Back to Insights
SECURITY ARCHITECTURE

Building a Practical Zero Trust Roadmap

Key steps to assess, design, and implement Zero Trust in phases that align with your business and risk priorities.

7 min read

At a glance

Prioritize

Start with critical resources, business impact and real access paths.

Decide

Evaluate identity, device, resource, action and current context.

Enforce

Place controls close to applications, data, services and workloads.

Assure

Measure policy coverage, exceptions, containment and risk reduction.

A practical roadmap for leaders who need stronger resource protection without turning Zero Trust into a technology purchasing exercise. The document connects critical services, identity, devices, workloads, data, policy enforcement, visibility and response into one staged operating model.

Primary audience

  • Executive sponsors
  • Security leaders
  • Enterprise architects
  • Identity teams
  • Platform owners
  • Application leaders
  • Risk functions

Leadership objective

Select a small number of critical resources, approve the target access model, implement in phases and measure operating outcomes before wider expansion.

01Executive perspective

Zero Trust protects resources, not trusted locations

Treat Zero Trust as a resource protection and operating model change.

Zero Trust removes implicit trust based only on network location, asset ownership or prior access. It evaluates the subject, device, resource, action and current context before permitting access and while the session remains active.

Many organizations begin Zero Trust by purchasing a new access product or by renaming existing network security initiatives. That approach misses the operating change. The real objective is to ensure that every important resource has an explicit access policy, a reliable enforcement point, current decision inputs and a tested response when risk changes.

Leadership test

Can the organization explain who or what can access each critical resource, from which device, for which action and under which policy?

The strongest starting point is not the entire enterprise. It is a short list of critical services, applications, data sets and administrative paths where misuse would cause material business harm. Each selected use case should have a named owner, a clear access purpose and measurable risk outcomes.

Avoid this trap

Do not define progress only by product deployment. A capability is useful when policy is enforced, exceptions are controlled and evidence is available.

Zero Trust at a glance

  • Verify explicitly: evaluate identity, device, resource, action, risk and purpose for every important access request.
  • Use least privilege: grant only the resource, action and time required for the approved business purpose.
  • Assume compromise: limit exposure, prevent lateral movement and preserve tested containment options.
02Architecture principles

Translate seven durable Zero Trust tenets into design decisions

Translate the seven NIST tenets into practical design decisions.

NIST describes seven tenets that guide a Zero Trust Architecture. The implementation can vary, but each protected resource should be governed through explicit policy, current context and enforceable access decisions.

  1. Resources are protected: data, services, applications, devices and workflows are treated as resources.
  2. Communication is secured: traffic is protected regardless of network location or ownership.
  3. Access is per session: authorization is granted for a specific resource and session.
  4. Policy is dynamic: decisions reflect identity, device, resource, behavior and threat signals.
  5. Asset state is monitored: the organization measures device, workload and service integrity.
  6. Authentication is enforced: subject and device assurance are evaluated before access and renewed as needed.
  7. Telemetry improves policy: activity and posture data refine decisions, investigation and response.

Architecture rule

Every protected resource needs an owner, policy, enforcement point, decision inputs, logging and response path.

Policy scope

  • Who or what
  • Which device or workload
  • Which resource
  • Which action
  • Which purpose
  • Which conditions

Design implication: a resource is not protected because it sits on a private network. It is protected when access is explicitly evaluated, enforced, observed and terminated when risk becomes unacceptable.

Traditional assumptionZero Trust response
Inside the network is trustedLocation becomes one signal and never the sole basis for access.
A user login proves legitimacyIdentity, device, resource, action and context are evaluated together.
Network access grants application accessPolicy enforcement is placed close to the protected resource.
Access remains valid until logoutSessions can be reevaluated, constrained or terminated as risk changes.
03Assessment and prioritization

Start with critical services, access paths and control gaps

Select critical services and access paths before selecting technology.

A practical programme begins with discovery. Identify what must be protected, who and what accesses it, how access is currently granted and which failures create unacceptable business impact.

