> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tied.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Canned responses

> Reusable text you insert with a slash. The fastest thing in the composer, and the easiest to let rot.

A canned response is a saved piece of text with a short code. Type `/` in the composer, type the code, and the text drops in — where you can edit it before sending.

**Settings → Canned responses.**

## What a canned response is

| Field | What it does |
| - | - |
| **Short code** | What you type after `/`. Keep it short and guessable — you are typing it a hundred times a day |
| **Content** | The text. Rich text, so links and formatting survive |

Insert one by typing `/` at the start of the composer and picking from the list, or typing the code directly.

<Info>
  **The text is inserted, not sent.** It lands in the composer for you to edit. That is the whole
  point — a canned response is a starting draft with the boilerplate already typed, not an
  auto-reply.
</Info>

## Canned response, macro, or agent?

Three things get confused constantly. They are not alternatives; they do different jobs.

| | Does | Use when |
| - | - | - |
| **Canned response** | Inserts **text** into the composer for you to edit | You type nearly the same paragraph often |
| **[Macro](/inbox/macros)** | Runs a set of **actions** — label, assign, resolve, send a message | A situation always needs the same several steps |
| **[AI agent](/agents)** | Reads the conversation and **writes the answer itself** | The question can be answered from knowledge or a system lookup |

The useful test: if you always send the *same words*, that is a canned response. If you always *do the same things*, that is a macro. If you find yourself needing a canned response for something a customer could have been told automatically, that is a [knowledge gap](/agents/knowledge#the-learning-loop) and the agent should be answering it.

## Writing ones people actually use

**Name the code after the situation, not the content.** `/refund-approved` beats `/refund2`. You will search this list under pressure.

**Leave the specifics blank and obvious.** Put a marker where the agent has to type — `[order number]`, `[date]` — so a half-filled response is visibly half-filled rather than quietly wrong. A canned response sent with someone else's order number in it is worse than no canned response.

**Keep them short.** A long canned response gets sent unedited, and a customer can tell. The best ones are two or three sentences that save the tedious part and leave the human part to you.

**Do not can an apology.** *"I'm so sorry for the frustration this has caused"* arriving verbatim to a customer who has had it before is worse than saying nothing. Can the facts; write the feeling.

## Keeping them honest

Canned responses rot faster than anything else in the workspace, because they contain your policies at the moment they were written and nothing tells you when a policy changes.

<Warning>
  **Every time a policy changes, search the canned responses for it.** A response still promising a
  30-day return window after you moved to 14 is a commitment your team is making in writing, at
  speed, without noticing — and your customer has it in writing too.
</Warning>

A reasonable habit: review the list quarterly, and delete anything nobody used. An unused canned response is not free — it is one more row somebody has to scroll past while a customer waits.

<Tip>
  **A canned response you use every day is a candidate for the agent.** If your team sends the same
  paragraph fifty times a week, that is fifty conversations the [AI agent](/agents) could have
  resolved. Put the content in your [Help Center](/help-center) instead of a canned response, and
  the agent can answer from it — which also keeps one copy correct rather than two.
</Tip>

## Next

<CardGroup cols={2}>
  <Card title="Macros" icon="zap" href="/inbox/macros">
    Reusable actions rather than reusable text.
  </Card>

  <Card title="Insert an article" icon="book-open" href="/inbox/insert-article">
    Sending a Help Center article instead of retyping it.
  </Card>

  <Card title="Knowledge and retrieval" icon="book-open" href="/agents/knowledge">
    Turning a repeated answer into one the agent gives.
  </Card>

  <Card title="Keyboard shortcuts" icon="keyboard" href="/keyboard-shortcuts">
    The rest of the composer.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.