Ed Zitron is right that AI currently can’t prove ROI—but wrong about why.
Only twenty-nine percent see significant returns despite ninety-seven percent of executives reporting AI agent deployment – not a vendor opacity issue. [6] The real visibility gap is fragmented accountability spreading cost ownership across four or more functions—information technology at twenty-five percent, risk at eighteen percent, cross-functional teams at seventeen percent, dedicated units at ten percent—each optimizing for their own budget rather than total system cost. [1] Shadow AI compounds this: fifty-seven percent hide AI use, forty-eight percent upload data to public systems, creating invisible cost centers workforce teams cannot see. [2] When teams cannot see spending, they assume the worst about priorities, creating engagement, retention, and innovation costs. Structural blindness exists, trust erodes.
The market consensus follows Zitron’s logic closely, though. Leadership teams demand detailed token pricing, usage dashboards, and infrastructure cost breakdowns. They assume that once opaque vendor numbers are transparent, the return on investment puzzle solves itself.
But the 2026 governance data contradicts this story. Accountability fragments across information technology, risk management, cross-functional teams, and dedicated AI governance teams. [1]
Fifty-eight percent of leaders believe their controls are keeping pace with AI adoption, yet only eighteen percent have active mitigation covering most or all identified risks. [1] The perception gap runs deeper than the fragmentation.
The real problem is organizational opacity rather than vendor opacity.
Technology opacity is the downstream effect of organizational opacity — the symptom, not the cause. Organizations hide structural costs through governance vacuums, decision delays, and coordination friction more effectively than any vendor hides token costs.
The organization sees governance where gaps actually exist.
On top, trust deficits emerge from organizational architecture that fragments visibility and erodes workforce confidence. Trust erosion multipliers create structural blindness that prevents senior leadership from seeing real costs, regardless of vendor transparency.
Organizations live structural blindness. They hide costs far more effectively than AI vendors.
AI Vendor Transparency is the Wrong Conversation
The chase for AI vendor transparency follows a familiar pattern. Leadership teams demand numbers clarity from vendors – external accountability.
But the same AI investment produces wildly different outcomes across organizations. One company surfaces costs and adjusts; another drowns in identical data. The difference is organizational structure, structural accountability, not vendor transparency.
AI agent deployments mirror the opacity they encounter. When organizations fragment accountability, hoard information, and obscure decision rights, AI agents adapt to the structural fog rather than creating it. They work within ambiguous boundaries because the organizational architecture offers no clear lines to align with.
Structural blindness begins with the wrong question. Leaders ask why vendors hide costs of AI deployment when they should ask what governance structures make their organizations vulnerable to hidden costs.
They chase external answers to internal architecture problems. The transparency gap lives in architecture, not pricing models.
This is an architectural failure, not a procurement gap.
Accountability Fragmentation Creates Systemic Gaps
No single function owns more than twenty-five percent of AI governance responsibility. Information technology holds twenty-five percent, risk management holds eighteen percent, cross-functional teams hold seventeen percent, and dedicated AI governance teams own just ten percent. [1] When cost ownership spreads across four or more functions, each unit optimizes for its own budget rather than total system cost.
Chief AI Officer adoption exploded from twenty-six percent to seventy-six percent in a single year [3], while dedicated AI governance teams own only ten percent of governance responsibility [4]. The title creates an illusion of authority while the actual decision rights remain scattered.
The organization sees governance where gaps actually exist. Decision latency accumulates while each function waits for another to move. This is structural blindness in its purest form.
The perception gap runs deeper than the fragmentation.
Fifty-eight percent of leaders believe their governance controls are keeping pace with AI adoption, yet only eighteen percent have active mitigation covering most or all identified risks. [1]
The transparency gap lives in architecture, not pricing models.
Structural blindness comes from gaps in governance, no AI vendor can cross.
Shadow AI Operates Beyond Organizational Sight Lines
Top-down mandates roll out without roadmaps — and often without the structured accountability or clear decision rights needed for execution. [5]
Leaders announce AI adoption programs while the operational infrastructure to support, track, or govern those programs remains undefined. The authority gap stays invisible to boards because the mandate itself creates an illusion of control. [5]
The result: Organizations cannot see their own deployments. Fifty-seven percent of employees hide AI use from their employers. Sixty-six percent do not verify outputs.
Forty-eight percent upload company data into public AI systems. [2]
Information latency compounds the problem. Frontline teams make choices without visibility into what other units have already deployed. The left hand builds what the right hand cannot see.
This is an architectural alignment failure, not a procurement gap.
Trust Erosion Multiplies the Blindness Effect
When teams cannot see how resources flow, they fill the void with assumption. They assume the worst about leadership priorities, budget allocations, and strategic intent.
Trust is the multiplier that determines whether any investment generates returns.
That assumption erodes trust at the exact moment the organization needs coordination most. The costs surface slowly: engagement drops, retention weakens, and innovation stalls before anyone traces them back to opacity.
Trust is the multiplier that determines whether any investment generates returns. When trust drops, the numerator of value collapses regardless of how robust the technology or transparent the vendor.
Teams work at cross-purposes because the architecture fragments how their decisions connect to the whole. No vendor dashboard can repair that fracture.
How to reverse structural blindness
Structural blindness resolves only when organizations redesign what they look at. AI vendor transparency addresses pricing models.
Organizational architecture addresses who owns decisions, who sees costs, and who coordinates action. The redesign required is governance architecture, not vendor negotiation.
Leaders who shift their question from “Why do AI vendors hide costs?” to “What structures make us vulnerable to hidden costs?” begin to surface what the system conceals. They map fragmented accountability.
They clarify decision rights. They build the structural conditions for transparency rather than demanding it from external parties.
Structural blindness is reversible. The cure lives in organizational architecture, not vendor contracts.
Sources
- Optro: AI Governance Stats — 2026 — Primary source for accountability fragmentation (IT 25%, Risk 18%, Cross-functional 17%, Dedicated AI teams 10%) and perception gap (58% believe controls keep pace, 18% have active mitigation)
- Ian Xie via Ana Petras: KPMG–University of Melbourne Shadow AI Research — 2026 — 48,000+ respondents, 47 countries (57% hide AI use, 48% upload data to public systems)
- IBM via Optro: Chief AI Officer Adoption Data — 2026 — CAIO adoption exploded from 26% to 76% in a single year
- Optro: Dedicated AI Governance Team Ownership — 2026 — Only 10% of governance responsibility owned by dedicated teams
- Atif Rafiq/Pao: Leadership Vacuum Widens as Enterprises Mandate AI Without Roadmaps — 2026 — Top-down mandates without structured accountability or decision rights
- Writer: AI Adoption Survey 2026 — 2026 — 97% executives deploy AI agents, only 29% see significant returns
Please note: 51even is an AI-first organization. We embrace AI at every step of our value creation and build our processes with a deep integration of human-AI capability. Humans always have the last decision. But this text was heavily built with AI.