guide
40 Canned Responses You Should Delete From Your Help Desk
Every macro library rots. Here is what to take out of yours.
Macro libraries only ever grow. Somebody adds one for an edge case, it stays for four years, and the policy it references changed in 2024.
This is the subtraction pass. Forty things worth taking out, grouped by why they fail.
1. The ones that announce they are a template
These are detectable in under a second and they set the tone for everything after.
| Delete | Because | Use instead |
|---|---|---|
| I hope this email finds you well | Nobody has ever written this to one person | Nothing. Start with the answer |
| Thank you for reaching out to us | Fine once; a tell across a thread | Their name, then the answer |
| We appreciate your patience | Used where you kept them waiting | I am sorry this took [n] days |
| Rest assured | Almost always precedes no action | State the action |
| We value your feedback | Says nothing, costs credibility | What you did with it |
| Please do not hesitate to reach out | Nobody hesitates. They just leave | Reply here and it comes to me |
| As per our policy | Reads as a shield | The reason behind the policy |
| Unfortunately, at this time | Softening that fools nobody | I cannot do X. Here is what I can |
The pattern in all eight: they occupy the position where the answer should be. A customer opens an email to find out something, and the first line tells them nothing.
2. The ones that commit you without meaning to
These create obligations your team cannot see. The second contact is where they surface.
- "We will get back to you shortly" — shortly is whatever the customer decides it means, and they will hold you to it.
- "Our team is looking into this" — if nobody is, this is a lie with a delay attached.
- "This should be resolved soon" — should and soon are both unfalsifiable, which is why they get used.
- "I have escalated this" — to whom, and what happens next? Escalated with no named owner means nothing happened.
- "We are working on a fix" — if it is not on a roadmap, it is not being worked on.
- "I will personally make sure" — you cannot, unless you can, and you probably cannot.
- "This is a known issue" — implies a fix exists. Say whether one does.
- "We have passed this to the relevant team" — a customer cannot chase 'the relevant team'.
Replace every one with a time and an owner. "I will update you by Thursday whether or not there is news" is a promise you can keep, and it is the one customers actually want.
3. The ones that are out of date
Not a phrasing problem. These are the responses that were correct when written and now quietly misinform.
- Any macro naming a price. Prices move; macros do not.
- Any macro naming a timeframe that has changed — returns windows, response times, delivery estimates.
- Any macro linking to a page that has moved. Check every link in the library once a quarter.
- Any macro referencing a feature that has been renamed or removed.
- Any macro naming a specific colleague who has left.
- Any macro referencing a promotion that ended.
- Any macro with a year in it.
- Any macro mentioning a third party you no longer use.
This category is the argument for auditing rather than only adding. A library that nobody prunes fills with responses referencing policies that changed, and a confidently wrong answer is worse than no macro at all.
4. The ones that are actively costing you
The four worth deleting today, ahead of everything above.
Anything that asks a question you can already answer
"Could you provide your order number?" — sent to someone who included it. "Which product is this about?" — when it is on the account. Every one of these adds a full round trip and reads as not having been read.
Anything that closes without resolving
"I am marking this resolved. Let us know if you need anything else." — sent while the customer is still waiting. It converts a support problem into a trust problem, and the reopened ticket arrives angrier.
The reflexive apology
"Sorry for any inconvenience" on a routine question where nothing went wrong. It devalues the apology you will need later, when something actually did.
The upsell in a complaint thread
Any macro that mentions upgrading, attached to a conversation about something being broken. Some help desks suggest these automatically. Turn that off for complaint tags.
How to run the audit
- Export the whole library. Most teams have never seen it in one list and the count is usually a surprise.
- Sort by last used. Anything untouched in six months is a deletion candidate before you even read it.
- Check every link and every number. These are the silent failures.
- Search for the eight phrases in section one and delete on sight.
- For anything referencing a timeframe or an owner, add both or remove it.
- Then stop. Do not replace what you deleted — the library got too big by adding.
The number to aim for
Fewer than you have. A support team of ten does not need four hundred macros; it needs perhaps sixty that are current, and a habit of writing the rest by hand.
The cost of an oversized library is not storage. It is that agents stop reading them, pick the nearest match, and send something slightly wrong — which is the same failure mode as an AI answering from stale documentation, for the same reason.
This connects directly to automation. If you are pointing an AI at your help centre or your macro library, everything above is now training material. Stale macros produce confidently wrong AI answers, and unlike a stale macro, nobody reads those before they send.
Frequently Asked
What canned responses should I delete?
Start with four: anything asking a question you can already answer, anything closing a ticket that is not resolved, reflexive apologies on routine questions, and any upsell attached to a complaint thread.
Why do canned responses sound robotic?
Usually because the opening line says nothing. "I hope this email finds you well" and "thank you for reaching out" occupy the position where the answer should be, and customers notice within a second.
How many canned responses should a support team have?
Fewer than most have. A team of ten needs perhaps sixty current ones rather than four hundred. Past a certain size agents stop reading them and pick the nearest match.
How often should I audit my macro library?
Quarterly, checking every link and every number. Prices, returns windows and response times change; macros do not update themselves.
What is wrong with saying 'we will get back to you shortly'?
Shortly means whatever the customer decides it means, and they will hold you to their interpretation. Give a time and an owner instead.
Do canned responses hurt CSAT?
Stale or badly-phrased ones do, particularly across a thread where identical phrasing is obvious. Current ones used as structure rather than script raise consistency and speed.
Should I delete unused macros?
Anything untouched in six months, yes — before you even read it. If nobody has needed it in half a year it is more likely to be wrong than useful.
How does this affect AI support tools?
Directly. If you point an AI at your macro library or help centre, stale content becomes training material — and unlike a stale macro, nobody reads an AI answer before it sends.
Tools Mentioned
Full reviews, pricing tiers and where each one breaks.
Help Scout
Chosen for simplicity rather than power. AI drafts and summaries, not an autonomous agent.
Freshdesk with Freddy AI
Cheaper than Zendesk with a comparable feature list. The trade shows up in depth rather than breadth.
Front
Treats support as team email rather than tickets, with AI on top as drafts and summaries.
Zendesk AI Agents
The default incumbent, now consolidating the category by acquisition. Strongest if you already live in Zendesk.
You Can Also Look Into
70 Customer Service Email Templates (Free, No Signup Required)
Seventy templates you can send after changing the bracketed parts. Grouped by share of your queue, with a note on which half an AI can handle.
How to Train an AI Support Agent on Your Help Docs (2026)
Why teams with mediocre tools and good docs beat teams with the reverse, and what to fix first.
12 AI Customer Service Mistakes (and How to Fix Each One) in 2026
Best-practice lists are usually a vendor describing its own feature set. These are the failure modes teams actually report, and what prevents each one.
What Is a Good CSAT Score? Benchmarks by Industry (2026)
The cross-industry average, benchmarks for eight sectors, and the three ways a CSAT score gets inflated before it reaches a dashboard.
SOURCES
WRITTEN BY AR · UPDATED 2026-08-04
I read the fine print. Vendor pricing pages, billing definitions, terms, funding filings and acquisition notices — then I do the arithmetic nobody publishes: what a platform actually costs at your volume, what its headline metric is really counting, and who owns it now. I do not run benchmarks, and no page here pretends otherwise.