BYOA and trillions of agents will strain enterprise governance

At a glance

What this is: ArmorCode says BYOA, Shadow AI, and low-cost workflow creation are pushing enterprise environments toward trillions of agents, turning agent governance into a scale problem rather than a tooling problem.

Why it matters: IAM, PAM, and AI governance teams need to treat agent creation, access, context, and closure as lifecycle controls, because unmanaged agents can expand enterprise exposure faster than traditional review models can track.

Context

BYOA, or Bring Your Own Agent, describes the use of self-built or employee-created AI agents that can trigger workflows, make probabilistic decisions, and access enterprise systems. The governance gap is that these actors can be created quickly, move across systems, and persist outside normal application ownership and lifecycle assumptions.

The article is really about governance failure under scale: enterprises are adding agents faster than they can assign ownership, constrain context, or prove closure. That intersects directly with IAM and NHI governance because these agents behave like non-human actors with delegated access, yet many organisations still manage them as if they were just another software tool.

Key questions

Q: What breaks when employees can create their own AI agents without governance?

A: The control that breaks is lifecycle ownership. When employees can create agents quickly, the organisation loses a stable way to bind access, context, and accountability to a named owner. That turns each agent into a non-human identity problem, because the access path may persist after the business need changes or the creator leaves.

Q: Why do multi-hop AI agent workflows create more risk than single-agent automation?

A: Because each additional hop creates another place where scope can drift, tokens can be exchanged, and intent can be altered without a clear human checkpoint. The risk is not only compromise. It is semantic loss of control across a chain that still looks authenticated at every step.

Q: How do security teams know if agent governance is actually working?

A: It is working only if the team can answer three questions quickly for any agent: what it can reach, what it did recently, and whether that behaviour matches intent. If any of those answers require manual reconstruction, governance exists on paper but not in operations.

Q: Should IAM teams treat BYOA differently from ordinary application onboarding?

A: Yes. Ordinary application onboarding assumes a relatively stable system with a predictable owner and lifecycle. BYOA introduces distributed, fast-moving, employee-created actors that may outlive the role or team that created them. IAM teams should govern them as living identities with creation, review, revocation, and closure controls.

Technical breakdown

Why BYOA changes the access model for agents

BYOA collapses the gap between software creation and business use. A non-technical employee can spin up an agent that accesses enterprise systems, chains API calls, and persists across tasks without passing through the controls normally used for application onboarding. That creates a governance problem similar to unsanctioned service accounts, except the actor may also carry memory, task intent, and delegated decision logic. In identity terms, the issue is not simply whether the agent authenticates, but who owns its authority, what it can reach, and how it is retired when the business need ends.

Practical implication: Treat employee-created agents as governed identities, not informal automations.

Why persistent memory makes agent sprawl harder to contain

The article distinguishes between passive Shadow AI and persistent agents. Passive use leaks information, but persistent agents store context, act laterally, and continue operating without explicit approval at every step. That matters because memory turns a one-time prompt into an ongoing exposure surface. Once the agent can remember prior interactions, it can inherit business context, credentials, and workflow state across systems. This is exactly where conventional IAM review logic struggles: certification looks backward, while agent behaviour can change dynamically between one task and the next.

Practical implication: Bind agent memory, scopes, and authorisations to a short-lived lifecycle with explicit ownership.

Why detection alone cannot govern autonomous workflows

ArmorCode’s core claim is that the bottleneck is no longer finding issues, but deciding what matters, who owns it, and when it is closed. In agentic environments, vulnerability discovery may scale, but closure becomes the limiting factor because thousands of autonomous workflows can create overlapping exposure. Governance has to move from static assessment to continuous context-aware control. For IAM and PAM teams, that means the relevant question is not only whether access exists, but whether access is still justified for an agent that may self-spawn, self-chain, and self-retire outside human review windows.

Practical implication: Shift controls toward ownership assignment, context tagging, and closure workflows at issuance time.

NHI Mgmt Group analysis

BYOA is an identity governance problem before it is an AI problem. The article shows that when employees can create agents in minutes, the organisation inherits a new class of non-human actor that behaves more like a delegated identity than a conventional app. That moves the issue into IAM and NHI governance, because the real control question becomes who owns the agent, what it can do, and how it is offboarded. Practitioners should treat agent creation as a lifecycle event, not a convenience feature.

