Service outage communication template for chat, email and social
A service outage communication template for a small business whose booking system, website, ordering, internet, hosting or software is down while the chat and inbox fill up. Eight tabs from first notice to follow-up, plus instructions for your AI agent. Every message carries the same fields: confirmed impact, workaround, next update time, resolution and follow-up. Word template, Excel incident log and a filled example.
Made by Anatolii Ulitovskyi 
Fact-checked by RH Fardin Updated 1. First notice in chat and email: within 15 minutes
Send firstThis service outage communication template starts with one message, sent within 15 minutes of confirming the outage: you saw the failure yourself, or the person fixing it says so. It says what is broken, what still works, who is affected, what to do instead and when the next update comes. No cause, no fix time.
Chat replyHi first name, thank you for telling us. Yes, what is broken has been down since start time. What still works. Who is affected. For now, workaround. We will post the next update here by next update time, even if there is no news yet.
Email subjectService is down: what still works, next update by time
EmailHi first name,
What is broken has not been working since start time today. We are fixing it with who is fixing it, for example our booking provider.
What still works: what still works.
Who is affected: who is affected.
What you can do now: workaround.
Next email by next email time, or sooner if it is fixed. Updates every 30 minutes: where, for example our website chat.
Sorry for the trouble.
Your name, role
Business name, phone
Social postService has been down since start time. What still works. Workaround. Next update here by next update time. Questions: chat link or phone.
If it is not confirmed yetThank you, first name. We are checking a report that service is not working. What do you see on your screen? I will reply here within 15 minutes.
Fill in before sending
- Confirmed impact: what is broken, what still works, who is affected.
- Workaround: one you have tried yourself, or 'none yet'.
- Next update time: a clock time 30 minutes from now. Then log the message.
Never promiseA fix time, a cause, or 'back in a few minutes'.
2. Update with no news yet: give the next update time
Send this every 30 minutes until there is real news. Atlassian recommends never going more than one hour without an update while customers cannot use the service, always saying when the next one comes, and posting 'nothing new' rather than nothing1. The time is the one promise you can keep.
Chat and social updateUpdate time: what is broken is still down. Who is fixing it is still working on it, and we do not have a fix time yet. What still works works normally. Until then: workaround. Next update here by next update time.
Reply to 'Any news?'Not yet, first name, and I would rather say so than guess. Next update by next update time.
If the outage passes 2 hoursUpdate time: what is broken is still down, and who is fixing it expects a longer repair. From now on we post every 60 minutes, next at next update time, and right away if anything changes. Workaround still works.
Fill in before sending
- Time of this update, so customers see it is fresh.
- What changed since the last update, or 'no change': only what the person fixing it told you.
- The workaround again, and the next update time. Set a calendar alert 5 minutes before it.
Never promiseA fix time nobody gave you, or 'shortly'. Never skip an update for lack of news. PagerDuty's guidelines allow fewer updates after two hours, but every update still names the next one2.
3. Update with an ETA: only when confirmed, with a buffer
Give a fix time only when the person fixing it, your provider, developer or IT support, has confirmed it, ideally in writing. Then add a buffer: at least 30 minutes, or half the estimate for a long repair. Customers remember the time you said.
Chat and social updateUpdate time: who is fixing it has found the cause and expects what is broken to work again by ETA with buffer. Until then, workaround. We will confirm here as soon as it works. Next update by next update time.
Email subjectUpdate: service expected back by ETA with buffer
EmailHi first name,
Who is fixing it has found the cause of the what is broken problem and expects it to work again by ETA with buffer today.
Until then: workaround.
What it means for you: for example, bookings made before the outage are kept.
We will email you when it is fixed, or by next email time at the latest.
Your name, role
Business name
If the ETA slipsUpdate time: the fix is taking longer than expected, so ETA with buffer no longer holds. New ETA with buffer, or: we do not have a new fix time yet. Workaround still works. Next update by next update time.
Fill in before sending
- ETA source: who confirmed it, when and how. Note it in the log.
- Next update time: still every 30 minutes, even with an ETA.
Never promiseA time nobody confirmed, or one without a buffer. If it will slip, say so before it passes.
4. Workaround and what still works, by type of outage
Most outages break one thing, not everything. Name what still works in every message, and offer only a workaround you have tested.
