If your organisation is still running on-prem Exchange in Australia, the immediate decision is not only how fast to patch or whether to move to Exchange SE. It is which systems still trust, connect to or depend on the mail server, and which of those should be cut away first. That matters more after the September 2026 ISM release, because ASD updated incident guidance to recommend using trusted systems separate from a compromised system, and its email hardening guidance continues to treat email environments as systems that need strong controls and reduced exposure. ASD also states that Exchange environments are continuously targeted for compromise and should be considered under imminent threat. If migration will not land this quarter, the fastest risk reduction comes from isolating the mail-connected systems that extend Exchange compromise into identity, endpoints, applications and operations.
01What changed in September 2026, and why it changes the containment plan
The September 2026 ISM is a current Australian Government cyber security framework, and the published change log notes an important update to incident handling: the control now recommends using trusted systems separate from a compromised system for remediation planning and coordination. In practice, if Exchange is the suspected foothold, you should assume the surrounding communication paths, admin tooling and integrated services may also be exposed.
ASD's current email hardening guidance says hardening email environments reduces the risk of compromise and limits the impact of cyber incidents. Its Exchange-specific guidance goes further, stating that Exchange environments are continuously targeted for compromise and should be considered under imminent threat. For teams still on Exchange 2016 or 2019, this is made worse by Microsoft's support status: both versions ended support on 14 October 2025.
That means the right question is no longer only, 'Is Exchange patched?' It is, 'If Exchange is breached tomorrow, which attached systems let that compromise spread, persist or block recovery?'
- Treat Exchange as a high-value compromise point, not just a messaging service
- Move incident coordination, privileged admin and security monitoring off mail-dependent paths
- Prioritise integrations that carry credentials, privileged workflow or broad internal reach
- Use segmentation to reduce lateral movement, not only perimeter filtering
02Which mail-connected systems should be isolated first
Start with systems whose compromise would either expand attacker reach or slow down containment. The order below is based on business impact and the speed at which you can usually reduce exposure without a full mail migration.
ASD's guidance on network segmentation and segregation says segmentation reduces the potential consequences of compromise. Its implementation guidance also says organisations should segment sensitive data, hosts and services from the environment in which users access external resources, in particular the web, email and other internet services. Applied to Exchange, that means isolating trusted services from the email zone first.
| Priority | System type | Why isolate first | Fast containment move |
|---|---|---|---|
| 1 | Privileged admin workstations, PAWs, jump hosts, incident comms tools | These are needed to investigate and recover. If they depend on Exchange or sit in the same trust path, you lose clean control. | Move admin and incident coordination to trusted separate systems, block Exchange-side authentication dependencies where possible. |
| 2 | Identity-linked services using SMTP relay or mailbox access | Apps that use stored Exchange credentials or service accounts can become persistence points. | Rotate credentials, move relays behind controlled gateways, disable unused mailbox impersonation and broad service permissions. |
| 3 | Internet-facing applications that send mail directly through Exchange | A compromised app can become an internal bridge to mail infrastructure, or mail compromise can expose app credentials. | Replace direct Exchange submission with a restricted SMTP relay or mail security gateway with allow-listed sources. |
| 4 | Line-of-business systems that receive alerts, approvals or password flows via Exchange | These create operational dependency on a suspect system and can block containment decisions. | Move critical alerts and approvals to alternate channels first, then reduce mailbox dependency. |
| 5 | Scanners, MFDs, legacy devices and building or plant systems | They usually use weak auth, broad relay rights and poor logging. | Put them on a separate VLAN, force relay through a broker, retire direct server access. |
| 6 | Backup, monitoring and ticketing workflows that email into action | If response relies on email, attackers can suppress, observe or manipulate response. | Re-route critical notifications to portals, mobile push, SMS or non-Exchange collaboration channels. |
03First isolate trusted operations systems, not user mail
The fastest mistake is to spend two weeks hardening Outlook access while leaving the incident response path itself tied to Exchange. The September 2026 ISM change is useful here because it makes the principle explicit: use trusted systems separate from a compromised system.
So the first systems to isolate are not ordinary user mailboxes. They are the systems your security, infrastructure and service desk teams need when mail cannot be trusted. That includes privileged admin workstations, jump hosts, incident chat, runbooks, admin password vault access, case tracking and forensic storage. If password resets, approval chains, MFA fallbacks or admin notifications are still routed through Exchange-dependent paths, fix that before you touch lower-risk integrations.
This is as much an operating model issue as a network one. Your recovery team needs a clean collaboration path, separate credentials where possible, and logging that does not rely on the environment being investigated. If your current tooling sits on the same domain, the same mailbox notifications and the same network segment as Exchange, you have not really isolated anything.
This is also where a targeted cybersecurity risk assessment is more useful than a broad mail strategy workshop. The immediate task is identifying which operational controls still fail closed when Exchange is suspect.
- Separate incident communications from Exchange-connected systems
- Ensure jump hosts do not require mail-based approvals or notifications
- Review MFA fallback and password reset paths for Exchange dependency
- Keep investigation notes and evidence storage on trusted separate services
- Confirm the service desk can operate without Exchange-distributed tickets or alerts
04Then isolate application relay, service accounts and legacy device mail paths
The next priority is every application or device that talks to Exchange with a stored credential, wide relay permission or mailbox-level access. These are often ignored because they look low-risk: scanner-to-email, ERP alerts, website contact forms, HR workflow notifications, monitoring alarms and multifunction devices. In reality, they are common places to find stale credentials, over-broad permissions and almost no ownership.
