36 customer service chat response templates (and when to use each)

Updated November 6, 2026

Copyable customer service chat response templates for greetings, delays, refunds, angry customers, handoffs and closings, each with a note on when it fits and what to change before you send it.

These customer service chat response templates are written to be copied, edited once for your business, and saved as canned replies. Each one has a bracketed slot for the specific detail, because a template that names the customer's actual problem reads like a person wrote it, and one that does not reads like a template. Numbers are grouped by situation, with a short note on when to use each. Replace the brackets, cut anything that does not sound like your team, and keep the whole library under one page if you can.

Greetings and first replies. (1) 'Hi [name], thanks for reaching out. I'm [agent] and I'll help you with this.' Use when a chat arrives with no question yet. (2) 'Hi [name], I can see you're asking about [topic]. Give me a moment to check.' Use when the question is already there; naming it proves you read it. (3) 'Hi, thanks for waiting. I'm picking this up from our assistant and I've read what you wrote about [issue].' Use for every handoff from a bot, so nobody repeats themselves.

Clarifying questions. (4) 'To make sure I get this right: is it [option A] or [option B]?' Use when the question has two likely meanings; a choice is faster to answer than an open question. (5) 'Could you send me your order number? It's in the confirmation email and starts with [format].' Telling them where to find it saves a round trip. (6) 'Could you share a screenshot of what you're seeing? It helps me spot the problem faster.' Use for anything visual or any error message.

Delays and holds. (7) 'I need about [5] minutes to check this with [team]. I'll reply right here.' Always give a number. (8) 'This is taking longer than I said, sorry. I should have an answer in [10] more minutes.' Send before the first estimate runs out, not after. (9) 'This needs [team] to look at it, and they work [hours]. I've passed on everything you told me, and you'll get a reply by [time] at [email]. You don't need to keep this window open.' Use when the answer will not come during this chat.

When you don't know yet. (10) 'Good question, and I don't want to guess. Let me confirm with [team] and come back to you by [time].' Honest uncertainty beats a confident wrong answer every time. (11) 'I've checked and we don't have a documented answer for that yet. I've flagged it so we add one, and here's what I can tell you now: [partial answer].' Use when your help content has a gap, and then actually fill the gap.

Refunds and returns. (12) 'Yes, you can return this. Here's how: [steps]. Your refund goes to [payment method] within [X] business days after we receive it.' Lead with the yes. (13) 'I've issued a refund of [amount] to your [card ending/method]. Banks usually take [X] business days to show it.' Use the moment it is done, with the amount. (14) 'This order is outside our [30]-day return window, so I can't do a standard refund. What I can do is [store credit / repair / partial refund]. Would that work?' The no goes first, the nearest yes second.

Billing and pricing. (15) 'You were charged [amount] on [date] for [plan/item]. That's [reason].' Use when someone does not recognize a charge; specifics calm people faster than reassurance. (16) 'I can see the double charge. I've reversed the extra [amount], and you'll see it back in [X] business days.' (17) 'Our pricing for that is [price and basis]. The full details are on our pricing page, and I'm happy to help you work out which plan fits.' Never paraphrase prices loosely; quote them.

Saying no without losing the customer. (18) 'We don't support [request], and I don't want to leave you guessing. The closest option is [alternative].' (19) 'I can't change that from my side because [real reason]. Here's what you can do: [next step].' A real reason, not 'policy'. (20) 'That isn't something we offer right now. I've passed your request to the product team; I can't promise it, but they do read these.' Only say they read them if they do.

Angry or frustrated customers. (21) 'You're right to be frustrated. [Specific thing] shouldn't have happened, and I'm going to fix it now.' Agree with what is true; skip the generic apology. (22) 'I'm sorry this is the third time you've had to contact us about [issue]. I'm staying on this until it's resolved.' Acknowledge the history if you can see it. (23) 'I want to fix this, and I can do that faster with [order number/detail].' Redirects energy toward the solution. (24) 'I understand. I'm not able to continue while the language is abusive, but I do want to help with [issue] when you're ready.' Use only for real abuse, and keep it calm.

Handoff to a human or another team. (25) 'I'm connecting you with [name/team], who handles [topic]. I've passed on the details, so you won't need to repeat anything.' (26) 'Our team is offline right now; they're back at [time, time zone]. Leave your email and they'll reply there first thing.' Use after hours instead of letting a chat hang. (27) 'This needs a specialist. [Name] will pick it up here in this chat within [time].' Same thread, real time estimate.

Technical troubleshooting. (28) 'Let's try one quick thing first: [step]. Tell me what you see after that.' One step at a time in chat, never a wall of ten. (29) 'Thanks, that tells me it's [likely cause]. The fix is [steps].' Explain what you learned from their answer. (30) 'I've reproduced it on our side, so it's not something you did. I've logged it with our engineers as [reference], and I'll update you here or by email when it's fixed.'

Sales and pre-purchase questions. (31) 'Yes, [product] does [feature]. Here's how it works for [their use case].' Answer before you ask for anything. (32) 'Honestly, if you need [requirement], we might not be the best fit; [what we do instead].' Candor here prevents churn later. (33) 'Want me to send you the comparison or book a short call with [name]? What's the best email?' Ask for contact details after you have helped, not before.

Closing the chat. (34) 'Is there anything else I can help with today?' Use once, at the end, and only if the main issue is solved. (35) 'Glad that's sorted. If [issue] comes back, reply to this chat and it'll come straight to us.' Tells them the door stays open. (36) 'Thanks for your patience with this one, [name]. Here's a summary: [what was done, what happens next, by when].' Use after long or complicated chats; the summary prevents a second contact.

How to keep templates from sounding canned. Read each one out loud; if you would not say it to a customer at a counter, rewrite it. Keep every template to two or three sentences. Never send two templates back to back without a line of your own in between. Review the library every quarter and delete anything unused, because a long library is a library nobody searches.

Where AI changes the template list. On an AI-first chat, the repetitive answers (hours, shipping times, how-to steps) are better handled by the AI answering from your help content, so the human template library can shrink to the hard moments: refunds, exceptions, angry customers and handoffs. Your best templates are also good raw material for the knowledge base. A crisp, accurate answer about your return window is one review away from being content the AI can use around the clock.

Where ZChat fits. In ZChat, canned replies sit in the inbox composer for the human side, while the AI answers repetitive questions from your website crawl, files, help articles and Q&A pairs, and hands off to your team in the same thread with the transcript attached, which is exactly what templates 3 and 25 assume. Resolved tickets can be turned into Q&A suggestions for review, so your best human answers can become AI answers once someone approves them.

Every template needs a slot for the customer's specific detail; naming the problem is what makes it read as human.

The templates that matter most are the pressure moments: delays with real times, refunds that lead with the decision, a no with the nearest yes, and handoffs that carry context.

On AI-first chat, let the AI handle repetitive answers and keep the human library short and sharp.