A compromised Microsoft 365 account can give an attacker far more than access to one inbox. It may expose shared files, finance systems, customer information and internal conversations. To implement zero trust access is to reduce that opportunity by making every request for data, applications and systems earn its access rather than assuming it is safe.
For small and mid-sized organisations, zero trust is not a single product to buy or a switch to turn on. It is a practical way of improving how identities, devices and information are protected. Done well, it strengthens security without making ordinary work needlessly difficult.
What zero trust access means in practice
Traditional network security was built around a boundary: staff and systems inside the office network were generally trusted, while anything outside it received more scrutiny. That model is less useful when people work from home, access cloud platforms directly and use mobile devices across several locations.
Zero trust replaces assumed trust with ongoing checks. Before granting access, the organisation considers who is making the request, the device they are using, what they need to reach and whether the circumstances look normal. Access should be limited to what is needed for the task, and it should be reviewed as risk changes.
This does not mean treating every member of staff as a threat. It means recognising that passwords can be stolen, devices can be lost and accounts can be misused. Good controls protect people from the consequences of those events while allowing them to do their jobs.
Start with the business, not the security tool
The most common mistake is beginning with a technology purchase. A sensible zero trust programme starts by understanding the systems that matter most to the organisation and the people who use them.
Identify where sensitive or business-critical information sits. This may include Microsoft 365, finance and payroll applications, customer relationship management systems, production platforms, file shares and remote access tools. Then establish which roles genuinely need access, including suppliers and temporary staff.
The aim is not to document every permission perfectly on day one. Start with the accounts and systems where a compromise would cause the greatest disruption, financial loss or reporting obligation. For a manufacturer, that may be production and design data. For a school or academy trust, safeguarding and pupil records will demand particular attention. For a charity, donor data and finance platforms may be the priority.
This initial work often exposes a familiar issue: access that has accumulated over time. Former responsibilities, shared logins and broad administrator rights are convenient until they become a route into the organisation.
Define normal access before enforcing it
Before applying tighter policies, agree what normal working looks like. Consider approved devices, common locations, expected applications and the roles that require elevated permissions. These decisions should involve operational leaders as well as IT.
A policy that blocks an engineer working at a customer site, or a manager approving payroll while travelling, is not a success. Zero trust controls need to reflect how the organisation operates, while creating a clear process for exceptions that is monitored and time-limited.
Build identity security first
Identity is usually the most effective place to begin. If an attacker cannot successfully sign in, they cannot move easily into email, cloud files or connected business applications.
Multi-factor authentication should be standard for all users, particularly for email, remote access and administrator accounts. App-based authentication methods are generally stronger than text messages, although the right approach depends on user needs and the systems in place. Staff also need clear guidance on authentication fatigue attacks, where criminals repeatedly send approval prompts in the hope that someone accepts one.
Next, remove shared accounts wherever possible and make sure staff joiners, movers and leavers are handled consistently. A new starter should receive only the access their role requires. When someone changes jobs, permissions should change with them. When they leave, access must be removed promptly, including access provided through third-party software.
Privileged access deserves separate treatment. Administrator accounts should not be used for ordinary email, browsing or document work. Give elevated access only when it is needed, require stronger authentication and keep a record of who has it. For some organisations, time-limited administrator access is a worthwhile next step once the basics are established.
Make device health part of the decision
A valid password and authentication code are not enough if the device itself is unmanaged or compromised. Zero trust access should take account of whether a laptop, desktop or mobile device meets your security requirements.
At a minimum, managed devices should use supported operating systems, disk encryption, endpoint protection and automatic security updates. They should also be protected by a screen lock and backed up where appropriate. Mobile device management can help apply these standards consistently, particularly where staff use company phones or tablets.
Conditional access policies can then allow or restrict access based on device compliance and sign-in risk. For example, a managed, encrypted company laptop may be allowed to access SharePoint and Teams normally, while an unmanaged personal device can be limited to browser access or blocked from downloading sensitive files.
There is a trade-off here. Some organisations rely on bring-your-own-device arrangements because they are practical or cost-effective. That does not rule out zero trust, but it changes the policy design. The controls may need to focus more heavily on browser sessions, application protection and preventing local downloads, rather than full device management.
Apply least privilege to data and applications
Once identity and devices are better controlled, review what users can reach after signing in. Least privilege means people receive the minimum access necessary to complete their work, not every permission they might conceivably need.
In practice, this means using role-based groups rather than granting permissions individually, separating confidential folders from general collaboration spaces and reviewing external sharing. It also means checking that old project sites, Teams channels and shared mailboxes do not remain open indefinitely.
Not every file needs the same protection. A public brochure is different from a board paper, an HR record or a technical drawing. Classifying information by sensitivity helps the organisation apply proportionate controls, such as restricted sharing, encryption or additional authentication for highly sensitive information.
Avoid attempting to redesign every permission at once. Begin with the most sensitive data sets and the applications that create the greatest operational risk. Early improvements are easier to manage, and they give staff time to adapt to new ways of working.
Monitor, test and improve the policies
Zero trust access is not a one-off implementation. People change roles, new software is introduced and threats evolve. Policies need regular review, supported by meaningful monitoring.
Look for unusual sign-ins, repeated authentication failures, impossible travel alerts, unexpected administrator activity and large downloads of sensitive data. These signals do not always indicate an attack, but they provide the context needed to investigate quickly.
Testing matters too. Run access reviews for key systems, test account removal when staff leave and rehearse what happens if an administrator account is compromised. A documented incident response plan is useful only if the people involved know their responsibilities and can act without delay.
User feedback should shape the process. If a control creates repeated friction for a legitimate team, investigate the cause rather than encouraging workarounds. The objective is secure, reliable access that supports productive work, not security for its own sake.
A realistic route to implement zero trust access
For many organisations, the right route is phased. Begin by securing identities with multi-factor authentication, stronger administrator controls and sound joiner, mover and leaver processes. Then bring managed devices up to a defined standard and use conditional access to protect key cloud services.
The next phase is usually access rationalisation: removing unnecessary permissions, improving data sharing rules and introducing regular reviews. More advanced measures, such as data classification, privileged access management and detailed risk-based policies, can follow when the foundations are working reliably.
This approach is more sustainable than a large, disruptive programme. It also makes costs clearer, because investment can be directed towards the risks that matter most rather than towards features the organisation is not ready to use.
A trusted technology partner can help translate these decisions into workable policies, particularly where internal IT teams are already managing day-to-day support. The value lies in balancing security requirements with the realities of staff, systems and budget.
Zero trust should leave your organisation more confident, not more constrained. Start with the access that would cause the most harm if it were misused, make the next improvement achievable, and build from there.

