PwC AI Research: The 56% Problem — Most CEOs See No Return from AI

See also (wiki): workflow-redesign, agentic-ai-governance, training-architecture, productivity-rcts, firm-size-ai-outcomes

Executive Summary

• PwC’s 29th Global CEO Survey (n=4,454, Jan 2026) delivers the sharpest number in the enterprise AI debate: 56% of CEOs report zero financial return from AI investments — neither revenue gains nor cost savings. Only 12% have achieved both.
• The AI Jobs Barometer (nearly 1 billion job ads, Jun 2025) shows the opposite side of the coin: industries most exposed to AI see 3x higher revenue-per-employee growth and a 56% wage premium for AI-skilled workers. The value is real — but it accrues unevenly.
• PwC’s own chairman, Mohamed Kande, names the root cause: “People forgot that the adoption of technology, you have to go to the basics.” Companies with strong AI foundations are 3x more likely to report meaningful financial returns.
• The 20/80 rule surfaces again: PwC estimates technology delivers only 20% of an initiative’s value — the other 80% comes from redesigning work around what AI makes possible.

The CEO Reality Check

PwC’s 29th Global CEO Survey, fielded September–November 2025 across 4,454 CEOs in 95 countries, paints a sobering picture of where enterprise AI actually stands.

The headline: more than half of CEOs (56%) say their company has realized neither higher revenues nor lower costs from AI. Only 30% report revenue increases. Only 26% report cost decreases. And only 12% — the “vanguard” — report both. Meanwhile, 22% say AI has actually increased their costs.

This is not a technology failure. It is a deployment failure. The vanguard companies that have achieved both revenue and cost benefits share a common profile: they deploy AI more extensively across business areas (44% have applied AI to products and services vs. 17% for others), and they have invested in foundations — technology environments that enable integration, defined AI road maps, formalized responsible AI processes, and organizational cultures that enable adoption.

The survey also reveals how shallow most deployments remain. Fewer than a quarter of CEOs say AI is applied to a “large or very large extent” in any major business area: demand generation (22%), support services (20%), products and experiences (19%), strategic direction (15%), demand fulfillment (13%). Only 14% of workers use generative AI daily.

CEO confidence in near-term growth has hit a five-year low (30%, down from 56% in 2022), and their top concern is whether they are transforming fast enough to keep up with technology.

The Labor Market Signal

PwC’s 2025 Global AI Jobs Barometer offers a counterpoint to the CEO gloom — but one that reinforces the same asymmetry. Based on nearly 1 billion job ads and thousands of company financial reports across six continents, the Barometer finds that AI is creating measurable value at the industry level, even as most individual companies fail to capture it.

Industries in the most AI-exposed quartile have achieved 27% cumulative productivity growth (revenue per employee) from 2018–2024, compared to 8.5% for the least exposed quartile. That gap opened sharply after 2022 — suggesting that the ChatGPT wave did translate into real economic activity, but only in sectors positioned to absorb AI into their workflows.

The wage data tells the same story. Workers with AI skills now command a 56% wage premium, up from 25% the prior year. Wages are growing 2x faster in AI-exposed industries than in AI-insulated ones. AI-skill job postings grew 7.5% year-over-year even as total job postings fell 11.3%. Skills required in AI-exposed occupations are changing 66% faster than in non-exposed roles.

The Barometer’s most important finding may be its rebuttal of the automation-as-job-killer narrative: job numbers are growing in virtually every AI-exposed occupation, including highly automatable ones. Workers in automatable roles are not being replaced — they are being redirected from routine tasks to higher-complexity work.

Responsible AI as Business Enabler

PwC’s 2025 Responsible AI Survey (n=310 US business leaders, October 2025) finds 61% of organizations at the “strategic” (28%) or “embedded” (33%) stage of Responsible AI integration — actively built into operations and decision-making. Among respondents, 60% say Responsible AI boosts ROI and efficiency, and 55% report improved customer experience. Organizations at the strategic stage are 1.5–2x more likely to describe their Responsible AI capabilities as effective compared with those still in the training stage.

This aligns with the CEO Survey finding: the 12% vanguard has formalized responsible AI processes. Governance is not overhead — it is a performance differentiator.

