Quick answer: A major IT incident communication workflow names the incident commander, identifies affected services and audiences, establishes an approved message owner, sends factual updates on a defined cadence, and records recovery and post-incident actions. A status agent may remind coordinators about outstanding updates; it must not determine root cause, declare systems safe, disclose sensitive evidence or issue unauthorised customer notifications.
When a critical system fails, technicians are busy restoring service while customers ask what happened and whether they can continue working. The service desk may have partial facts, an account manager may promise an unconfirmed recovery time, and a vendor may send an update that describes only its own infrastructure. These gaps create a second incident: confusion about what anyone knows and who is allowed to say it.
A major incident customer communication plan should reduce that uncertainty. It is not a promise that every update will contain good news. It is a controlled way to deliver accurate information, show what is being investigated and give customers a reliable channel for the next update.
Establish severity and authority before drafting a message
Your organisation’s incident classification should reflect business impact and urgency, not merely the number of support tickets. Consider which services are affected, which customers are exposed, whether essential operations are blocked, and whether data or security risks are suspected. A qualified incident manager decides severity using the approved procedure. Avoid letting an automated summary elevate or downgrade an incident from a single caller’s impression.
Assign an incident commander or major-incident lead, a technical owner, a customer communications owner, a service-desk liaison and an escalation contact for legal, privacy or security issues. One person may cover multiple roles in a small MSP, but the responsibilities should still be explicit. Confirm who may issue external notices, approve estimates and declare service restored. During a fast-moving event, clarity of authority matters more than elaborate organisational charts.
Maintain a verified operational picture
Build a source-of-truth incident record with start time, affected services, reported impact, confirmed scope, next technical checkpoint, current mitigation, customer channels and message approvals. Mark every assertion as confirmed, provisional or unknown. “Customers cannot access the portal” may be verified by monitoring and customer reports; “the database failed because of a cyberattack” is a different and potentially unsupported claim.
Maintain the technical investigation in the organisation’s approved security and service systems. A broadly visible account-management board should display the customer-relevant status and link to appropriately restricted evidence. Do not paste logs with credentials, personal information, exploit details or unreviewed internal allegations into automatic messages or task notes.
Publish updates with a consistent structure
A useful external update answers five questions: What service is affected? What impact has been confirmed? What is the team doing? Is there an action the customer should take? When will the next update arrive? If there is no reliable resolution estimate, say so. Replace speculation with a specific next-update time and an explanation of what is being investigated.
Avoid absolute assurances such as “no data was affected” until the responsible security or privacy personnel can support them. Where reporting requirements may apply, involve the designated legal and compliance team. An operational incident note is not the same as a legally required breach notification. Different audiences may need different messages: senior client contacts, affected end users and internal support teams should share the same verified facts but not necessarily identical detail.
Use a cadence that continues even when nothing changes
The incident lead should set an update rhythm according to severity and contract commitments. The interval may change as the incident develops; it must not be guessed from a generic template. A “no material change; investigation continues” update can still be useful if it preserves trust and specifies the next planned communication. Publish it only after the current facts have been checked.
Assign a primary and backup message owner. If the primary owner is engaged in a technical escalation, the backup can prepare the draft and route it for approval. Keep an update log showing version, audience, channel, author, approver, timestamp and factual changes. This protects against different clients receiving conflicting explanations from separate employees.
A monday.com communications board for incident coordinators
One incident item can connect to audience-specific communication tasks. Useful fields include Incident ID, Commander, Affected service, Confirmed impact summary, Customer group, Communications owner, Approval owner, Update due, Draft link, Approval status, Sent timestamp, Next update and Closure confirmation. Store the technical ticket reference separately and restrict access according to policy.
The communications states should be Draft required → Draft prepared → Factual review → Approved to send → Sent → Next update due. “Sent” only records dispatch; it does not prove that each recipient received or understood the message. Distinguish a service-status update from the formal final incident report. The latter may require deeper review after recovery.
