SailPoint leadership change signals tighter focus on agentic identity

At a glance

What this is: SailPoint has named Levent Besik chief product officer, and the announcement frames agentic AI identity governance as the company’s key product focus.

Why it matters: IAM, NHI, and security teams need to separate leadership signals from control reality, because agentic identity demands governance models that go beyond human-centric access assumptions.

👉 Read SailPoint's announcement on its new chief product officer and agentic identity focus

Context

The identity governance gap is widening as enterprises add agents, workloads, and service identities to environments that were built around human access patterns. In this context, SailPoint’s leadership move matters because the company is explicitly tying product direction to agentic AI identity security rather than treating it as a side capability.

For IAM practitioners, the practical question is whether current governance models can handle identities that act across clouds and platforms with runtime variability. That pushes the conversation toward lifecycle control, privilege boundaries, and accountability for non-human and agentic actors, not just traditional user access.

Key questions

Q: How should security teams govern AI agents alongside service accounts and human users?

A: Treat AI agents as a distinct governed actor class, not as a renamed service account or a user proxy. Separate policy, logging, approval, and review paths by actor type, then align access scope to the task being performed. Governance only works when the programme can tell who or what executed the action and why that identity had access.

Q: When does zero-standing privilege matter most for non-human identities?

A: Zero-standing privilege matters most when machine identities touch sensitive data, production systems, or cross-cloud workflows. Persistent access creates avoidable blast radius and weakens accountability because entitlements stay valid long after the original need has passed. Time-bound access should be the default wherever operations can tolerate it.

Q: What do identity teams get wrong about agentic AI governance?

A: They often assume existing IAM controls will scale if the new actor simply receives a role or credential. Agentic systems change the problem because the actor can initiate actions at runtime across tools and environments, so static provisioning logic no longer fully describes risk. Governance must reflect runtime behaviour, not just account setup.

Q: Who should own accountability when AI agents operate across multiple systems?

A: Accountability should sit with the team that defines the agent’s scope, data access, and failure handling, not with the platform team alone. Cross-system agent use creates shared responsibility, so ownership has to cover policy design, operational oversight, and incident response. If no owner can explain the actor’s boundary, governance is already failing.

How it works in practice

Agentic identity governance across clouds and platforms

Agentic identity governance is the problem of controlling software actors that can access data, call tools, and execute tasks across multiple environments. Unlike static service accounts, agents may need contextual access that changes by session, workload, or task. The core challenge is not simply authentication, but managing what an agent can do, where it can do it, and under which oversight model. That requires identity systems to treat the agent as a governed actor rather than a fixed credential with a single role.

Practical implication: map agent access paths by environment and task scope before allowing production use.

Zero-standing privilege for non-human identities

Zero-standing privilege removes persistent access and replaces it with on-demand, time-bound entitlements. For NHIs and AI agents, that matters because long-lived credentials create unnecessary attack surface and make attribution harder when access is reused across systems. In agentic settings, the security question is not only whether access exists, but whether it exists only when needed and only for the narrowest useful purpose. Without that, governance becomes backlog management instead of risk reduction.

Practical implication: reduce standing entitlements for machine and agent identities wherever task-scoped access is feasible.

Contextualized risk as a control model

Contextualized risk means access decisions are shaped by identity type, resource sensitivity, and behavioural context, rather than by a one-time role assignment. That is increasingly relevant when AI agents operate alongside human users and machine identities in the same workflow. The control model needs to distinguish who initiated the action, which identity actually executed it, and whether the resulting access path matches policy intent. This is where identity governance and security operations start to converge.

Practical implication: require policy decisions to account for actor type, resource sensitivity, and session context together.

NHI Mgmt Group analysis

Agentic identity is becoming a product and governance boundary, not just a feature request. SailPoint’s announcement shows that agentic AI is now shaping identity security roadmaps at the leadership level, which means practitioners should expect more pressure to govern software actors as first-class identities. The market is moving from abstract AI readiness language to concrete control requirements around access, accountability, and lifecycle management. Practitioners should evaluate whether their programmes can classify and govern agentic actors separately from human users and routine workloads.

Zero-standing privilege is moving from a best practice to a baseline expectation for machine access. The vendor’s own positioning reflects a broader shift in the identity market: long-lived non-human access is becoming harder to defend when environments are increasingly dynamic and distributed. That does not mean every credential can be ephemeral, but it does mean persistent entitlements now require explicit justification. Practitioners should re-check where standing machine privilege remains embedded in operational workflows.