Agent sprawl creates governance debt faster than security teams can review it. Persistent agents, memory, and cross-system chaining break the assumption that access is stable long enough to be certified after the fact. The article is right that detection scales faster than remediation, but governance is the deeper failure mode: organisations are creating digital actors without a durable ownership model. Practitioners need controls that assign responsibility at creation time and preserve that responsibility through closure.

Trillion-agent forecasts are less important than the operating model they imply. The number itself is directional, but the structural point is clear: heterogeneous estates, M&A fragmentation, SaaS sprawl, and agent creation all converge on the same governance layer. This is where agent ownership drift becomes the named concept to watch: the gap between who creates an agent, who benefits from it, and who remains accountable for its access over time. Practitioners should build governance around that drift before scale makes it unmanageable.

The market is moving from detection-centric security to control-plane governance. The article frames security as an analytics and orchestration problem, which is the right direction for environments full of autonomous actors. For NHI and agentic AI programmes, that means the differentiator is not how many agents can be found, but how reliably their context, access, and closure are governed across systems. Practitioners should prepare for agent control planes to sit alongside identity controls rather than outside them.

Shadow AI and BYOA are converging into one lifecycle risk. The article separates unsanctioned AI use from persistent agents, but in practice the boundary will blur as people keep the agents they build and reuse them across roles and employers. That makes offboarding, revocation, and ownership transfer critical. Practitioners should assume that unmanaged agents will outlive the context in which they were created unless lifecycle controls are explicit.

From our research library:

• Only 13% of organisations feel extremely prepared for the reality of agentic AI despite the majority racing toward autonomous adoption, according to the 2026 Infrastructure Identity Survey.
• Read next: Agentic AI Identity Guide

What this signals

Agent ownership drift: the central governance issue in BYOA is not merely that agents exist, but that ownership, context, and closure can separate as they move across tools and teams. Once that drift starts, access review becomes a lagging indicator rather than a control. Practitioners should design lifecycle controls that force an owner to remain attached to the agent until retirement.

The 2026 Infrastructure Identity Survey found that only 13% of organisations feel extremely prepared for the reality of agentic AI despite the majority racing toward autonomous adoption. That gap suggests most programmes are still building governance reactively, after the operating model has already shifted. Practitioners should assume the control plane must mature before the agent population does.

For practitioners

• Define agent ownership at creation Require every employee-built agent to have a named owner, business purpose, and explicit approval path before it can reach enterprise data or systems.
• Bind agent access to short-lived scopes Limit each agent to task-scoped permissions and revoke access when the task, context, or sponsoring business need ends.
• Track agent memory and context retention Inventory where agents store memory, what context persists between sessions, and which data classes can enter that memory layer.
• Treat agent offboarding as a lifecycle control Remove agent credentials, API grants, and workflow hooks when an employee leaves, changes role, or no longer needs the agent.
• Create a closure workflow for autonomous exposures Route unresolved agent exposures into an explicit remediation queue so ownership, priority, and closure do not depend on ad hoc follow-up.

Key takeaways

• BYOA turns employee-created agents into governed actors, which means IAM and NHI teams need lifecycle controls, not just visibility.
• The article’s scale argument matters because persistent agents, memory, and cross-system chaining make ownership and closure the real bottlenecks.
• The most defensible control response is to assign ownership at creation, constrain scope, and enforce offboarding for every autonomous workflow.

Key terms

• Bring Your Own Agents (BYOA): A pattern where employees build and deploy AI agents outside central security review. These agents often connect directly to enterprise systems, creating identity, access, and accountability issues that look like shadow IT but behave more like shadow AI.
• Ownership Drift: Ownership drift occurs when the person or team recorded as responsible for an NHI no longer matches operational reality. It often appears after reorganisations, platform migrations, or inherited service accounts. Drift weakens response, rotation, and certification because the governance record no longer points to the right decision-maker.
• Persistent Agent: A persistent agent is a non-human identity that can continue checking, adjusting, or revisiting its work over time. Unlike one-shot automation, persistence expands the control problem because the agent can keep acting after the original task has changed or failed.
• Closure Workflow: The operational process used to assign, track, and complete remediation for an access issue or exposure until it is formally resolved. In agentic environments, closure must be tied to ownership and scope, or exposures can accumulate faster than teams can clear them.

Deepen your knowledge

NHI Foundation Level course, the industry's only accredited NHI security programme, covers NHI governance, machine identity security, and agentic AI identity. It helps practitioners translate identity controls into operating models for autonomous systems and non-human actors.