Business criticality

Service purpose, customers, revenue, safety, regulatory exposure, recovery need and accountable owner.

Data sensitivity

Classification, residency, legal obligations, sharing constraints, retention and loss consequence.

Access patterns

Employees, administrators, partners, customers, service identities, devices and actions.

Architecture and exposure

Applications, APIs, networks, cloud services, dependencies, public paths and legacy constraints.

Control readiness

Identity, device posture, enforcement points, telemetry, response, ownership and evidence.

Change dependency

Application, identity, network and process changes required to implement the target policy.

Discovery outputs

  • Critical resource register
  • Access path map
  • Identity and privilege inventory
  • Device and workload inventory
  • Policy and telemetry gap list
  • Prioritized use case backlog

Selection rule

Choose use cases where business value, risk reduction, technical feasibility and accountable ownership are all present.

First use cases: privileged administration, remote workforce access, partner access, sensitive data access and workload communication are common starting points.

Decision factorQuestions to resolvePriority signal
Business consequenceWhat happens if access is abused, disrupted or granted to the wrong party?High financial, regulatory, safety or customer impact.
Exposure and reachHow many users, devices, partners, networks and interfaces can reach the resource?Broad or poorly understood access path.
Identity and privilegeAre privileged roles, service identities and lifecycle events controlled?Standing privilege, shared credentials or weak ownership.
Control feasibilityCan policy be enforced and can reliable context be collected?A viable path to measurable improvement.
04Target architecture

Place policy decision and enforcement around protected resources

Connect policy decision, policy administration and policy enforcement.

A Zero Trust Architecture separates policy decisions from policy enforcement. Identity, device, resource and environmental signals inform the decision, while an enforcement point establishes, constrains, records or terminates the approved connection.

Vendor neutral Zero Trust reference architecture

  • Governance, compliance and risk appetite drive policy optimization and security posture assessment.
  • Request sources: human identities, partners, service identities, corporate, personal and server endpoints.
  • Strong authentication, device compliance and workload posture inform each request.
  • Risk and behavior assessment feeds the policy decision point.
  • Policy engine evaluates policy, context and risk; a policy administrator creates or revokes access.
  • Threat protection provides continuous assessment, threat intelligence, forensics and response.
  • Enforcement permits, denies, limits, records and terminates access across public, private, brokered and segmented paths.
  • Protected resources include data, applications and infrastructure across cloud, containers and internal services.
  • Telemetry, security analytics, policy evidence, threat intelligence, investigation and response close the loop.

Architecture rule

The subject does not connect directly to the protected resource. Access is mediated by an enforcement point using an explicit decision.

  • Decision inputs: use authoritative identity, device, workload, resource, behavior, threat and session signals.
  • Operating evidence: preserve the request, policy result, enforcement action, resource activity and response record.
05Identity and privilege

Make identity the primary control plane for human and service access

Make human and service identity the primary control plane.

Identity is more than authentication. A practical Zero Trust programme governs who can receive access, how credentials are issued, how privilege is elevated and how access is removed when employment, role or risk changes.

Identity governance

Authoritative source, joiner, mover and leaver workflow, ownership, review and separation of duties.

Strong authentication

Phishing resistant methods for high risk access, risk signals and secure recovery processes.

Privileged access

Dedicated administration identities, temporary elevation, session control, approval and audit.

Federation and partners

Trusted identity providers, contractual responsibilities, claims, lifecycle and access boundaries.

Service identities

Unique workload identity, managed credentials, short validity, rotation and nonhuman ownership.

Session policy

Reauthentication, step up, time limits, activity control and termination when risk changes.

High value controls

  • Phishing resistant authentication
  • Dedicated privileged identities
  • Temporary elevation
  • Service identity ownership
  • Rapid session revocation

Policy principle

Grant the minimum resource, action and duration required for an approved purpose.

Risk signal

Shared accounts, long lived secrets and permanent administration roles weaken both prevention and accountability.