Contextualized identity controls matter more when access is shared across clouds, tools, and actors. AI agents and NHIs create overlapping execution paths that traditional role models do not fully describe, especially when the same system can be touched by a human, a workload, and an agent in one chain. The governance problem is therefore not just access grant or revocation, but differentiating intent and responsibility at runtime. Practitioners should rework policy design so actor type is part of every access decision.

Human IAM lessons are not enough for autonomous and machine identity governance. The announcement underscores a category shift: the same enterprise identity stack now has to deal with people, workloads, and agentic systems under one governance model. That creates a stronger need for lifecycle discipline, privileged access review, and entitlement scoping that apply consistently across actor types. Practitioners should stop treating NHI governance as a separate island from broader identity governance.

Agentic identity governance will increasingly determine platform credibility. As more vendors align product strategy to agentic AI, identity teams will need to judge platforms by whether they can express policy, scope access, and preserve accountability across non-human actors. The field is moving toward identity systems that can support both operational speed and provable control boundaries. Practitioners should track whether their chosen stack can govern agents without collapsing them into generic service accounts.

From our research:

• 97% of NHIs carry excessive privileges, increasing unauthorised access and broadening the attack surface, according to Ultimate Guide to NHIs.
• Only 5.7% of organisations have full visibility into their service accounts, which makes entitlement governance and accountability materially harder across machine identities.
• Forward pivot: 79% of organisations have experienced secrets leaks, with 77% of these incidents resulting in tangible damage, according to Ultimate Guide to NHIs.

What this signals

Agentic identity programmes will be judged on whether they can separate actor type from access method. The next governance failure is not missing authentication, but misclassifying an agent as a normal workload or user proxy. Teams that keep those boundaries explicit will be able to apply policy, review, and containment differently across humans, NHIs, and autonomous systems.

Ephemeral access only helps when the surrounding lifecycle is also coherent. If creation, approval, review, and retirement still follow human-paced processes, machine and agent identities will outgrow the controls meant to govern them. The practical test is whether the programme can explain who owns each non-human identity from issuance to offboarding.

According to Ultimate Guide to NHIs, 92% of organisations expose NHIs to third parties, which means agentic workflows are increasingly inheriting external trust relationships as well as internal ones. That makes the boundary between internal governance and supply-chain control much thinner than most IAM roadmaps assume.

For practitioners

• Reclassify agent identities separately Inventory AI agents, service accounts, and human accounts as distinct governance classes so policy, review, and escalation paths do not blur actor type.
• Review standing privilege in machine workflows Identify any persistent entitlements used by workloads, integrations, or agents and determine where time-bound access can replace them without breaking operations.
• Add actor type to access decisions Require policy logic to include whether the requester is human, NHI, or agentic, then align approval, logging, and recertification accordingly.
• Test lifecycle controls against agent behaviour Validate whether joiner, mover, and leaver processes still work when the identity can be created, changed, or retired by software rather than a person.

Key takeaways

• This announcement is a market signal that identity security is shifting toward agentic and non-human governance, not just human access management.
• Persistent machine privilege remains a major control weakness, especially as identity stacks absorb more cloud and AI-driven workflows.
• Practitioners should test whether their IAM programme can distinguish humans, NHIs, and agents at policy, lifecycle, and accountability levels.

Key terms

• Agentic Identity: An agentic identity is the identity representation used by software that can decide and act at runtime, including when to call tools, access data, or continue a task. In governance terms, it needs lifecycle, policy, and accountability controls that can follow dynamic behaviour rather than fixed user patterns.
• Zero-Standing Privilege: Zero-standing privilege is a model where access is not continuously present but granted only when needed for a specific task or session. For NHIs and agents, it reduces persistent exposure and makes governance more precise because access exists for execution, not as a permanent entitlement.
• Contextualized Risk: Contextualized risk is an access decision model that considers the actor, the resource, the session, and the business purpose before granting or continuing access. It is especially important where humans, workloads, and agents operate in the same environment and simple role mappings no longer describe real risk.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are responsible for identity security strategy or NHI governance in your organisation, it is worth exploring.

This post draws on content published by SailPoint: SailPoint appoints Levent Besik as Chief Product Officer. Read the original.