Layered security, not a single product
No single product makes an environment secure. A defensible program layers identity controls, endpoint protection, email security, data governance, monitoring and a tested incident response plan. The tools (Defender, Purview, Entra ID, Intune) are the implementation; the architecture is Zero Trust — never trust, always verify, assume breach.
Microsoft Defender threat protection
The Defender suite spans endpoints (Defender for Endpoint), identity (Defender for Identity), Office 365 (Defender for Office 365) and Cloud Apps (Defender for Cloud Apps). Apply the preset Standard and Strict policies in Defender for Office 365 before hand-tuning. Enable attack surface reduction rules in Defender for Endpoint, and turn on automated investigation and remediation for high-confidence threats.
- Apply preset Standard/Strict policies before custom rules
- Enable Safe Links for internal mail, not just inbound
- Turn on ASR rules in block mode after auditing
- Enable automated investigation for high-confidence detections
Purview compliance and eDiscovery
Microsoft Purview is the compliance operating layer. Use retention labels for content lifecycle, sensitivity labels to mark and protect confidential data, and DLP policies to stop risky sharing. For investigations, use Purview eDiscovery (Standard for routine, Premium for large or custodian-heavy cases). Enable unified audit logging on day one — you cannot retroactively audit what was never logged.
| Purview capability | Purpose | Start with |
|---|---|---|
| Retention labels | Content lifecycle | One label per business value |
| Sensitivity labels | Mark + protect confidential data | Public/Internal/Confidential |
| DLP | Stop risky sharing | One high-value policy (e.g. credit cards) |
| eDiscovery | Legal/HR investigations | Standard; Premium for large cases |
| Audit | Investigation baseline | Enable on day one |
MFA and Conditional Access policies
Conditional Access is where identity security actually lives. Build a baseline that requires MFA for all admins, blocks legacy authentication, and requires compliant devices for sensitive apps. Then layer report-only policies to measure impact before enforcing. Use authentication strength to require phishing-resistant MFA (FIDO2, Windows Hello) for admins and passwordless for users where possible.
Connect-MgGraph -Scopes "Policy.ReadWrite.ConditionalAccess"
$params = @{
displayName = "Require MFA for admins"
state = "enabled"
conditions = @{ roles = @("62e90394-69f5-4237-9190-012177145f10") }
grantControls = @{ operator = "OR"; builtInControls = @("mfa") }
}
New-MgIdentityConditionalAccessPolicy -BodyParameter $paramsZero Trust security architecture
Zero Trust is not a product — it is a set of principles: verify explicitly, use least-privilege access, and assume breach. In practice that means identity as the primary perimeter (Entra ID + Conditional Access), device compliance gating access (Intune), micro-segmentation of networks, and continuous monitoring. Every access decision is evaluated at request time, not granted once at the network edge.
- Verify explicitly — identity, device, risk at every request
- Least-privilege access — JIT/JEA for admin roles
- Assume breach — segment, monitor, contain
- Device compliance gates access via Intune + Conditional Access
Compliance frameworks: HIPAA, GDPR, SOC 2
Map controls to the framework, then implement once and evidence for many. HIPAA governs healthcare PHI (encryption, access controls, audit, BAAs with vendors). GDPR governs EU personal data (lawful basis, data subject rights, breach notification in 72 hours, DPIAs). SOC 2 governs service organisation security (the Trust Services Criteria — security, availability, processing integrity, confidentiality, privacy). Use Microsoft Purview Compliance Manager to track control implementation across frameworks.
| Framework | Scope | Key obligation |
|---|---|---|
| HIPAA | US healthcare PHI | Encryption, BAAs, audit, breach notification |
| GDPR | EU personal data | Lawful basis, DSAR, 72h breach notice, DPIA |
| SOC 2 | Service organisations | Trust Services Criteria + auditor attestation |
| ISO 27001 | Infosec management system | Risk-based ISMS + certification |
Security incident response
An incident response plan that has never been tested is a hope, not a plan. Define roles (incident lead, comms, forensics), severity levels and escalation paths. Run tabletop exercises quarterly and a full technical drill annually. Integrate alerts from Defender, Sentinel and the SIEM into a single queue with clear ownership and SLAs. Post-incident, run a blameless retrospective and update the runbook.
- Define IR roles, severity levels and escalation paths
- Tabletop quarterly, full technical drill annually
- Single alert queue with ownership and SLAs
- Blameless retrospective after every incident — update the runbook
Lessons learned
- Layer controls — no single product makes you secure
- Enable unified audit logging on day one
- Require phishing-resistant MFA for admins
- Test the incident response plan — an untested plan is a hope
Written by
Jamshed Alam
Global MSP & IT Infrastructure Consultant with 20 years across Exchange, Microsoft 365, Azure, security and automation. Available for project and contract work worldwide.