tool for the people working the inbox, not something the AI sends on its own.

Anatomy of a canned response

Each entry has four fields:

Shortcut - the trigger you type after the slash, e.g. thanks or returns. Required. Limited to letters, numbers, hyphens, and underscores (no spaces), up to 64 characters.
Title - a human-readable label so you can tell responses apart in the list. Required, up to 255 characters.
Body - the actual text that gets inserted. Required. Multi-line is fine; line breaks are preserved.
Locale - a language code, defaulting to en. This lets you keep separate versions of the same shortcut per language.

Adding and updating responses

The page lists your existing responses at the top - each shows the /shortcut chip, the title, the locale, and the body. Below that, the Add or update response form lets you create a new one: fill in shortcut, title, locale, and body, then click Save response.

Create your first response

Pull up the three questions your team answers most - those are the snippets that repay the setup time fastest.
In the Add or update response form, give each a short shortcut (e.g. ship) and a clear title so teammates can tell entries apart.
Write the body exactly as it should land in the chat, then click Save response.
Open a conversation in the inbox and type /ship to confirm it expands before telling the team it exists.

Saving the same shortcut overwrites it

Shortcuts are unique per locale and are upserted on save. If you save a response using a shortcut + locale that already exists, it updates the existing entry rather than creating a duplicate. To keep two variants, give them different shortcuts or different locales.

Deleting a response

Each row in the list has a Delete button. Removing a canned response takes effect immediately; agents can no longer expand that shortcut. There is no separate confirmation step, so delete deliberately.

Practical tips

Keep shortcuts short and memorable (ship, refund, vip) so agents can fire them from muscle memory.
Use a consistent naming convention across the team - agreeing on /returns vs /return-policy avoids near- duplicates.
For multilingual stores, reuse the same shortcut with different locale values and check which localized result the composer offers before sending. Do not assume automatic translation or a fallback language.
Canned bodies are static text. If you want answers that pull live order or policy data, that’s the AI agent’s job - see How the AI works.

Canned replies vs. AI replies

Canned replies are for human agents who want to insert a vetted snippet quickly. The saved wording is fixed until you update it; agents can review and edit an inserted draft before sending. The AI agent, by contrast, drafts a fresh answer grounded in your catalog and policies. Many teams use both: let the AI handle the bulk, and keep a handful of canned responses for the exact-wording moments (refund approvals, escalation acknowledgements) where consistency matters.

Canned responses are inserted from the inbox. To set response- time expectations for that inbox, see SLA targets. For when a conversation should move from AI to a human, read Human handoff.

Test a localized shortcut

Create a reviewed English and French shipping snippet with distinct wording. Open an appropriate test conversation, insert the shortcut, and verify the chosen language and actual dispatch terms before sending. Deleting or overwriting a saved row has no documented restore action here; keep approved copy separately if you need version history.