Microsoft's own support pages make clear that organisations still on Exchange 2016 or 2019 after 14 October 2025 end of support are outside normal security support. If those environments are also carrying dozens of anonymous or weakly controlled SMTP dependencies, your blast radius is much larger than the server count suggests.
The practical move is not to rebuild every workflow. It is to insert a narrow control point. Put legacy senders behind a restricted relay tier or gateway. Allow-list source IPs. Require authenticated submission only where the device can support it. Remove direct mailbox access for applications that only need to send. Rotate all service credentials that have touched Exchange. Disable old receive connectors and SMTP auth exceptions that nobody can justify.
If you need to preserve these workflows during a staged exit, the clean pattern is to reduce Exchange to a controlled dependency behind an integration boundary. That is the same principle used in API integration work and legacy modernisation: introduce a choke point first, then migrate callers one by one.
- Inventory every system using SMTP relay, EWS, IMAP, POP or mailbox impersonation
- Classify each by business criticality and credential risk
- Move unaudited senders to a dedicated relay segment
- Revoke direct Exchange rights that exceed send-only needs
- Document owner, source IP, auth method and fallback plan for each sender
05Separate business-critical workflows that stop recovery when mail is down
Some systems are not dangerous because they can break into Exchange. They are dangerous because your business cannot make decisions without them once Exchange is isolated. Think approval chains, customer support notifications, maintenance alerts, payroll exception messages, ERP batch outcomes and security alarms that only arrive by email.
ASD's email hardening guidance focuses on reducing compromise and limiting impact. Limiting impact means moving the critical workflows that should continue during a mail incident onto alternative channels before you need them. This is often the highest-value 30-day programme because you can keep Exchange in place while making the business less dependent on it.
A useful triage rule is simple: if a workflow creates a financial, legal, safety or customer outage when email is unavailable for 24 hours, isolate it now. That may mean redirecting alerts to a service management portal, mobile push, SMS or a separate collaboration platform. It may mean changing an application so approvals can be actioned in-app instead of from mailbox links. It may mean stopping email-driven password resets for administrators entirely.
This work often sits between infrastructure and application teams, which is why it gets delayed. If the internal team is thin, a focused pod covering cloud infrastructure, application changes and mail routing can usually move faster than waiting for a full tenant migration programme to be approved.
- List workflows that fail if email is unavailable for one business day
- Mark which of those affect revenue, customer operations, payroll, safety or security response
- Provide a non-Exchange channel for each critical workflow
- Test those paths during a tabletop or controlled mail outage drill
06What to ask your team or vendor this week
If you need decisions quickly, ask questions that expose dependency and recovery risk, not just patch status.
First, ask for a list of systems that send mail, receive mail programmatically, read mailboxes, create Exchange connectors, use service accounts against Exchange APIs, or require Exchange to complete admin or security workflows. If no one can produce that list in a few days, the integration surface is already too opaque.
Second, ask which of those systems sit on the same network zones, identity paths or admin trust boundaries as Exchange. ASD's segmentation guidance is clear that organisations should separate sensitive systems from everyday activities such as web browsing, email and internet access. If the answer is 'most of them', ring-fencing has not started.
Third, ask which dependencies can be cut by configuration only, which need a relay or gateway inserted, and which need application change. This gives you a realistic 30, 60 and 90 day sequence. For many organisations, the fastest wins are connector clean-up, credential rotation, alternate alert routing and segmenting devices that should never have talked directly to Exchange in the first place.
- Which systems can still authenticate to Exchange with stored credentials?
- Which service accounts have mailbox read, impersonation or admin rights?
- Which apps and devices send directly to Exchange rather than through a controlled relay?
- Which operational processes fail if Exchange is quarantined today?
- Which trusted admin or incident systems still sit in the Exchange blast radius?
- What can be isolated by network policy only, and what needs code or workflow change?
07A practical 30-day isolation plan
Week 1: identify. Build a dependency register for mail-connected systems. Include protocol, auth method, owner, source network, privilege level and business criticality. Confirm whether you are still on a version that passed end of support on 14 October 2025, or whether a path to Exchange Server Subscription Edition is even realistic in your environment.
Week 2: protect trusted operations. Move incident coordination and privileged administration off Exchange-dependent paths. Rotate Exchange-linked service credentials. Remove mailbox rights that are not essential. Lock down receive connectors and SMTP auth exceptions.
Week 3: segment and broker. Place legacy devices and application senders into dedicated segments. Insert a restricted relay or gateway. Reduce direct server reachability. Apply allow-listing and logging at the relay boundary. ASD's network segmentation overview is useful here because it frames segmentation as a business resilience control, not just a network engineering task.
Week 4: test business continuity. Simulate Exchange quarantine for a controlled window. Check whether security alerts, customer communications, approvals and key operational notifications still reach the right teams through alternate paths. Then document the remaining dependencies that force a bigger migration or application change.
Sources
Topics
SyncTrix Editorial
SyncTrix · SyncTrix
Writes about the engineering decisions behind production systems - architecture, delivery and the trade-offs that only show up at scale.
Building something like this?
SyncTrix engineers AI, SaaS, platform and cloud systems for enterprises and high-growth teams. Tell us what you're shipping and we'll scope it with you.