4.1Booking system down
Still worksYou are open; existing bookings are usually kept (check with your provider).
WorkaroundBook or cancel by phone, chat or email; check in by name.
SayOur online booking is down, but we are open and all appointments run as booked. To book or cancel, reply here or call phone.
4.2Online ordering down
Still worksPhone orders, pickup, orders placed before the outage.
WorkaroundTake orders by phone or chat and confirm each one in writing.
SayOnline ordering is down. You can still order by phone at phone or here in chat, and orders placed before start time go out as planned.
4.3Website down
Still worksPhone, email, social messages, your shop or office.
WorkaroundA pinned social post and an outage voicemail greeting.
SayOur website has been down since start time, but we are open as usual. Call phone or message us on social channel. Next update by next update time.
4.4Card payments down
Still worksThe service itself; cash or bank transfer if you accept them.
WorkaroundConfirm the order now and send the payment request later.
SayCard payments are not going through right now. Your order or booking is confirmed. We will send the payment request when it works, and you will not be charged twice.
4.5Internet, phone or software down
Still worksMobile phones, mobile data and paper.
WorkaroundA mobile hotspot, a second number, a printed list of today's appointments.
SayOur phone line has been down since start time. Please message us here or email email address; we reply within 30 minutes. Today's appointments run as planned.
5. Angry or urgent customers mid-outage: short replies
Some customers have a real cost: a missed appointment, an order for tonight. Reply in 2 or 3 sentences: acknowledge it, give the facts and the workaround, say who decides the rest.
5.1Angry about the outage
WhenUpset, but asks for nothing specific.
Hand overIf they ask for a person, threaten to leave or are still angry after one reply.
ReplyYou are right to be frustrated, first name, and I am sorry. What is broken has been down since start time; for now, workaround. Next update by next update time.
5.2A deadline today
WhenAn appointment, order or delivery today that the workaround cannot save.
Hand overYes, to the incident lead.
ReplyI understand that what they need cannot wait. I am passing this to name with your details, who can what they can do. Name will reply here by time.
5.3Refund, credit or fee waiver
WhenMoney back, an SLA credit or a waived fee.
Hand overAlways. Only approver decides.
ReplyThat is a fair question. I cannot decide on refunds or credits in this chat, so I have passed your request to approver. You will hear back by time.
5.4Data, payment or safety worry
WhenIs my data, payment or account safe?
Hand overYes, right away.
ReplyThat is an important question, and I do not want to guess. I have passed it to name, who will answer you by time.
Handover ruleHand over for money, a deadline today, a legal threat, data or safety, a threat to leave, or a second request for a person. Everyone else gets the facts, the workaround and the next update time.
6. Resolved: all clear on every channel you used
Send the all clear only after your own test passes (book, order or load the page on a phone) and 15 quiet minutes, on every channel that got the first notice. PagerDuty's guidelines say the final message confirms full recovery and states clearly whether any data was lost2.
Chat and socialFixed: what was broken has worked again since fix time. Lasting effects, for example no bookings were lost. What to do, if anything. We will email who gets the follow-up by follow-up date with what happened.
Email subjectFixed: service works again
EmailHi first name,
What was broken has worked again since fix time today. We tested it ourselves at test time.
- Lasting effects, for example: no bookings were lost and no data was lost
- What you need to do, or: nothing
- Any decision already made, for example no late-cancel fees for the outage hours
We will send a short follow-up by follow-up date with what happened and what we are changing.
Thank you for your patience, and sorry again.
Your name, role
Business name
Before you send
- Your own test passed, then 15 minutes without errors.
- Resolution: fix time, what was fixed, and what was lost, or 'nothing was lost'.
Never promiseThat it will never happen again, or a cause nobody confirmed.
7. Follow-up after the incident: what happened, what changed
Completion checkWithin 1 business day, email everyone affected or who contacted you. Atlassian's outline: acknowledge the problem and apologize, explain what went wrong, explain the fix and how you prevent a repeat, and apologize again1. Add the credit decision and who approved it.
Email subjectWhat happened with service on date, and what we changed
EmailHi first name,
On date, what was broken did not work from start time to fix time. We are sorry: the cost to customers in one sentence.
