The Non-Human Workforce Has No Manager

Enterprises have spent the past eighteen months deploying AI agents at a pace that has outrun every governance structure designed to control them. A survey of 1,879 IT leaders by OutSystems, published in its 2026 State of AI Development report, puts the paradox in sharp relief: 97% of enterprises are actively exploring agentic AI strategies, and 49% describe their internal capabilities as advanced or expert. Yet only 12% have centralized control over the agents they have built and deployed. The gap between operational ambition and institutional readiness is not a rounding error. It is the defining risk profile of enterprise AI in 2026, and it is widening faster than most boardrooms have acknowledged.

The governance problem is structural, not technical. Current identity and access management frameworks were built for human workers — discrete individuals who authenticate once, operate within defined hours, and leave an auditable trail of decisions with a name attached. As Forbes reported on April 7, the next challenge is IAM for a non-human workforce, one that may soon outnumber the human one in many organizations. Agent interactions occur in milliseconds, autonomously, crossing every network boundary without pause. There is no badge swipe, no login prompt, no manager approval. A single misconfigured agent can propagate decisions across dozens of systems before any human has had a chance to intervene. The question of who is responsible — the vendor, the integrator, the CISO, the business unit that commissioned the deployment — remains almost entirely unresolved in most enterprise legal and compliance frameworks. The network is emerging as the only layer offering full visibility into environments where autonomous agents operate, and most organizations have not yet recognized this as an accountability architecture problem.

The trust deficit is measurable, and the figures are more alarming than most boardrooms have acknowledged. Even in financial services — the sector most accustomed to operating in high-stakes, high-accountability environments — complete trust in autonomous agents stands at just 12%, the highest reading of any industry surveyed. The implication is not that financial services is behind; it is that financial services, with its decades of compliance infrastructure and regulatory scrutiny, represents the ceiling of institutional confidence. Every other sector is below it. Meanwhile, Anthropic launched its Managed Agents infrastructure on April 8, offering out-of-the-box tooling to deploy fleets of Claude agents at eight cents per agent runtime hour plus model usage, with early adopters including Notion, Rakuten, and Asana. Anthropic's annualized revenue reached $30 billion in March 2026, up from $9 billion at the end of last year. The infrastructure for mass agent deployment is now a commodity available at any scale. The governance infrastructure to match it does not yet exist.

The organizational strain is showing in the survey data and in the language executives are using to describe it. Writer's 2026 Enterprise AI Survey, drawn from 1,200 C-suite executives and 1,200 employees, found that 79% of organizations face challenges adopting AI — a double-digit increase from 2025. More striking still: 54% of C-suite executives describe AI adoption as tearing their company apart. That is not hyperbole about integration difficulty. It is the language of institutional friction — between the pace at which agents are being deployed and the speed at which management structures, accountability chains, and decision rights are being updated to reflect the new operational reality. Fortune and Okta framed it directly on April 13: AI agents are now operators, acting on their own accord without the need for a human manager, while corporate structures still treat them like software. The real threat, as that coverage put it, is not how intelligent agents are, but how much authority executives delegate to them without first building the frameworks to oversee that authority.

Business Implications

The governance gap has immediate and concrete consequences across the C-suite. For the CFO, the risk is balance-sheet exposure that has not yet been priced: agents operating without centralized oversight can commit resources, execute contracts, and trigger workflows with financial consequences that surface only after the fact, creating liability that existing insurance frameworks were not designed to cover. For the CIO, the operational risk is compounded by the vendor landscape — enterprise SaaS incumbents including Salesforce, ServiceNow, and Workday have seen their stocks fall more than 30% since January 2026, with roughly $2 trillion in market capitalization erased as organizations recalibrate how much they are willing to pay for platforms that agents are beginning to displace or bypass entirely. CIOs taking a harder line with vendors are simultaneously inheriting responsibility for the agent fleets their own organizations are deploying without equivalent governance discipline applied to their internal infrastructure. For the board, the governance gap is a fiduciary matter: directors who have approved AI investment strategies without requiring parallel investment in agent oversight, identity management, and accountability frameworks are exposed. The OutSystems data — 97% exploring agentic strategies, 12% with centralized control — describes an organization in which the accelerator has been pressed to the floor while the governance engineers are still reading the manual.

ZeroForce Perspective

The AI governance gap is not a failure of imagination. It is the predictable consequence of organizations that treated agent deployment as a technology rollout rather than an organizational redesign. Every governance framework in existence was built around a human workforce — individuals with job titles, managers, performance reviews, and legal accountability. None of those constructs map cleanly onto a fleet of agents executing tasks in parallel, at machine speed, across every system the enterprise operates. The Zero Human Company thesis does not require a dystopian reading of this moment; it requires a clear-eyed one. Organizations that solve the governance problem — that build the accountability architecture, the identity infrastructure, and the decision rights framework for a non-human workforce before their competitors do — will be positioned to scale agent deployment without the institutional friction that is currently tearing 54% of C-suites apart. The companies still treating agents like software will discover, one incident at a time, that autonomous operators require autonomous governance. The window to build that infrastructure proactively is closing. The enterprises that move now are not managing a technology risk. They are designing the operating model of the next decade.