75% of Executives Admit Their AI Strategy Is 'Just for Show'

There is a quiet admission buried in Writer's 2026 enterprise AI report that should reframe how every board talks about artificial intelligence: 75% of executives concede that their company's AI strategy is "more for show" than a genuine guide to how the organization actually operates. Read that again. Three out of four leaders looking at their own AI plan describe it as theater. Not a failure of execution they are working to fix — a performance they are knowingly staging. The deck exists, the press release went out, the steering committee meets, and almost none of it touches the way work gets done.

What "Performative AI Strategy" Actually Looks Like

Performative AI strategy is not the absence of activity. It is the presence of activity that produces no change in how value is created. It is busy, well-funded, and frequently celebrated internally — which is precisely why it survives so long before anyone notices the P&L never moved. If you have sat through the last two years of enterprise AI initiatives, you have seen its signature moves.

There is the AI steering committee that meets monthly, produces a roadmap slide, and ships nothing into a customer-facing or revenue-bearing workflow. There is the all-hands demo, where a polished prototype draws applause and is then never seen again because no one owns the path from demo to production. There is tool sprawl: fourteen AI subscriptions across departments, a six-figure annual spend, and not one workflow that was redesigned around the tools. And there is the press release announcing an "AI-first transformation" that, when you trace it to the income statement, has no associated line — no cost taken out, no revenue added, no cycle time reduced.

The tell in every case is the same: the strategy describes an aspiration, not an operating change. A real operating change is falsifiable. You can point to the process that used to take five steps and now takes two, the support queue that used to need twelve people and now needs eight, the proposal that used to take three days and now takes three hours. A performative strategy describes a posture — "we are becoming an AI-native company" — that no one can disprove because it never specified what would be different on Tuesday.

The Numbers Behind the Theater

The "just for show" admission would be easier to dismiss as a rhetorical flourish if the surrounding data did not corroborate it so precisely. Writer's 2026 report also found that 48% of executives now describe AI adoption as a massive disappointment — up from 34% the year before. That is not a small drift. In a single cycle, the share of leaders who feel let down by AI grew by fourteen percentage points, a clear sign that the gap between the strategy on the slide and the outcomes in the business has become impossible to paper over internally.

The disappointment is not irrational, because the return numbers are genuinely thin. According to the same report, only 29% of organizations report significant ROI from generative AI, and just 23% report it from AI agents. Put plainly: more than seven in ten companies cannot point to meaningful financial return from the generative AI they have deployed, and more than three-quarters cannot from the agents they have piloted. After two years of aggressive investment, that is a striking shortfall — and it is the financial reality the performative strategy was built to obscure.

And yet the potential is not in doubt. The same body of research that documents the disappointment also finds that AI super-users — the individuals who have genuinely rewired how they work around the tools — deliver roughly 5x productivity gains. The capability is real and the leverage is enormous. The failure is not in the technology. It is in the distance between what a handful of power users achieve and what the organization, governed by a for-show strategy, manages to capture. That distance is the whole story, and we unpack the mechanics of it in our companion piece on the rise of AI productivity theater.

Why Performative AI Strategy Happens

Theater is not a character flaw. It is the predictable output of a specific set of pressures acting on leaders who have no clean way to satisfy them honestly. Understanding the mechanism is the first step to dismantling it.

Board and FOMO Pressure

The first force is external. Boards, investors, and analysts have spent two years asking every leadership team the same question: "What is your AI strategy?" The cost of answering "we are being deliberate and we have not found a high-ROI use case yet" is socially and reputationally high, even when it is the most honest answer in the room. The cost of producing a confident deck with a roadmap, a committee, and a vendor logo wall is low. So leaders optimize for the cheap answer. The strategy is built to be presented, not to be operated, because the immediate audience is the board, not the business.

No Accountable Owner

The second force is internal and structural. Most AI strategies have a sponsor but no owner — someone who blesses the initiative in the abstract but no one whose compensation, headcount, or quarterly objectives depend on a specific, measurable outcome. When everyone is responsible, no one is accountable. The steering committee diffuses ownership across a dozen calendars, and a diffuse owner produces a diffuse result: a strategy that lives in slides because no single person is on the hook to make it live in production.

Mistaking Adoption for Transformation

The third force is conceptual, and it is the most damaging. Many organizations measure their AI progress in seats, licenses, and logins — how many employees have access, how many prompts were run, how many tools were procured. These are adoption metrics, and adoption is not transformation. Giving every employee a chat assistant changes the average workday by a few minutes; it does not change the unit economics of the business. Transformation requires redesigning a workflow so the work itself is different, and that is hard, slow, and unglamorous. Adoption is easy to report. So adoption gets reported, the dashboard turns green, and the for-show strategy is declared a success while the P&L sits untouched.

