Most small businesses don’t run servers any more. Their mail, files, chat, contacts and finance live in SaaS platforms: Microsoft 365, Google Workspace, a CRM, an accounting app. That moved a lot of security work from hardware to settings.
Each of those platforms has hundreds of settings. Who can share files outside the company. Which apps can connect, and what they can read. Who’s an admin. Whether old sign-in methods still work. Get one wrong and an attacker doesn’t need to break anything. They just use the door you left open.
SSPM (SaaS security posture management) is the practice of checking those settings, admins and connected apps continuously, and fixing what drifts. This post explains what SSPM means, what it covers, how it differs from similar terms, and why AI agents make it a bigger job than it was two years ago.
SSPM meaning, in plain words
SSPM is the SaaS version of a regular health check. Instead of asking once a year whether Microsoft 365 is set up well, it asks every day:
- Settings: are sharing, sign-in and security settings where they should be?
- People with power: who are the admins, and did that change?
- Connected apps: which outside apps can read or change data, and do they still need that access?
- Drift: what changed since yesterday, and did anyone mean to change it?
The “posture” part is the overall shape of all those settings together. The “management” part is the important bit: finding a problem is only useful if someone fixes it and checks it stays fixed.
What SSPM looks at
Here’s what a good SSPM review covers in Microsoft 365 and Google Workspace.
Sharing
Files and sites shared with everyone in the company, or by open links anyone can use. Sharing rules loosened for one project and never put back. Guests who still have access long after the work ended.
Sign-in and admin settings
Whether MFA is actually required, not just available. Whether old protocols that skip MFA still work. How many admins there are, and whether a new one appeared without a ticket. See the Microsoft 365 security checklist for the settings that matter most.
Connected apps
This is the part most reviews miss. Every time someone clicks “Allow” on a sign-in screen, an outside app may get access to mail, files or calendars. That access is an OAuth grant, and it often outlives the reason it was given.
Both platforms let you control this, but the defaults are open:
- Microsoft Learn says users can consent to apps for permissions that don’t need an admin unless you change it, and recommends limiting user consent to apps from verified publishers. New tenants now start with a Microsoft-managed consent setting, so check what each tenant actually uses.
- Google’s Admin Help says that, by default, users can sign in to any third-party app, and those apps can request unrestricted Google data. Admins can mark apps as trusted, limited or blocked under API controls.
Change over time
A one-time review is a photo. SSPM is a video. The question isn’t “is it set up well?” but “what changed, who changed it, and is that all right?”
SSPM vs CASB vs CSPM vs ITDR
These terms get mixed up. Here’s the short version.
| Term | What it watches | Question it answers |
|---|---|---|
| SSPM | Settings, admins and connected apps in SaaS platforms | Is this SaaS platform set up safely, and is it staying that way? |
| CASB | Traffic and activity between users and cloud apps | Who is using which cloud apps, and what data is moving? |
| CSPM | Cloud infrastructure, such as Azure, AWS or Google Cloud | Are our cloud servers, storage and networks configured safely? |
| ITDR | Sign-ins and account activity | Is someone other than the real person using this account? |
They work together. SSPM closes the doors. ITDR catches the person who walks through one anyway. Read what is ITDR for that side.
Why AI agents change SSPM
Two years ago, a connected app was usually something like a calendar tool or an e-signature app. You could list them, check them once a quarter and move on.
AI changed that in three ways.
1. There are many more of them. AI note-takers, writing assistants, chatbots and browser extensions all ask for access to mail and files. Staff connect them in seconds, often with a personal sign-in. Read shadow AI detection for how fast this grows.
2. Agents act, not just read. An AI agent can send mail, move files, raise tickets and call other apps. Its access is no longer just a privacy question. It’s a question of what it can do on its own.
3. Apps now connect to apps. An agent in one platform can hold a grant to another. If one app is compromised, its access can be used to reach the next. Posture now has to include app-to-app access, not just person-to-app.
