A new Protiviti survey of 852 C-suite executives reveals a stunning divide: CIOs and CTOs report 61% confidence that AI is driving business transformation, while CEOs and boards clock in at just 34%. That's not a rounding error. That's a 27-point chasm between the people building AI systems and the people signing the checks for them.
Three major surveys dropped in the last 30 days — Protiviti's Global Transformation Survey (852 C-suite executives, conducted with the University of Oxford), PwC's 29th Global CEO Survey (4,450 CEOs across 95 countries), and Deloitte's 2026 State of AI in the Enterprise. Together they paint a consistent picture: enterprise AI adoption is accelerating, but the leadership teams inside these organizations are not looking at the same scoreboard.
The question isn't just academic. When CEOs and CIOs disagree on whether AI is working, budgets get cut, projects stall, and the enterprises that could be capturing real value spend another year in pilot mode instead.
The Numbers Behind the Divide
Let's start with the Protiviti data because it's the most specific.
When asked about confidence in AI driving revenue growth, CIO/CTOs land at 61%. CEOs and boards land at 30%. That's not just a perception gap — it's a two-to-one difference in how the same technology is being evaluated inside the same organizations.
The PwC CEO survey reinforces the pessimism at the top. Of the 4,450 CEOs surveyed across 95 countries, 56% say they have not realized either revenue or cost benefits from AI. Only 30% have measured tangible revenue increases. Only 26% report lower costs. Just one in eight companies qualifies as what PwC calls the "vanguard" — organizations achieving both revenue gains and cost reductions through extensive AI deployment.
That is a brutal number. One in eight. After years of investment, headlines, and vendor promises, only 12.5% of enterprises can point to both top-line and bottom-line returns from AI.
And yet the Deloitte data tells a different story in the same market. Sixty-six percent of organizations report productivity and efficiency gains from AI. Fifty-three percent cite enhanced insights and decision-making. Forty percent report cost reductions. Worker access to AI tools rose 50% in 2025 alone, and the number of companies with at least 40% of their AI experiments running in production is on track to double within the next six months.
So which survey is right?
Both Are Right — And That's the Problem
The resolution to this apparent contradiction is straightforward once you understand what each survey is measuring.
Deloitte is measuring AI activity — adoption rates, usage breadth, experimentation volume. On those metrics, progress is real and accelerating. More employees are using AI tools than ever before. More experiments are reaching production.
PwC is measuring AI outcomes — specifically, revenue increases and cost reductions that hit the income statement. On those metrics, results are sparse and unevenly distributed.
CIOs are optimizing for adoption velocity. CEOs are optimizing for financial returns. When adoption doesn't convert to returns fast enough, you get the 27-point confidence gap Protiviti found.
This is the core misalignment. CIOs measure success when AI is deployed. CEOs measure success when AI changes the P&L. These are different things, often separated by 12 to 24 months of organizational change, workflow redesign, and measurement work — none of which happens automatically when you deploy a model.
Who's Right? The CIO Isn't Wrong, But the CEO Isn't Lying
In conversations with technology leaders at large enterprises, I keep hearing a version of the same frustration: "We're seeing real productivity gains, but leadership keeps asking where the revenue growth is."
That frustration is valid. The CIO isn't imagining the efficiency gains — 66% of organizations reporting productivity improvements in the Deloitte survey is a real signal. But there's a gap between "employees are more productive" and "the company is making more money." Most enterprises haven't done the work to connect those dots.
The classic example: a legal team uses AI to cut contract review time by 40%. That's a real, measurable productivity gain. But if the company doesn't redeploy those attorneys to higher-value work, reduce headcount, or accelerate deal cycles, the financial benefit doesn't materialize. The CIO sees a productivity win. The CFO sees the same legal budget with no improvement in throughput or cost.
This is what PwC calls "innovation theater" — AI activity that resembles transformation but produces no tangible value at the business level. Less than 10% of surveyed CEOs say their organizations have adopted five or more of the six practices PwC identifies as markers of real innovation maturity: structured experimentation, rapid customer testing, high risk tolerance for innovation, defined incubation processes, and active portfolio pruning.
The COO Surprise
One data point from the Protiviti survey deserves special attention: COOs are the most enthusiastic AI believers in the C-suite. Forty percent of COOs selected AI as the capability with the greatest potential to drive revenue growth — outpacing even CIOs on AI optimism.
This makes operational sense. COOs live in the world of process efficiency, supply chain optimization, and customer service throughput. These are exactly the domains where AI is delivering the most documented returns: NVIDIA's 2026 survey found that 99% of telecom respondents said AI had improved employee productivity, with a quarter describing the improvement as major or significant.
When AI improves a specific, measurable operational process — call deflection rates, logistics routing, document processing — the COO can see the financial return more directly than most. The problem is that this operational ROI often doesn't aggregate cleanly into the CEO-level revenue and cost metrics that drive overall confidence.
