Topic starter
24/06/2026 11:32 pm
TL;DR: AI for access administration uses access intelligence, role mining, provisioning automation, and behavioural analysis to reduce privilege creep and improve certification quality as teams manage hundreds of SaaS apps, according to SecurEnds. The real shift is not automation alone but tighter governance over how access decisions are made, reviewed, and evidenced.
Editorial analysis by NHI Mgmt Group, based on content published by SecurEnds: “AI for Access Administration: From Promise to Practice”.
Key questions
Q: How should organisations use AI in IAM without weakening governance?
A: Use AI for pattern detection, summarisation, and query generation, but keep access approval, role creation, and policy enforcement under human accountability.
Q: Why do manual access reviews fail in growing enterprises?
A: Manual reviews fail because the review process cannot keep up with the rate of entitlement change.
Q: What signals show that access administration is drifting out of control?
A: Common signals include outdated roles, dormant permissions, long approval queues, repeated review exceptions, and accounts that still hold access after job changes or exits.
Practitioner guidance
- Standardise identity data before automation Clean HR attributes, job titles, department mappings, and group membership data before letting AI influence access decisions or review recommendations.
- Use AI to surface role drift Compare AI-mined permission clusters against current job functions and retire roles that persist only because they are familiar, not because they still match work patterns.
- Tighten certification criteria Require reviewers to assess whether a permission is actively used, aligned to the role, and still needed rather than approving long entitlement lists by default.
Bottom line: Manual access administration does not scale well when applications, roles, and entitlements multiply across the enterprise.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
25/06/2026 8:29 am
Manual access administration fails because review cadence was designed for a slower identity environment. Spreadsheets and manager approvals assume access state changes slowly enough to be understood after the fact. That assumption breaks in SaaS-heavy estates where roles, entitlements, and job context shift continuously. The implication is not that human review disappears, but that it can no longer be the primary mechanism for keeping access current.
A question worth separating out:
Q: How should organisations decide whether AI can auto-provision access or only recommend it?
A: Use auto-provisioning only where the entitlement is low risk, the role model is stable, and the identity data is reliable. Anything high impact, ambiguous, or exception prone should remain recommendation-only until the organisation can prove the decision pattern is consistent and defensible.
👉 Read our full editorial: AI for access administration exposes the limits of manual IAM