SignalPolicy useEvidence
Identity assuranceConfirm the subject, authentication strength and credential risk.Authentication record, identity source and recovery history.
Role and entitlementLimit the resource and action to approved business responsibility.Role design, access approval and review record.
Privilege stateRequire temporary elevation, additional approval and enhanced monitoring.Elevation request, session record and command trace.
Lifecycle statePrevent access after termination, supplier expiry, role change or compromise.Source event, removal record and exception closure.
06Device, application and workload trust

Use differentiated access paths for managed, privileged and unmanaged devices

Evaluate health, integrity and identity for every access participant.

Zero Trust does not require every device to receive the same access experience. The access path should reflect device ownership, posture, user privilege, resource sensitivity and the organization's ability to enforce and monitor policy.

Validated resource access: device ownership and posture determine the permitted route, assurance level and accessible resource scope. User access devices range from managed devices with validated posture and direct access to approved corporate resources, to privileged access workstations that are hardened and dedicated, to personal or unmanaged devices that receive restricted access through a controlled path. Identity assurance covers authentication, role, privilege and risk; device and workload trust covers compliance and integrity, both evaluated at a policy enforcement point before reaching protected resources such as business applications, sensitive data and workflows, and cloud and infrastructure spanning on premises and multi cloud environments.

Three access paths

  • Managed device path: allow direct access only when the device is known, managed, healthy and appropriate for the requested action.
  • Privileged path: use a dedicated identity and workstation, temporary elevation, restricted destinations and complete session evidence.
  • Unmanaged path: use browser controls, virtual desktops, isolation or reduced resource scope rather than broad corporate access.
07Network and resource access

Constrain paths and lateral movement without treating location as trust

Constrain paths and lateral movement without treating location as trust.

Networks remain important for exposure control, segmentation and containment, but a private address or internal route should not grant broad access. Policy should be enforced close to the protected resource.

Identity aware access

Broker access to applications according to user, device and resource policy.

Macro segmentation

Separate business zones, environments, management planes and high consequence resources.

Micro segmentation

Limit workload communication to approved service identities, ports, protocols and flows.

Application and API gateways

Authenticate, authorize, validate and record access at application interfaces.

Service mesh policy

Apply service identity, mutual authentication and granular policy between workloads.

Egress control

Restrict outbound destinations, protocols, data movement and unauthorized command paths.

Network role

Reduce reachability, isolate high consequence resources and provide rapid containment when policy or identity fails.

Do not rely on

  • Private addressing
  • Corporate network location
  • One time VPN authentication
  • Broad subnet access

Design target

Users and services should reach only the specific resource and action required for the approved purpose.

Access pathMinimum design expectationEvidence
Employee to applicationFederated identity, device context, resource policy, secure session and activity logging.Policy result, authentication, posture and application record.
Administrator to platformDedicated identity, approved elevation, hardened path, restricted commands and session audit.Approval, elevation, command trace and review.
Partner to serviceFederation or managed identity, defined claims, limited resource scope and expiry.Agreement, identity mapping, policy and recertification.
Workload to workloadUnique service identity, explicit authorization, protected channel and flow telemetry.Identity record, policy, token and service logs.
08Data protection

Make data sensitivity, purpose and consequence part of every access decision

Make sensitivity, purpose and consequence part of access policy.

The purpose of Zero Trust is to protect resources, and data is often the resource of greatest consequence. Classification, entitlement, purpose, handling and output controls should shape access policy.

Inventory and ownership

Locate sensitive data, name the owner and record systems, interfaces, copies and processors.

Classification and labeling

Apply practical labels linked to access, sharing, retention and protection requirements.

Access and purpose

Authorize who can view, change, export, administer or share data and for which purpose.

Protection and handling

Encryption, key control, masking, rights management, loss prevention and secure deletion.

Activity and response

Monitor access, export, sharing, deletion and unusual behavior with defined actions.

Output consequence

Apply approval, validation and human review where an action can create material harm.

Data policy test

Can the organization explain why access was allowed, what action was taken and whether data left the approved boundary?

