SOC 2 · Control requirements

SOC 2 Control Requirements

Practical answers for the IT, engineering and security teams who build and run SOC 2 controls: what auditors expect to see, and the evidence that proves it.
SOC 2 controls

System and scope

1. How do we establish an accurate inventory of the systems and assets in scope?

List every application, server, database, cloud account, device, and third-party service that stores, processes, or protects in-scope data, with an owner for each. Pull the list from your cloud consoles, identity provider, and endpoint tools rather than memory, and review it at least quarterly. An accurate inventory is the foundation of your scope; anything missing is a gap the auditor can find.

2. How should we document our commitments to customers and system requirements?

Collect the security, availability, and privacy promises you make in contracts, service agreements, terms of service, and your website, such as uptime targets, encryption, and data retention. Record them in your system description and map each to the controls that meet it. Auditors test your controls against these commitments, so promises you can’t support become findings.

3. How do we address customer responsibilities in our SOC 2 report?

List the controls your customers must operate for your service to stay secure, such as managing their own user access or protecting their API keys. These appear in the report as complementary user entity controls. Stating them clearly sets fair expectations and keeps you from being held responsible for controls you don’t run.

4. How do we decide whether to include or carve out a subservice organization?

A subservice organization is a vendor that performs controls on your behalf, such as your cloud host. Most companies use the carve-out method: the vendor’s controls are excluded, and you monitor them through the vendor’s own SOC report. Inclusion is rare because the vendor must take part in your examination. With a carve-out, the report still names the vendor and the controls you expect it to perform.
SOC 2 controls

People and access

5. How are remote employees and contractors addressed in our SOC 2 controls?

They are held to the same controls as on-site staff: background checks where appropriate, signed policies and confidentiality agreements, security training, managed devices with encryption and endpoint protection, multi-factor authentication, and timely offboarding. For contractors, define these requirements in the contract and check them. Auditors sample remote and contract workers the same way they sample employees.

6. What access controls and onboarding and offboarding processes do we need?

Grant access based on role and least privilege, with manager approval and recorded in a ticket or workflow. Require multi-factor authentication on critical systems. On departure, remove access promptly, ideally the same day, using a checklist that covers every system. Auditors sample new hires and leavers, so every step needs a dated record.

7. What security awareness training should employees complete?

Everyone in scope should complete security awareness training at hire and at least annually, covering phishing, password and MFA use, data handling, incident reporting, and acceptable use. Add AI use guidance if staff use AI tools. Keep completion records, because auditors test them. Phishing simulations are a strong addition, not a requirement.

8. How do we demonstrate that periodic user access reviews are effective?

Review who has access to each in-scope system on a set schedule, commonly quarterly. The system owner confirms each user still needs their access, and removals are completed and recorded. Keep the reviewed user list, the reviewer’s sign-off and date, and proof that flagged access was removed. A review with no follow-up does not count as effective.

9. How should we manage privileged accounts and service accounts?

Limit administrator access to the few people who need it, use separate admin accounts with multi-factor authentication, and log their activity. Assign each service account an owner, grant only the minimum permissions they need, and store secrets in a vault and rotate them. Review both types in your access reviews. Privileged access is one of the first areas auditors test.
SOC 2 controls

Operations

10. How should we document incident response and test our response procedures?

Write an incident response plan that defines roles, severity levels, escalation, customer and regulator notification, and lessons learned. Test it at least annually with a tabletop exercise, and record who took part, the scenario, and any changes you made afterward. Log real incidents in tickets from detection to closure. Auditors look for a plan that is tested and actually followed.

11. What vulnerability scanning and penetration testing should we complete?

Run automated vulnerability scans of your infrastructure and applications regularly, typically monthly or continuously, and fix findings within severity-based timeframes. SOC 2 does not list penetration testing as a strict requirement, but an annual third-party penetration test is common and strong evidence. Keep the reports and proof that findings were fixed.

12. How do we demonstrate that security monitoring and logging controls are working?

Collect logs from your cloud accounts, identity provider, applications, and critical systems in a central place, retain them for a defined period, and set alerts for suspicious activity. Then show the alerts are reviewed and acted on: tickets for investigated alerts, review records, and escalations. Logging that nobody watches does not satisfy auditors.

13. What backup and disaster recovery evidence should we maintain?

Keep your backup configuration and schedule, records showing backups succeed, and the results of periodic restore tests. Maintain a business continuity and disaster recovery plan with recovery time and recovery point objectives, and test it at least annually. A restore test proves backups work; a backup that has never been restored is an untested control.

14. How should we document and approve changes to applications and infrastructure?

Run every production change through a defined process: a ticket or pull request, peer review, testing, approval, and a record of deployment. Infrastructure as code and protected branches make this evidence automatic. Define how emergency changes are handled and reviewed afterward. Auditors sample changes and trace each one from request to approval to deployment.
SOC 2 controls

Engineering

15. What encryption and key management practices should we evaluate during readiness?

Confirm that customer data is encrypted in transit with current TLS and at rest in databases, storage, and backups. Check that encryption keys are managed in a key management service or vault, access to keys is restricted and logged, and keys are rotated on a schedule. Document the standards in an encryption policy so auditors can match practice to policy.

16. What evidence supports secure software development and code review practices?

Auditors look for a documented secure development policy, peer-reviewed pull requests with required approvals, branch protection, automated security testing such as code and dependency scanning, and developer security training. Your repository history provides most of this evidence automatically, so configure your tools to enforce review instead of relying on habit.

17. How should we separate development, testing, and production environments?

Keep production in separate accounts, projects, or networks from development and testing. Limit production access to those who need it, and deploy through your pipeline rather than making manual changes. Keep real customer data out of non-production environments, or mask it. Separation reduces the risk that a test mistake or a compromised developer account reaches customer data.
Next step

Tell us where you stand

Tell us what you’re trying to accomplish. We’ll reply with clear next steps.