A mid-size manufacturer completes a company-wide AI literacy rollout. Two hundred employees earn certifications. The AI Center of Excellence publishes a framework. The governance committee holds its quarterly meeting. Twelve months later, the CFO asks a simple question: where is the productivity gain? No one can answer.
This is the second of five constraints we see between AI investment and earnings. We call it Competence Theater: the accumulation of credentials, frameworks, and governance structures that signal readiness without producing it.
The pattern
This is not an isolated story. It is the dominant pattern in enterprise AI adoption right now. Organizations are investing heavily in the signals of AI readiness while the underlying operations remain unchanged. The credentials accumulate. The outcomes do not.
BCG’s 2026 CEO survey found that 72 to 75 percent of organizations report positive returns on their gen AI investments. When examined carefully, those returns consist of productivity gains and cost avoidance, not margin expansion. McKinsey’s State of AI (November 2025) provides the earnings picture directly: 78 percent of companies have deployed gen AI, and 80 percent report no measurable earnings impact from it. The ADR Center’s 2025 survey of enterprise AI governance found that 87 percent of organizations have formal governance principles and policy structures in place, but only 22 percent say those structures work effectively in practice; most describe governance as possessing good structures but inconsistent execution. Coverage drops sharply across the AI lifecycle: 72 percent apply governance at development, 44 percent during post-deployment monitoring, and 4 percent at system retirement. The infrastructure of readiness is in place. The discipline of outcomes is not.
Certifications validate awareness, not application. When organizations stop at awareness, they have not built capability. They have built the appearance of it.
What it costs
Training investment is a particularly clear illustration. Corporate AI literacy spending increased substantially in 2024 and 2025, driven by genuine recognition that workforce capability matters. But workforce surveys consistently show that training completion and behavioral change are not the same thing. Employees finish modules and return to existing workflows that were never rebuilt to use what they learned: 84 percent of companies have not redesigned a single job to fit AI, so the training lands on processes that cannot absorb it (Deloitte, 2026).
The confidence gap is its own tell. In one 2026 readiness study, 87 percent of organizations reported having the infrastructure they needed while 42 percent named that same infrastructure as their single biggest obstacle, a contradiction that only makes sense if the readiness is declared rather than demonstrated (Drexel LeBow and Precisely, 2026).
The pattern also creates a governance liability. Committees that meet but do not decide, frameworks that exist but are not enforced, checklists completed but not monitored: these structures create the documentation of diligence without the substance of control. As regulatory scrutiny of AI systems increases across jurisdictions, organizations that have invested in the form of governance without the function face real exposure.
What distinguishes organizations that escape it
What distinguishes organizations that escape this pattern is not more training or better frameworks. It is outcome discipline applied from the start. The firms that convert AI investment into earnings share a specific orientation: they define measurable success criteria before any capability-building begins, they tie every training and governance activity to a specific deployment target, and they reduce oversight dependency over time by design rather than by accident.
The question is never “did employees complete the curriculum?” It is “what changed in how work gets done, and what did that change produce?” That distinction sounds simple. In practice, it requires a different kind of program architecture than most organizations have built.
The five constraints that prevent AI investment from converting to earnings (Strategy Without Validation, Competence Theater, The Governance Gap, Invisible Spend, and The Accountability Vacuum) are structural, not technical. Competence theater is the one most organizations are least willing to name, because the investments involved feel responsible. The problem is not that the investments are wrong. The problem is that without outcome discipline they are incomplete, and incomplete looks identical to effective until the CFO asks the question that no one can answer.
This is the second in a five-part series on the constraints between AI investment and earnings. We named these constraints using the same diagnostic discipline we run on our own practice before we bring it to a client. The two-week Readiness Assessment names which are operating in your organization, verifiable against your own data, before Q3 planning.