High consequence actions

  • Bulk export
  • External sharing
  • Privilege change
  • Irreversible deletion
  • Model or agent execution

Owner decision

Data owners should define acceptable use, required assurance and consequences that require approval or dual control.

Decision inputQuestionPolicy response
Data sensitivityWhat business, privacy, regulatory or customer consequence applies?Require stronger assurance, tighter scope and greater monitoring.
Subject and purposeIs the requester entitled and is the intended use approved?Permit only the role, action and purpose needed.
Device and applicationCan the access path enforce handling requirements?Restrict download, copy, print, export or offline use.
Location and transferDoes the destination comply with residency, supplier and sharing requirements?Deny, require approval or apply protected transfer.
Behavior and volumeIs the activity consistent with normal use and business need?Step up, limit, alert, suspend or investigate.
09Visibility and response

Correlate signals so access can be evaluated and misuse contained

Correlate signals so policy can adapt and misuse can be contained.

Zero Trust depends on current information. The organization should collect signals that affect policy, investigation and containment, then connect them across identity, devices, workloads, applications, networks and data.

Identity

Authentication, entitlement, privilege, federation, token and lifecycle events.

Device

Management, posture, configuration, vulnerability, malware and compromise signals.

Application

Authentication, authorization, administration, API activity and business actions.

Workload

Service identity, process, image, runtime, dependency, secret and communication activity.

Network

Reachability, DNS, flow, proxy, gateway, segmentation and unusual egress events.

Data

Access, search, bulk use, export, sharing, deletion and policy violations.

Policy and enforcement

Permit, deny, step up, exception, policy change and failed enforcement events.

Threat and response

Threat intelligence, detections, investigations, containment and recovery evidence.

Risk conditionRapid actionHuman decision
Credential compromiseRevoke sessions, disable tokens, require new authentication and protect related accounts.Confirm scope, reset access and approve restoration.
Untrusted deviceRestrict resources, isolate the endpoint and preserve posture evidence.Assess compromise and authorize return to service.
Policy bypassBlock the path, raise severity and preserve relevant configuration and logs.Approve emergency control and determine residual risk.
Unusual data movementLimit export, suspend sharing and preserve data activity evidence.Validate business purpose, privacy impact and reporting need.

Policy feedback loop

Telemetry should improve policy. Repeated exceptions, false positives, blocked business actions and missed detections should drive controlled changes.

Containment options

  • Revoke identity sessions
  • Remove privilege
  • Isolate device
  • Block workload path
  • Restrict data export

Evidence goal

Investigators should reconstruct who accessed which resource, from what context, under which policy and with what outcome.

10Implementation roadmap

Move through four risk aligned delivery phases

Move through four risk aligned delivery phases.

Zero Trust is implemented through selected use cases and reusable capabilities, not a single enterprise cutover. The phases below should be adjusted to service criticality, dependencies, change capacity and evidence of operating effectiveness.

Days 0 to 60: mobilize and prioritizeMonths 2 to 4: strengthen identity and device trustMonths 4 to 8: enforce resource accessMonths 8 to 12: scale, automate and assure
Name the sponsor and architecture owner. Inventory critical resources and access paths. Select initial use cases and success measures. Close immediate privileged access gaps. Gate: approved use case charter and baseline.Improve lifecycle and authentication. Reduce standing privilege. Establish device inventory and posture. Define policy ownership and exception handling. Gate: reliable identity and posture inputs.Place enforcement around priority resources. Limit network and application paths. Protect workload communication and sensitive data. Integrate session evidence and containment. Gate: policy operates consistently for selected use cases.Expand reusable policy patterns. Correlate telemetry across control domains. Test containment and automate approved actions. Measure risk outcomes before wider rollout. Gate: operating evidence supports controlled expansion.

Roadmap principle

A phase is complete when the selected access path operates with approved policy, reliable inputs, enforceable action, current evidence and tested response. Product installation alone is not a delivery gate.