That’s why SSPM increasingly overlaps with non-human identity security. Every agent and connected app is an identity with access. It needs an owner, a reason to exist and a review date, the same as a person’s account.
A worked example: a 40-person marketing agency
This is an illustrative example, not a real client. An MSP runs its first SSPM-style review of an agency’s Microsoft 365 tenant.
Settings. Sharing is open to anyone with the link by default. The MSP changes the default to people in the organisation and sets open links to expire.
Admins. Five Global Administrators, including a freelancer who left in the spring. The MSP removes the freelancer and cuts the rest to two, with separate admin accounts.
Connected apps. 31 outside apps hold access. Eleven are AI tools: four AI note-takers, three writing assistants, two chatbots and two “AI agent” apps. One note-taker has read access to every mailbox, approved by a junior designer. One agent app can send mail as the account manager who connected it. Nobody owns either.
The fix. User consent is limited to verified publishers and low-risk permissions. Every AI app gets an owner, or its access is removed. The two that read or send mail across the company go to the agency owner for a decision.
Then the important part. Three weeks later, a new AI scheduling assistant appears with calendar and mail access. It’s flagged the same day, not at next year’s review.
SSPM checklist for Microsoft 365 and Google Workspace
- Default sharing is “people in the organisation”, not “anyone with the link”.
- Open links on sensitive sites and folders are removed or set to expire.
- Guests are reviewed and removed when the work ends.
- MFA is required for everyone, and old sign-in methods are blocked.
- Admins are few, named and reviewed monthly.
- User consent to apps is limited (Microsoft) or apps are set to trusted, limited or blocked (Google).
- Every connected app has an owner and a reason, or its access is removed.
- AI apps and agents are on the same list, with what each one can reach.
- App-to-app access is checked, not just person-to-app.
- Changes are watched continuously, with someone alerted when a setting drifts.
How Centeye helps
Centeye Synthwatch (early access) brings agentic SaaS security to Microsoft 365, Google Workspace and connected apps. It finds the agents and copilots inside your SaaS platforms, governs the AI apps connected through OAuth, and traces app-to-app access before one compromised app reaches the next. Every agent gets an identity, an owner and an off switch.
It works alongside Centeye Lumen, which keeps a register of every AI tool your people use, and Centeye Gatewatch, which watches the sign-ins and admin changes that SSPM settings are meant to prevent.
To see where one tenant stands today, try the free AI Exposure Check or the Identity Exposure Check.
Frequently asked questions
What does SSPM stand for?
SaaS security posture management. It means checking the settings, admins and connected apps of SaaS platforms continuously, and fixing what drifts.
Do small businesses need SSPM?
They need what SSPM does. A small business on Microsoft 365 or Google Workspace has the same settings and connected-app risks as a large one, with less time to check them. Many get it as part of a managed service from their MSP.
Is SSPM the same as CSPM?
No. CSPM checks cloud infrastructure, such as servers and storage in Azure or AWS. SSPM checks SaaS platforms, such as Microsoft 365, Google Workspace and CRM tools.
How do AI agents affect SaaS security posture?
Agents connect to SaaS platforms through OAuth, often with broad access, and can act on their own. Each one adds an identity to manage. Posture now has to include what every agent can reach and who owns it.
How often should SaaS settings be reviewed?
Continuously, if possible, with a full review at least monthly. Settings, admins and connected apps change every week, and a yearly review misses most of it.
New to the terms? See SSPM, OAuth grant, shadow IT and non-human identity in our glossary.
Microsoft user consent defaults and recommendations from Microsoft Learn (Configure how users consent to applications), November 2026. Google Workspace third-party app access defaults and API controls from Google Workspace Admin Help, November 2026. Settings and defaults change, so check each tenant. The worked example is illustrative. Microsoft, Microsoft 365 and Entra are trademarks of the Microsoft group of companies. Google Workspace is a trademark of Google LLC.