The 20/80 Rule and Agentic AI

PwC’s 2026 AI Business Predictions quantify a principle that runs through every serious enterprise AI study: technology delivers only about 20% of an initiative’s value. The other 80% comes from redesigning work so that agents handle routine tasks and people focus on what drives impact.

PwC expects successful agentic deployments to coalesce around centralized “AI studios” — hubs that combine reusable technology components, use-case assessment frameworks, sandboxes for testing, deployment protocols, and skilled operators. The alternative — decentralized, siloed agent experiments — produces the same outcome the CEO Survey documents: cost without return.

PwC estimates AI agents can now perform roughly half of tasks that people currently do. But agentic workflows are spreading faster than governance models can address their risks — a gap that the Responsible AI data suggests only a minority of companies are closing.

Key Data Points

What This Means for Your Organization

The 56% number should not be read as evidence that AI does not work. It should be read as evidence that most companies have not done the prerequisite work that makes AI work. PwC’s own data shows that the 12% achieving results have invested in foundations — technology integration, road maps, responsible AI processes, and culture — before scaling deployment. The labor market data confirms the payoff is real for those who get there.

For a mid-market company with 200–2,000 employees, three implications stand out. First, the 20/80 rule means that selecting an AI tool is the smallest decision in the chain. Redesigning workflows, retraining staff, and building governance consume 80% of the effort and produce 80% of the value. Second, the 56% wage premium and 66% skill-change acceleration mean that workforce investment is a competitive necessity, not a perk — companies that delay will pay more to hire the same talent later. Third, the Responsible AI data suggests that governance maturity correlates with financial performance, not just compliance — making the business case for governance frameworks before regulators make it for them.

If these numbers raise questions about where your organization stands relative to the 12% vanguard, that is a conversation worth having — brandon@brandonsneider.com.

Sources

Source credibility note: PwC has direct commercial interest in AI strategy, governance, and transformation engagements. Its research findings — including the 56% “zero return” statistic — are credible because they are self-damning rather than promotional: a consulting firm publishing that most AI investments fail has commercial incentive to be accurate (it validates their services). Treat quantitative claims as directionally reliable but cross-reference key figures against independent sources (McKinsey, BCG, Stanford HAI). Self-reported maturity scores (Responsible AI Survey) are subject to response bias.

1. PwC, “Leading Through Uncertainty in the Age of AI: PwC’s 29th Global CEO Survey” (Jan 2026). n=4,454 CEOs, 95 countries. PDF: https://www.pwc.com/gx/en/ceo-survey/2026/pwc-ceo-survey-2026.pdf. Credibility: HIGH — largest annual CEO survey; independent methodology; adjusted for sector, size, geography. Note: PwC has commercial interest in AI consulting and transformation engagements; 56% “zero return” finding is directionally credible as it is self-damning and corroborated by McKinsey (6% high performers, Nov 2025) and BCG (5% substantial gains, 2025).

2. PwC, “The Fearless Future: 2025 Global AI Jobs Barometer” (Jun 2025). ~1 billion job ads, 6 continents, data through end 2024. PDF: https://www.pwc.com/gx/en/issues/artificial-intelligence/job-barometer/2025/report.pdf. Credibility: HIGH — massive dataset; uses established Felten et al. AI Exposure index and IMF methodology. Caveat: measures AI exposure (ability to use AI) as proxy for uptake; actual adoption may lag.

3. PwC, “2025 Responsible AI Survey: From Policy to Practice” (Oct 2025). n=310 US business leaders, director+. URL: https://www.pwc.com/us/en/tech-effect/ai-analytics/responsible-ai-survey.html. Credibility: MEDIUM — smaller sample, US-only, self-reported maturity stages.

4. PwC, “2026 AI Business Predictions.” URL: https://www.pwc.com/us/en/tech-effect/ai-analytics/ai-predictions.html. Credibility: MEDIUM — opinion/forecast piece, not primary research; useful for framing.

5. Fortune, “56% of companies getting nothing out of AI, PwC research says” (Jan 19, 2026). URL: https://fortune.com/2026/01/19/pwc-global-chairman-mohamed-kande-ai-nothing-basics-29th-ceo-survey-davos-world-economic-forum/. Credibility: HIGH — direct quotes from PwC Global Chairman.

Brandon Sneider | brandon@brandonsneider.com April 2026