Keep workflow orchestration in the automation platform, but keep entitlement policy, review, and revocation authority in IAM or IGA. If the same system both moves the request and becomes the source of truth for access decisions, accountability becomes harder to prove and exceptions become harder to govern.
Why the control boundary matters
Security teams get the cleanest operating model when automation moves work and access control decides authority. Orchestration can open tickets, route approvals, trigger provisioning, and coordinate handoffs, but the policy engine should remain the place where entitlements are defined, reviewed, and revoked. That separation keeps the decision logic auditable instead of burying it inside a workflow tool.
When one platform becomes both the delivery mechanism and the source of truth for access, teams often lose a reliable answer to a simple question: who approved what, under which policy, and how was it removed later? A separate IAM or IGA control plane gives the business a stable governance record even when the automation layer changes frequently.
That distinction is also useful for exception handling. Automation is good at repeating a process, but access exceptions need policy context, ownership, and periodic review. If the automation system itself decides access, it becomes much harder to prove that a deviation was temporary, approved, and later cleaned up.
How to split workflow orchestration from entitlement authority
The practical split is straightforward: let the automation platform execute steps, collect inputs, and notify approvers, while IAM or IGA owns roles, entitlements, review campaigns, and revocation rules. This is the right design when the workflow is about moving requests or evidence, not about inventing the access model itself.
That separation works best when access decisions are expressed as policy, not embedded as brittle workflow branches. For a useful comparison of access model choices, Authorisation Models Guide is a strong reference point. For the governance side of the house, IAM and IGA Basics helps frame why provisioning, access review, and entitlement ownership should stay separate from orchestration.
At scale, the split also prevents role drift. If the same automation layer is allowed to create and interpret access logic, teams usually end up with hidden exceptions, duplicated rules, and unclear ownership. Keeping authority in IAM or IGA lets you standardise access outcomes while still allowing the automation system to vary the process around them.
What breaks when automation and access control are merged
The main failure mode is governance loss, not just technical complexity. Once workflow and authority are fused, audit evidence becomes harder to assemble, because the team must reconstruct whether the system merely moved a request or actually changed the entitlement model. That makes reviews slower and revocation less reliable.
A second failure mode is overreach. Orchestration systems are often built for speed, so they accumulate broad service permissions to reach downstream tools. If those permissions are also used to make access decisions, a compromise of the automation layer can turn into a privilege problem. For an access-control lens on that risk, Privileged Access Management Guide is useful because it shows why privilege, session control, and revocation need separate governance.
Finally, teams underestimate how quickly review quality degrades when the automation platform becomes the policy source. The reviewer may be looking at a completed workflow and assuming the access is still valid, when in fact the entitlement has no independent owner, expiry, or recertification path. That is where separation pays off operationally.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
NIST SP 800-53 Rev 5 sets the technical controls, while ISO/IEC 27001:2022 defines the regulatory obligations.
| Framework | Control / Reference | Relevance |
|---|---|---|
| NIST SP 800-53 Rev 5 | AC-6 — Least Privilege | Separating workflows from entitlement authority enforces minimal access in the access-control plane. |
| AC-2 — Account Management | The question concerns who owns provisioning, review, and revocation of access. | |
| IA-5 — Authenticator Management | Access control depends on managing credentials and related identity material separately from workflow orchestration. | |
| Recommendation — Apply AC-6 to keep entitlement decisions and revocation under least-privilege governance. Use AC-2 to centralise account lifecycle and review authority in IAM or IGA. Use IA-5 to govern credential lifecycle outside the automation workflow. | ||
| ISO/IEC 27001:2022 | A.5.15 — Access control | The subject is the separation of access policy from process automation. |
| A.5.18 — Access rights | Review and revocation authority are central to the question. | |
| Recommendation — Define access control ownership in the ISMS and keep orchestration outside entitlement decisions. Review and revoke access rights through a dedicated governance process, not the workflow engine. | ||
Practitioner Guidance
What to verify: Confirm that every access grant has an owner, a policy source, and a revocation path that do not depend on the orchestration tool staying online. If the same platform can approve, provision, and retain the authoritative record, treat that as a design smell.
Decision rule: If the automation system only routes work and records status, it can sit upstream of IAM or IGA. If it can alter entitlements or decide exceptions, move that authority back into the identity control plane and keep the workflow layer as an executor, not the decision maker.
Practitioner takeaway: The safest pattern is not “more automation everywhere,” but a clear split where automation accelerates the process and IAM or IGA remains the accountable authority for access.
Related resources from NHI Mgmt Group
- How should security teams run access reviews for non-human identities?
- How should security teams govern non-human identities that have persistent access?
- How should security teams govern API keys used for generative AI access?
- How should security teams separate access control from access management?