What happened: plain cause, for example our booking provider had a server failure.
What we did: what you did during the outage.
What we are changing: 1 to 3 concrete changes with dates.
Credit: the credit and who approved it, or delete this line.
If you think you were charged by mistake, reply by date and name will check it.
Thank you for sticking with us.
Your name, role
Business name
Credit request: your terms include oneThank you, first name. Our records show that what was down was unavailable from start time to fix time, duration in total. Under where the rule is written, that gives you credit. Approver has approved it, and you will see it on where and when.
Credit request: no credit in your termsThank you, first name. What was down was unavailable from start time to fix time, and I am sorry for the trouble. Our terms do not include a credit for outages, but I have passed your request to approver, who will reply by date.
Completion checkClosed only when the all clear went out on every channel used, this follow-up was sent and every row of the log is complete.
8. AI agent instructions during an outage
Most outage chats ask three things: is it down, what can I do, when is it back. An AI agent answers them in seconds, but only with the facts you give it. Paste the rules into its instructions and the note into its knowledge base.
Paste into InstructionsOUTAGE RULES (from start time on date until removed)
1. What is broken is down. Answer questions about it only with the facts in the OUTAGE NOTE in your knowledge base. Do not add causes, numbers or times.
2. Never give a fix time unless the note has one. If it has none, say we do not have a fix time yet and give the next update time.
3. In every answer about the outage, include what still works, the workaround and the next update time.
4. Never promise refunds, credits or fee waivers. Hand the chat to the team if the customer asks for these, has a deadline today, mentions legal action, data or safety, or asks for a person.
5. Keep outage answers to 2 or 3 sentences. Apologize once.
Paste into Knowledge BaseOUTAGE NOTE, updated time on date
Broken: what is broken since start time.
Still works: what still works.
Who is affected: who is affected.
Workaround: workaround.
Fix time: ETA with buffer, or none yet.
Decisions: for example no late-cancel fees for the outage hours, or none.
Next update: next update time in this chat and on social channel.
Every 30 minutes, and at the end
- Update the note firstNew time, fix time and next update; save, then post the same facts elsewhere.
- Test itAsk 'Is it down?' and 'When is it back?' and check the answers use the new note.
- After the all clearDelete both blocks, save, and check the AI agent no longer mentions the outage.
Never let the AI agentGuess a fix time, explain a cause nobody confirmed, or offer money.
For small sales and support teams
Let the AI agent answer the outage questions and hand the hard ones to you.
In CRMsoftware.pro you paste the outage note into your AI agent's Instructions and Knowledge Base, so every chat gets the same confirmed facts. Refund, credit and deadline chats come to Live Support. Status pages and outage emails stay in your own tools.
- AI chat 24/7 with a basic knowledge base on the free plan
- Human handover and the unified inbox for website chat, WhatsApp and more
- Launch, $19 a month billed yearly for 8 users; saved replies on Growth, $49
Free plan: 3 users, AI chat, basic knowledge base and basic tasks, no credit card.
Filled example
Outage messages to customers: a yoga studio's booking outage
Ironbark Yoga, a fictional studio with about 600 members, opens at 06:00. On Tuesday 6 October 2026 its booking app fails at 06:10. Theo, the manager, writes the outage note; the AI agent answers chats from it, Ana posts it on Instagram, Theo emails the 46 members booked before noon.
