A compromised identity is often all an attacker needs. It can provide a legitimate route into Microsoft 365, a VPN, a cloud console or a business-critical application, bypassing many controls designed to stop malware at the perimeter. For security leaders asking how to reduce identity risk, the immediate priority is not simply adding another authentication prompt. It is understanding which identities matter most, how they obtain access, and where that access exceeds the needs of the role.
Identity risk has expanded well beyond permanent employee accounts. It now includes administrators, contractors, suppliers, service accounts, application identities and former staff whose access was never fully removed. In a hybrid estate, identity controls have become a practical foundation for ransomware resilience, data protection and regulatory confidence.
Start with an identity risk assessment
Most organisations have an identity source of record, but fewer have a reliable view of every active account and its effective permissions. Before selecting technology, establish a baseline that combines directories, identity providers, cloud platforms, SaaS applications and privileged access systems.
The assessment should identify dormant accounts, shared credentials, excessive administrator rights, accounts without multi-factor authentication (MFA), stale guest users and service accounts with unclear ownership. It should also show where identity data is fragmented. A user may be disabled in HR and Active Directory while retaining an active licence and access to sensitive files in a SaaS platform.
Risk should be prioritised by impact, not just by the number of findings. A dormant marketing account is undesirable; a forgotten global administrator, domain administrator or finance-system approver is materially more urgent. The same is true for application identities that can read large volumes of confidential data or modify security settings.
This exercise also exposes an operational question that is easily missed: who owns the remediation? Security can identify risk, but HR, IT operations, application owners and procurement teams may each control part of the access lifecycle. Assigning accountable owners makes findings actionable rather than another unaddressed report.
How to reduce identity risk through stronger access controls
MFA remains one of the most effective controls against password-based attacks, but deployment quality matters. Apply it consistently to remote access, cloud administration, email, privileged accounts and high-risk business applications. Excluding older authentication methods or creating broad policy exceptions can leave a route around an otherwise sound control.
Where practical, move beyond basic push notifications. Phishing-resistant methods, such as FIDO2 security keys or certificate-based authentication, reduce exposure to adversary-in-the-middle phishing and MFA fatigue attacks. For some organisations, particularly those with frontline staff or legacy applications, a staged approach may be necessary. The right target architecture depends on user population, application compatibility and support capacity.
Conditional access should add context to authentication decisions. A sign-in from a managed, compliant device in the UK during normal working hours presents a different risk from an unfamiliar device accessing a sensitive application from an unusual location. Context-aware policies can require additional verification, restrict downloads or block access altogether when risk indicators are high.
However, conditional access is not a substitute for clean identity governance. Policies become difficult to audit when they contain years of emergency exclusions, duplicate groups and undocumented exceptions. Establish a change process, record the business justification for every exception and review it on a defined schedule.
Apply least privilege to people and systems
Least privilege means users, administrators and applications receive only the access needed to perform their current task. It is straightforward in principle and often difficult in established environments, where permissions accumulate through role changes, project work and temporary requests that were never withdrawn.
Role-based access control provides a useful starting point, but roles must reflect actual working practices. A generic group such as “all finance” may grant far more access than many team members require. For higher-risk systems, use narrower roles and time-limited access. Privileged access management (PAM) can provide just-in-time elevation, approval workflows, session recording and password rotation for sensitive accounts.
Service accounts need equivalent scrutiny. They commonly hold powerful, long-lived credentials and are rarely protected by the same lifecycle processes as user accounts. Each service account should have a named owner, a documented purpose, only the permissions required, and a defined credential or certificate rotation process. Where supported, managed identities and workload identity federation can reduce dependence on static secrets.
Make joiners, movers and leavers a security process
Identity risk often begins in ordinary business processes. A rushed new-starter request can create excessive access. A department move can leave permissions from the previous role in place. A poorly managed departure can leave active sessions, mobile access or SaaS accounts available after employment ends.
