Single Sign-On and Identity: Keeping System Access Under Control
A sales manager leaves on Thursday and can still log in to the CRM five weeks later. Single sign-on and a clear identity process fix that, if you set them up in the right order.

A sales manager resigns on Thursday. HR processes the paperwork, IT disables the email account, and everyone moves on. Five weeks later someone notices that the same person still has an active login to the CRM (customer relationship management system), the shared file platform and the pricing tool, because each system had its own username and nobody had a list.
This is the everyday access problem in growing companies. Each new tool adds another set of accounts, and the effort of keeping them all correct falls on whoever remembers. Single sign-on, combined with a clear identity process, is how you get that under control.
What single sign-on is
Single sign-on (SSO) means users sign in once to a central identity provider and that identity is trusted by the other systems. Instead of each application keeping its own password, the application hands the login to the identity provider and accepts its answer. Common identity providers include Microsoft Entra ID (formerly Azure AD), Google Workspace and dedicated services such as Okta.
The connection between the identity provider and each application usually uses one of two standards: SAML (Security Assertion Markup Language), which is older and widely supported by business software, or OpenID Connect, which is newer and common in modern cloud applications. You do not need to understand the protocol details, but you should ask every vendor which of the two they support and whether SSO is included in your licence tier or sold as an extra.
The benefits are practical. Users have one strong password plus multi-factor authentication instead of eight weak ones. IT can switch off one account and cut access to everything connected. And you get one place that shows who signed in to what and when.
SSO is half the answer: provisioning is the other half
SSO controls how people log in. It does not, by itself, control which accounts exist inside each application. If someone’s identity is disabled, they cannot sign in through SSO, but their account in the CRM may still exist, still own records and still count against your licences. Some applications also allow local logins that bypass SSO unless you switch that off.
Provisioning closes the gap. With automated provisioning, often using a standard called SCIM (System for Cross-domain Identity Management), the identity provider creates, updates and disables accounts in connected applications based on the user’s record. When HR marks someone as a leaver, the change flows to every connected system. Not all applications support it, and some only support it on higher licence tiers, so check before you assume.
For applications that support neither SSO nor provisioning, such as an older on-premise system, you need a documented manual step in the leaver process with a named owner and a deadline.
An illustrative scenario: a clinic group with six branches
Take a healthcare group in Amman and Irbid with six clinics, around 200 staff and a mix of systems: a clinical system, an accounting package, Microsoft 365, a laboratory portal, a scheduling tool and a handful of shared logins for equipment suppliers’ portals. Staff rotate between branches, and doctors often work part time across several sites.
An internal review finds former employees with active accounts, shared reception logins used by whoever is on shift, and no way to tell who changed a patient appointment. Given the sensitivity of health data and the obligations that come with Jordan’s Personal Data Protection Law of 2023, management wants this fixed properly.
The group uses Microsoft 365 already, so Entra ID becomes the identity provider. The scheduling tool and the laboratory portal support SAML and are connected first. The clinical system vendor offers SSO in a newer version, which goes onto the upgrade plan. Shared reception logins are replaced with named accounts and a fast sign-in on the reception PCs, such as a PIN or fingerprint through Windows Hello. HR’s staff list becomes the source of truth: a new starter, a branch transfer or a leaver is entered there first, and IT acts on that record rather than on a WhatsApp message.
A rollout sequence that avoids lockouts
Moving to SSO touches every user, so the order matters. A sequence that works for most mid-sized organisations:
- Clean the directory first. Make sure every current employee has exactly one account in the identity provider, with correct name, department, manager and job title. Disable accounts for people who have left. This is tedious and it is the foundation.
- Turn on multi-factor authentication. Use an authenticator app rather than SMS where possible. Communicate a week ahead and have IT available on the first morning.
- Inventory your applications. List every system, who owns it, how many users it has, whether it supports SAML or OpenID Connect, and whether it supports provisioning.
- Connect low-risk applications first. Start with a tool used by a small team, test sign-in and sign-out, then move to wider systems.
- Disable local passwords where possible. Otherwise people keep using the old login and the leaver gap remains.
- Keep emergency access. Create one or two emergency administrator accounts, excluded from normal sign-in rules, with credentials stored securely and use monitored. If the identity provider has a problem, you still need a way in.
Joiners, movers and leavers: the process behind the technology
Technology enforces access; the process decides it. Three events need a defined flow. A joiner gets access based on their role, approved by their manager, ready on day one. A mover, someone changing department or branch, gains the new role’s access and loses the old one. Movers are where access quietly accumulates, because people get new permissions and nobody removes the previous ones. A leaver loses everything on their last working day, with ownership of their records and files transferred.
Defining access by role makes this manageable. Instead of deciding permissions person by person, you define what a branch receptionist, an accountant or a sales rep needs, and assign the role. Then, every quarter or half year, ask each system owner to review the list of users and confirm each one still needs access. It takes an hour per system and regularly catches surprises.
Making access control part of how systems are run
A useful first check costs nothing: ask HR for everyone who left in the past year and look for each name in the user lists of your five most important systems. Whatever that finds is the case for single sign-on, tied to HR events, role definitions and regular reviews. That combination is part of how we approach Enterprise Systems work: designing systems and integrations so that access, data and ownership stay under control as the organisation adds tools, branches and people.