Outage note at 07:48: the fields every message copies
| Broken | Booking app and website booking since 06:10: no new bookings, cancellations or online payments |
| Still works | Studio open, every class runs, bookings made before 06:10 are kept |
| Who is affected | Members booking or canceling online; 46 members booked before 12:00 |
| Workaround | Check in by name at the desk; to cancel, reply in chat or email |
| Next update | 08:00 in chat and on Instagram; next email by 09:00 |
| Fix time | Provider: 08:30 (confirmed 07:24). We say: by 09:00 |
| Decisions | No late-cancel or no-show fees 06:00 to 12:00, approved by Rosa at 07:44 |
| Follow-up | Email by 7 October, 12:00, owner Theo |
Message log, 6 October: 17 messages, 3 channels
| Time | Channel and type | Next update |
|---|---|---|
| 06:29 | Chat: first notice | 07:00 |
| 06:31 | Email: first notice | By 08:30 |
| 06:34 | Instagram: first notice | 07:00 |
| 07:30 | Chat: update with ETA | 08:00 |
| 07:41 | Instagram: update, 11 min late | 08:00 |
| 07:48 | Chat: workaround and fee waiver | 08:00 |
| 08:44 | Chat: all clear | None |
| 08:46 | Instagram: all clear | None |
| 08:49 | Email: all clear | Follow-up by 7 Oct, 12:00 |
| 7 Oct, 11:00 | Email: follow-up | None |
Completion check and CRM record
| Item | Recorded | Status |
|---|---|---|
| All clear on every channel | Chat 08:44, Instagram 08:46, email 08:49 | Done |
| Follow-up sent | Email on 7 October, 11:00 | Done |
| Log complete | 17 rows, no field missing | Done |
| Outage note removed | Instructions and Knowledge Base, 08:45 | Done |
| Outage chats tagged | 41 chats, tag Outage 6 Oct, 3 handed over | Done |
| Follow-up task | Theo, deadline 7 October, 12:00 | Done |
Why it matters
Why a service outage communication template needs a clock
Atlassian recommends never going more than one hour without an update while customers cannot use the product, and always saying when the next one comes, because people left in the dark start to expect the worst1.
PagerDuty's public guidelines are stricter: a first message within 5 minutes of starting the incident response, updates at least every 20 minutes for the first two hours, each with the time of the next one, and a final message only when full recovery is confirmed2. That is the pace of a software company with an on-call team. A small business with one person on the chat needs a pace it can keep: first notice within 15 minutes of confirming, then every 30 minutes.
Atlassian also advises one primary channel that the others point to1. At small-business scale, that is one outage note with the five fields, written by one owner and copied into the chat, the email, the social post and the AI agent's instructions.
1 hour
longest gap between updates that Atlassian recommends
20 minutes
PagerDuty's update interval in the first two hours
5 fields
in every message of this template
Channels
Outage notification email, chat or social: what goes where
Write the five fields once in the outage note and copy them to each channel.
| Channel | What goes there | How often | Owner |
|---|---|---|---|
| Website chat (AI agent and team) | First notice, every update, all clear | Every 30 minutes: update the outage note | Incident lead |
| Email to affected customers | First notice, a confirmed ETA, all clear, follow-up | Each email names the time of the next one | Incident lead |
| Social (Instagram, Facebook, X) | First notice, every update, all clear; pin the latest | Every 30 minutes, same text as the chat | Second person |
| Phone and front desk | Outage voicemail greeting and a desk script | With each update | Front desk |
| Status page, yours or your provider's | Link to it from the other channels | Same as the chat | Incident lead |
If your website is down, the chat widget on it may be down too: then social, email and phone carry the updates.
Excel incident log
What is in the Excel incident communication log
Three sheets: Message log with the 17 messages from the example, Summary and Lists. Shaded columns fill in by themselves. Opens in Excel, LibreOffice and Google Sheets.
| Column | What it does |
|---|---|
| Sent at, channel, audience | When the message went out, where (Chat, Email, Social, Status page or Phone) and to whom. |
| Message type | First notice, Update, Workaround, All clear or Follow-up. |
| Confirmed impact, workaround | What is broken, what still works, who is affected, and what customers can do instead. |
| Next update due | The time you promised. Empty only for an all clear or a follow-up. |
| Owner, sent | Who sent it, from a list you edit, and Yes or No. |
| Next update on time Auto | Yes when the next message on that channel went out by the promised time, with 5 minutes of grace; No when it was late. |
| Complete Auto | No when a required field is empty. |
| Status Auto | On time, Next update overdue, Closed (an all clear, a follow-up or a newer message replaced it) or Not sent. |
The Summary sheet counts messages by type and channel, late and overdue updates, minutes from the first report to the first notice and time to the last all clear. Incident closed shows Yes only when every channel used got an all clear, the follow-up was sent and no field is missing. Sample: first notice after 15 minutes, 1 update 11 minutes late, last all clear after 2 hours 35 minutes.
How to use
How to use this outage communication template
- 1
Prepare before you need it
Name the incident lead, a backup and who approves credits. Save tabs 1 to 6 as drafts and saved replies.