100%0<15 minutesQuarterly
Selected critical resources with named owners and approved access policiesShared privileged accounts in the first implementation scopeTarget time to revoke a compromised high risk sessionLeadership review of policy coverage, exceptions and risk outcomes
11Practical checklist

Zero Trust readiness checklist

Confirm minimum readiness and operating evidence.

Use this checklist to confirm that minimum architecture, policy, enforcement, telemetry and governance components are in place. A checked item should be supported by current evidence and an accountable owner.

Readiness areaMinimum evidence
Business priorityCritical services, data and risk outcomes are approved and sequenced.
Resource inventoryApplications, APIs, data, devices, workloads, dependencies and owners are recorded.
Access path mapHuman, partner, privileged and service access paths are understood.
Policy authorityPolicy ownership, approval, exceptions and emergency changes are defined.
Identity lifecycleJoiner, mover, leaver, federation, privilege and service identity controls operate.
Device and workload trustPosture, integrity, identity and compromise signals influence access.
Policy enforcementSelected resources have a permit, deny, constrain and terminate path.
Network containmentExposure and lateral movement are reduced through segmentation and controlled routes.
Data policySensitivity, purpose, handling, export, sharing and deletion influence decisions.
Telemetry correlationIdentity, device, application, workload, network, data and policy events can be investigated together.
Response and recoveryCredential revocation, device isolation, workload containment and data protection actions are tested.
Metrics and assuranceManagement reviews policy coverage, denied access, privilege, exceptions, containment and corrective action.

Priorities

  • First priority: establish critical resource ownership, access path visibility and policy authority.
  • Second priority: strengthen identity, device trust, enforcement and telemetry for selected use cases.

Evidence rule

Do not mark an item complete without a current record, an accountable owner and evidence that the control operates.

12References and advisory

Reference sources and publication scope

Source guidance and publication scope.

The references below provide the basis for the architecture, maturity and implementation concepts used in this roadmap. They should be interpreted for the organization's sector, jurisdiction, technology estate and risk tolerance.

  1. NIST SP 800-207: Zero Trust Architecture and the foundational resource focused tenets.
  2. NIST SP 1800-35: Final practice guide with nineteen example Zero Trust implementations and lessons learned.
  3. NIST SP 800-207A: Application and service identity policy for cloud native and multi cloud environments.
  4. CISA Zero Trust Maturity Model, Version 2: Five pillars and three cross cutting capabilities for planning maturity.
  5. NSA Zero Trust Implementation Guidelines Primer: Methodology and linkage to NIST, CISA and defense Zero Trust guidance.
  6. NSA Zero Trust Implementation Guidelines Discovery Phase: Foundational discovery of data, applications, assets, services and access activity.
  7. NSA Zero Trust Implementation Guidelines Phase One: Activities that establish a secure foundation for Target level capabilities.
  8. NSA Zero Trust Implementation Guidelines Phase Two: Activities that integrate core Zero Trust solutions within the enterprise environment.
  9. Department of Defense Zero Trust Strategy and Roadmap: Strategic outcomes, capabilities and implementation direction for defense environments.
  10. NIST Cybersecurity Framework 2.0: Enterprise cybersecurity outcomes for governance, protection, detection, response and recovery.
  11. NIST SP 800-53 Revision 5: Security and privacy control families relevant to Zero Trust implementation.
  12. CISA Modern Approaches to Network Access Security: Guidance on Zero Trust, Secure Service Edge and improved remote access visibility.

How to use this brief

Use it as a vendor neutral architecture and delivery baseline, then tailor each use case to business impact, technology constraints and operating evidence.

June 2026 update

The reference set includes the final NIST implementation guide and the NSA Zero Trust Implementation Guidelines released in 2026.

About this advisory: this publication provides general, vendor neutral information and does not constitute legal, regulatory or product specific advice. Architecture and controls should be tailored to the organization's resources, access patterns, threat exposure and risk appetite. PFGsec supports Zero Trust strategy, architecture, identity, privilege, segmentation, cloud and workload access, data protection, visibility and phased implementation. Contact: insights@pfgsec.com, +1 (613) 402 6271.