Theater vs. a Real Operating Change

The clearest way to diagnose your own situation is to put the two modes side by side. They use the same vocabulary, attend the same conferences, and procure from the same vendors. What separates them is whether anything downstream of the strategy is measurably different.

Notice that the operating-change column is narrower and less impressive-sounding. It does not promise enterprise-wide transformation by next quarter. It promises one workflow, genuinely rebuilt, with a number attached. That modesty is the point. Real change compounds from concrete wins; theater inflates from abstract ambition. A leadership team that ships one P&L-positive workflow has more credibility — and a clearer path to the second and third — than one with a fifty-page strategy and nothing in production.

A Tale of Two Companies

Picture two companies in the same industry, both of which announced an AI strategy in the same quarter. The first stood up a cross-functional AI council, procured a suite of tools, rolled access out to every employee, and reported in the next board meeting that adoption had reached 80% — prompts run were up, seats were active, the dashboard was green. Eighteen months later, its margins are unchanged. The tools are used the way a faster word processor is used: real but marginal time savings, distributed thinly across thousands of slightly easier days, none of it visible in the financials.

The second company did something far less impressive to announce. It picked its support organization, named a director who already owned the support cost line, and gave her one objective: cut the cost per resolved ticket by rebuilding the resolution workflow around AI. She baselined the current cost, redesigned the process so an agent handled first-pass triage and drafting under a review harness, and measured the result against the baseline. There was no press release. But eighteen months later, the cost per resolved ticket is down by a third, the change is a defensible line on the income statement, and that one proven workflow is now the template being applied to billing and onboarding. The second company is in the 29%. The first is in the 48% calling AI a disappointment. They spent similar money. They differed only in whether one person owned one number on one workflow.

The Delegation Problem Underneath the Theater

One reason workflows never get rebuilt is that organizations have not learned to hand real work to AI systems with appropriate guardrails. Performative strategy and shallow delegation are two symptoms of the same root cause: a refusal to commit AI to anything consequential. As we argued in our analysis of the delegation gap in the orchestration era, teams use AI for a large share of their work but fully delegate only a thin slice of it — and that reluctance keeps AI parked in the role of a faster typewriter rather than a system that owns an outcome.

A workflow that has been genuinely transformed has crossed the delegation threshold: the AI does not merely assist a human step, it owns a bounded stage of the process under an evaluation harness that catches its failures. That is what allows headcount to be redeployed, cycle time to drop, and the change to land on the P&L. Without that delegation, the best you get is a slightly more productive workforce running the same process — which is exactly the 5x-individual, 29%-organizational gap the data describes.

How to Convert Performance Into Measured P&L

The exit from theater is not a bigger strategy. It is a smaller, harder, more accountable one. The leaders who are in the 29% that see significant ROI did not out-strategize the other 71%. They out-executed them on a narrow front. Here is the pattern that separates the two.

The discipline that makes this work is refusing to broaden scope until the first workflow has paid for itself. Performative strategy is broad and shallow by design — it has to cover the whole enterprise to look impressive on a slide. A real operating change is narrow and deep: one workflow, fully redesigned, measured against its own baseline, owned by a named person. Once that workflow is demonstrably P&L-positive, it becomes the template and the credibility for the next one. The strategy stops being a performance for the board and starts being an operating manual for the business.

What Boards Should Ask Instead

Part of the fix lives above the operators. If boards keep asking "what is your AI strategy?", they will keep getting theater, because that question rewards posture. The more useful question is "which workflow did AI change this quarter, what was the before-and-after, and who owns the number?" That question cannot be answered with a deck. It forces the conversation onto outcomes, and it makes the for-show strategy visibly hollow the moment it is asked. Boards that change their question change the incentives, and incentives are what produced the theater in the first place.

Conclusion: From the Slide to the Income Statement

The most honest thing about Writer's finding is that the executives said it themselves. Three-quarters of them looked at their own AI strategy and called it a performance. That candor is an opportunity, not an indictment. It means the gap between ambition and outcome is now visible at the top of the organization, and visibility is the precondition for change. The 48% who call AI a disappointment are not wrong about their results; they are wrong about the cause. The disappointment is not the technology underdelivering. It is a for-show strategy delivering exactly what a for-show strategy is built to deliver: a presentation.

The 29% who capture real ROI prove the alternative is available. The leverage that produces 5x gains for individual super-users is the same leverage that, when encoded into a redesigned workflow under an accountable owner, moves the income statement. The difference between the theater and the result is not budget, vision, or vendor selection. It is whether one person is on the hook for one number on one workflow — and whether the organization has the discipline to measure the outcome instead of announcing the intent.