- 2
Confirm, then send the first notice
Within 15 minutes of confirming: fill in the outage note, paste tab 8 into your AI agent, send tab 1 on every channel and log each message.
- 3
Update every 30 minutes
Tab 2 while there is no news, tab 3 once an ETA is confirmed. Update the note first, then the channels. Hand over the cases in tab 5.
- 4
All clear, follow-up, close
Test it yourself, send tab 6 on every channel, remove the outage note, send tab 7 within 1 business day. Close when the Summary sheet says Yes.
Set it up in CRMsoftware.pro
- My Bots > choose the bot > Basics tab: paste the rules from tab 8 into Instructions and the outage note into Knowledge Base. Click Save Bot and test in Try Bot. After the all clear, delete both and save.
- Handed-over chats arrive in Live Support with the status Awaiting agent. Claim the chat, reply, then Resolve. Turn on alerts in Notification Settings.
- Settings > Conversation Tags > New tag, for example 'Outage 6 Oct', and add it to each outage chat so you can find them for the follow-up.
- On Growth, save the chat messages as Saved Replies: Settings > Saved Replies > New reply, Shortcut (for example /outage-update), Title, Reply body, Share with team.
- Tasks > New Task: Task title 'Outage follow-up email', Assignee the incident lead, Deadline the next business day, Priority High, Status summary required on completion, so the credit decision is recorded before Mark done.
AI chat, a basic knowledge base and basic tasks: Free, 3 users. Human handover and Live Support: Launch, $19 a month billed yearly. Saved replies: Growth, $49. CRMsoftware.pro has no status page and does not send outage emails or broadcasts; use your own email tool and social accounts.
After every outage, 15 minutes
- Minutes from the first report to the first notice, against 15.
- Updates sent late, and why.
- Questions the AI agent could not answer: add them to your outage note draft.
FAQ
Questions about outage messages
What should a service outage communication template include?
Five fields in every message: the confirmed impact (what is broken, what still works and who is affected), a workaround, the next update time, the resolution once it is fixed and the follow-up. Add the start time and one owner who sends every update.
How do you write an outage email to customers?
Subject: the service and the next update time. Body: what is broken and since when, what still works, who is affected, what to do now and when the next email comes. Post the 30-minute updates in chat and on social instead.
Can you give an example of a system outage message?
From the filled example: 'Our online booking has been down since 06:10. The studio is open, every class runs and bookings made before 06:10 are kept. Check in by name at the desk. Next update here by 07:00.'
How often should we update customers during an outage?
Every 30 minutes in this template, and never more than an hour apart, as Atlassian recommends1. PagerDuty's own guidelines use at least every 20 minutes for the first two hours2. Each update names the time of the next one, even with no news.
Should we give customers an ETA during an outage?
Only when the person fixing it has confirmed one. Add a buffer of at least 30 minutes, and if it slips, say so before the time passes. Until then, give the next update time instead of a fix time.
How should live chat respond during a service outage?
Confirm the outage in the first sentence, give the workaround and the next update time, in 2 or 3 sentences. Hand refunds, credits, deadlines today, legal threats and data questions to a person.
Do we have to give customers a credit after an outage?
Only if your terms, contract or SLA say so; then apply it without making customers ask. Otherwise the owner decides, ideally once for everyone affected, and the follow-up email says what was decided and who approved it.
Can CRMsoftware.pro send an outage notice to all my customers?
No. CRMsoftware.pro has no status page or broadcast feature; send outage emails from your own email tool. In the app, the AI agent answers outage chats from the note in its Instructions and Knowledge Base and hands money and deadline questions to Live Support (Launch, $19 a month billed yearly).
Related templates
Free tools
Sources
- Update cadence, next update time, one primary channel and the post-incident outline, read 8 October 2026: Atlassian, Incident communication best practices
- First message, 20-minute updates, long incidents and the final message, read 8 October 2026: PagerDuty Incident Response Documentation, External Communication Guidelines
- Plans and prices: CRMsoftware.pro pricing
Give every customer the same facts while you fix the outage.
Start free with 3 users, AI chat and a basic knowledge base. Launch, $19 a month billed yearly, adds human handover and Live Support. Growth, $49 a month billed yearly for 20 users, adds saved replies.