Automating joiner, mover and leaver workflows reduces both delay and inconsistency. HR-triggered processes can provision a baseline role on arrival, prompt access review when someone changes position, and remove access promptly on departure. Automation should still include controls for non-standard arrangements, such as contractors, temporary workers, mergers and shared service teams.
Access reviews are essential where automation cannot determine whether access remains justified. Application owners should certify privileged roles, access to sensitive data repositories and external guest access at a frequency proportionate to risk. Quarterly review may suit administrator permissions; lower-risk access may be reviewed less frequently. Asking managers to approve a long, unprioritised entitlement list is unlikely to produce meaningful governance. Present high-risk access clearly, with role context and a simple action path.
Extend identity governance to SaaS and third parties
SaaS adoption creates a common blind spot. Teams can purchase and deploy applications outside central IT, then connect them to corporate identity systems or grant them permissions to read mailboxes, calendars and files. These integrations may continue long after the original business need has ended.
Maintain an inventory of sanctioned SaaS applications, connected OAuth applications and their consented permissions. Review whether integrations need broad tenant-wide access, whether they are actively used and whether the supplier's controls are suitable for the data involved. Restrict user consent where appropriate, and establish a process for reviewing new requests before access is granted.
Third-party access requires the same discipline. Supplier accounts should be individually assigned, time-bound and monitored. Shared supplier logins make accountability difficult and should be removed. For critical suppliers, agree access methods, logging expectations, incident notification requirements and offboarding responsibilities contractually, not only through technical configuration.
Detect identity attacks before access becomes impact
Preventive controls will not stop every attack. Security teams need visibility into suspicious authentication and privilege activity, including impossible travel, repeated MFA prompts, unusual token use, new forwarding rules, privilege escalation, mass file access and changes to conditional access policies.
Identity telemetry should be correlated with endpoint, email, network and cloud activity. An apparently valid sign-in may become clearly suspicious when the same account launches remote tools, accesses unfamiliar data stores or creates persistence on an endpoint. This is where EDR, XDR and identity security capabilities can complement one another, provided the organisation has the people and processes to investigate the resulting alerts.
Detection quality depends on tuning and response design. Set clear playbooks for events such as suspected account takeover or privileged-role changes: revoke sessions, reset credentials where appropriate, disable accounts, contain affected devices, preserve evidence and assess what data or systems were accessed. Testing these actions matters. A response plan that requires five teams to agree during an active compromise will be slower than the threat.
Choose identity security tools against your operating model
The market includes identity providers, MFA and conditional access platforms, PAM tools, identity governance and administration products, SaaS security tools, identity threat detection and response platforms, and broader XDR solutions. There is overlap, but they do not solve the same problem.
A useful evaluation begins with the primary gap. An organisation struggling to remove leaver access may need lifecycle automation and governance. One with too many standing administrator rights may prioritise PAM. Another facing phishing-led account takeover may benefit more from phishing-resistant authentication, conditional access and identity-focused detection. Buying a platform with an impressive feature list will not compensate for unclear ownership, incomplete integrations or no capacity to operate it.
Assess products against directory and application compatibility, coverage for cloud and on-premises identities, reporting needs, API maturity, deployment effort, support model and commercial predictability. Also ask how well the product supports evidence for audits and internal governance. Centralised reporting on access decisions, review completion and privileged activity can reduce the effort of demonstrating control effectiveness.
For organisations comparing a fragmented set of products, independent assessment is valuable. A vendor-neutral partner can map requirements to the relevant technology categories, identify overlap with existing controls and support a procurement decision that remains defensible after deployment.
The practical goal is not zero identity risk, which is unrealistic in a changing workforce and technology estate. It is an environment where access is deliberate, privilege is constrained, suspicious behaviour is visible and teams can remove access quickly when circumstances change. That is a far stronger position from which to manage both everyday operational risk and a serious cyber incident.