Where automated check-ins are appropriate
An approved scheduled call to the internal communications owner may ask, “Has the next customer update been drafted?” or “Is the vendor’s latest status available?” It can place a timestamped response on the board and alert a human when the due time approaches. Autoza AI for monday.com supports scheduled team status calls and mapped board updates. AI Workflow Design’s custom automation services can be considered for an agreed integration among an incident tracker, approval workflow and communications queue.
Neither tool should automatically generate incident facts from an unverified phone call, notify customers without required approval, declare recovery, or offer incident-specific legal assurances. The human incident team retains ownership of severity, security containment, recovery testing and external statements. A status record saying “vendor reports fixed” is not equivalent to service validation by the MSP.
Example: a customer portal outage with changing scope
Imagine a managed provider receives reports that two client organisations cannot log into a shared portal. Monitoring initially confirms intermittent authentication errors. The commander sets a provisional severity, contacts the relevant vendor and confirms the customer audience. The communications owner issues a checked message that describes the login impact and next update, without naming a cause.
An hour later, the vendor reports a configuration problem and a fix in progress. The MSP does not immediately publish “resolved.” Its technical team tests login and key transactions for affected accounts, then the commander validates recovery status. The approved next update communicates restoration and monitoring. A later retrospective reviews why the service desk initially listed only one client group, whether the update intervals were appropriate and which actions must be assigned. No invented root cause or promised recovery time is needed to make the process useful.
Close the communication loop after technical recovery
Resolution is not simply the disappearance of an alert. The incident leader should check the agreed recovery criteria, service behaviour, backlog of failed transactions and any customer-specific validation. Once approval is obtained, the communications owner can issue a service restoration notice that explains what is known, what may still require monitoring, and when further analysis will be shared if appropriate.
Open a post-incident action register for unresolved vendor explanations, customer follow-ups, monitoring improvements, runbook revisions, communication issues and relevant preventive work. Give each action a named owner, due date and evidence of completion. A retrospective document that sits unread in a folder is not an improvement programme. Where investigations remain open, keep technical and communications statuses distinct.
Measure clarity, not just message volume
Review whether the first approved customer update went out within the organisation’s own target, how often update deadlines were missed, whether messages later required correction, the share of incidents with clear ownership and the age of post-incident actions. Examine feedback from service-desk and customer contacts to learn whether updates answered their real questions. Do not use message count alone as proof that customers were informed.
The strongest process is one in which technicians can focus on containment and recovery, the communications team can say exactly what is verified, and clients always know when they will hear more. Automation supports the rhythm; accountable humans protect the facts.
Next step
Want these follow-ups handled consistently? Explore Autoza AI for monday.com or talk to AI Workflow Design about scheduled, permissioned status follow-ups that leave approvals with the right people.
Frequently asked questions
What belongs in a major IT incident customer update?
Include confirmed affected services, known customer impact, current response actions, any validated customer instructions and the next update time. Do not speculate about root cause, data exposure or a recovery estimate.
Who should approve external incident communications?
The role designated in the organisation’s incident procedure and customer agreements. Security, privacy, legal or executive reviewers may be required for particular incidents. Technical ownership does not automatically grant external-notification authority.
Should customers receive an update when there is no new information?
Often yes, if the agreed communication cadence calls for one. A concise statement that the investigation continues, with the next update time, can be more useful than an unexplained silence.
Can AI determine that an incident is resolved?
No. A routine status caller can collect task progress but cannot replace technical recovery testing, incident command decisions or security and privacy assessment.
When should the post-incident review be closed?
After the authorised owner has reviewed findings, recorded agreed actions and assigned follow-up responsibility under the incident-management process. The service restoration notice and the full retrospective are different deliverables.
Related reading
- Third-Party IT Vendor Delays: A Dependency and SLA Escalation Plan
- Backup Restore Exercises: A Task-Owner Follow-Up Checklist
- Patch Exception Expiry: A Follow-Up Plan for IT Risk Owners
- Privileged Access Reviews: Following Up on Application Owner Attestations
- SaaS License Renewals: A Client Approval and Vendor Follow-Up Checklist for MSPs