Canned Responses for Live Chat: 20 Examples That Do Not Sound Canned

The three rules

1. A canned response is a starting point, not a send button. Add one specific detail from the conversation — the visitor's name, their actual question — before sending. Ten words of personalization defeats the "robot" feeling entirely.
2. Write like your team talks. If nobody on your team would say "We apologize for any inconvenience this may have caused," do not save it as a canned response.
3. Every response that makes a promise needs an owner. "We'll follow up by email today" is only a good canned response if someone actually will.

Greetings

• "Hi {name} — happy to help. Give me a moment to look at your question."
• "Welcome back! I can see your earlier conversation, so no need to repeat anything."
• "Good question — checking that for you now."

Buying time honestly

• "That one needs a minute of digging. Stay in the chat — I'll be right back."
• "I want to give you the right answer rather than a fast one. Can I follow up by email within the hour?"
• "I'm checking with a teammate who knows this area better. Two minutes."

Troubleshooting

• "Let's narrow it down: does this happen every time, or did it start recently?"
• "Could you send the exact error text or a screenshot? That usually gets us to the answer fastest."
• "Try {step} and tell me what you see — I'll stay right here."

Refunds and policy

• "Here's our returns policy in short: {summary}. For your order specifically, let me check what applies."
• "I can start that refund now. You'll see it in 3–5 business days, and I'll send a confirmation to your email."
• "That's outside what I can approve in chat, but I'm escalating it now — you'll hear from {team} by {time}."

The frustrated visitor

• "You're right to be frustrated — that shouldn't have happened. Here's what I can do right now: {action}."
• "Thanks for laying it out clearly. Let me fix the immediate problem first, then make sure it doesn't repeat."

Saying no

• "We don't support that today, and I'd rather tell you straight than string you along. The closest thing we do have is {alternative}."
• "I can't do {x}, but I can do {y} — would that get you where you need to go?"

Closings

• "Anything else while I have you? If something comes up later, this transcript will be waiting in your email."
• "Glad that's sorted. I'm sending a summary of what we did so you have it on record."
• "Thanks for the patience on this one. Rate the chat if you have a second — it genuinely helps."

Make them shared, then make them better

Save these as team-shared canned responses so every agent starts from the same tone, then edit them monthly against real transcripts. The best canned library is written by your visitors: whatever they ask most, in whatever words confuse them least.

Escalation and handover

The moment a chat changes hands is where most of the goodwill is lost, and it is the easiest place for a saved response to help. The rule is that the visitor should never have to repeat themselves and should always know what happens next.

• “This one needs our billing team — I am passing them everything you have told me so you will not have to explain it again. Give me a moment.”
• “I am handing you to a colleague who works on this daily. They can see this conversation, so start wherever you left off.”
• “I cannot solve this from chat, so I am turning it into a ticket. You will get an email with the reference within a few minutes, and a reply by tomorrow.”
• “My shift ends in ten minutes and I do not want to leave you mid-sentence. Shall I hand this to the next agent now, or take your email and have them pick it up in the morning?”

After hours and proactive messages

Saved responses are not only for live conversations. The messages that fire when nobody is available do more work than most teams admit, because they are the ones a visitor reads when they are already slightly disappointed.

• Offline: “Nobody is on chat right now — we are back at 9am UK time. Leave your question and an email and you will have an answer first thing.” Give the real return time, not “shortly.”
• Busy: “We are on with a few people at once, so replies may be a couple of minutes apart. I am here — go ahead and tell me the problem.”
• Proactive on a stuck checkout: “Anything holding you up at this step? Happy to check stock or delivery dates.”
• Proactive on documentation: “This page trips people up around the API key step — want me to walk you through it?”

Notice that each proactive line names a specific reason for appearing. A generic “can I help you?” hovering over every page reads as an interruption; one that references the step the visitor is on reads as attention.

Keeping the library from rotting

A canned-response library decays in two directions at once. Old entries go stale as policies change, and new entries accumulate faster than anyone prunes them, until agents cannot find the right snippet and start typing from scratch — at which point the library is doing nothing but taking up a menu.

Keep it small deliberately. Twenty responses that everyone knows beat two hundred nobody can navigate. Review the list quarterly with the agents who actually use it, and be willing to delete: any entry not used in three months either has a discoverability problem or is not needed. Watch particularly for entries that hard-code a number — a price, an SLA, a return window — because those are the ones that quietly turn into misinformation, and they should carry the owner's name so there is somebody to ask when the policy moves.

The healthiest signal is agents editing snippets before sending. That means the library is a starting point rather than a script, which is exactly what it should be.

One practical note on finding them once the library is large: the console floats your pinned and recently used replies to the top of the list. Why your canned replies reorder themselves explains that ordering, and why renaming a reply quietly costs everyone their shortcut to it.