For enterprises navigating this divide, the COO is often the most productive ally for a CIO trying to build the business case for continued AI investment. Operational improvements translate to financial outcomes more visibly than abstract capability investments.
The agentic AI Timing Problem
Here's where the story gets more complicated — and more urgent.
Deloitte found that agentic AI (autonomous systems that act and decide with minimal human oversight) is poised for a sharp rise in enterprise use. Today, 23% of companies report at least moderate use of agentic AI. That figure is expected to grow substantially over the next two years.
The problem: only one in five companies currently has a mature governance model for these autonomous agents.
So just as enterprises are sorting out how to measure and demonstrate ROI from generative AI, the next wave — autonomous agents making decisions across customer interactions, financial processes, and supply chains — is arriving faster than governance frameworks can handle. This isn't a theoretical compliance concern. Agentic systems operating without adequate oversight represent direct operational risk at scale.
For CIOs, this creates a timing problem. The confidence gap with the CEO is already wide. Rolling out agentic AI before demonstrating returns from generative AI risks further eroding board-level trust — especially if an autonomous agent makes a costly mistake in a high-visibility process.
The smarter path: use the current wave of productivity wins to fund and justify the governance infrastructure needed for agentic AI. Don't deploy agents in production until you can show the board a clear oversight model. The ROI question is hard enough without adding "and also, our AI is making decisions without human review."
What the Alignment Advantage Actually Looks Like
The Protiviti survey found that organizations with strong executive consensus report significantly higher confidence in AI value realization. More specifically: confidence scores are below 20% in early-stage organizations and exceed 70% in organizations at advanced stages of transformation.
That's not a correlation you can ignore. Alignment isn't soft culture work — it's a leading indicator of whether your AI investments will produce measurable returns.
What does alignment actually look like in practice? Three things, based on the Protiviti findings:
Shared success metrics. Technology leaders and business leaders need to agree on what "AI working" means before deployment, not after. If CIOs measure success in model accuracy and adoption rates while CEOs measure in revenue per employee or margin improvement, you will always have a confidence gap. The fix is forcing that conversation early. What financial outcome are we targeting? What's the measurement period? What's the baseline?
Transparency on the value chain. If AI improves a process, what specific business outcome does that process improvement enable? This chain — AI capability → process improvement → financial outcome — needs to be documented and communicated consistently. CIOs often excel at the first link and skip the last two. CEOs need the full chain.
Investment sufficiency. PwC found that only 40% of CEOs say their level of AI investment is sufficient to meet their stated goals. This is a dangerous gap. Underinvesting in AI while expecting transformation-level returns is a setup for disappointment on both sides. If the board wants enterprise-scale AI outcomes, it needs to fund enterprise-scale AI infrastructure — including the talent, data governance, and change management that convert model capability into business results.
Practical Steps for CIOs Facing Skeptical Boards
If you're a CIO sitting across from a CEO who isn't convinced, these three moves matter most right now.
Translate productivity to dollars. Don't report "40% faster contract review." Report "our legal team processed 23% more contracts in Q1 without adding headcount, equivalent to approximately $X in avoided hiring cost." This is not spin — it's the translation layer your CEO needs to connect AI activity to financial reality.
Pick two or three flagship metrics and own them. Reporting 15 different AI impact metrics dilutes the story. Pick the ones that map most directly to what the CEO cares about — revenue per employee, gross margin, customer churn — and show consistent progress on those. Everything else is context.
Acknowledge the governance gap before the board asks. Proactively presenting your agentic AI governance framework — even if it's early-stage — demonstrates that you're thinking about risk, not just capability. Nothing erodes CEO confidence faster than a technology leader who seems unaware of the oversight questions that come with autonomous systems.
The Bottom Line for Business Leaders
If you're a CFO, CMO, or COO watching this debate from the sideline, the question is simpler: are the AI investments your CIO is proposing connected to outcomes you can measure in your domain?
If the answer is no, ask for the value chain before the budget gets approved. Not to be obstructionist — but because projects without clear business outcome linkage are exactly what PwC calls innovation theater. The best time to build that measurement framework is before deployment, not 18 months after when someone asks why the ROI didn't materialize.
If the answer is yes, push to accelerate. Deloitte's data shows that enterprises which reach 40%+ of AI experiments in production are on track to double that threshold within months. The gap between companies in the "vanguard" and those still in the pilot phase is widening. The confidence gap between CEOs and CIOs is real — but it's also a solvable problem, and the enterprises that solve it faster will compound the advantage.
The CIO isn't wrong that AI is working. The CEO isn't wrong that the returns are harder to find than the headlines suggest. Closing that gap is the most important AI work happening in boardrooms right now — and no model can do it for you.
Sources:
- Protiviti Global Transformation Survey 2026 — 852 C-suite executives, conducted in partnership with University of Oxford
- PwC 29th Global CEO Survey — 4,450 CEOs across 95 countries
- Deloitte 2026 State of AI in the Enterprise
- NVIDIA 2026 Enterprise AI Survey — 3,200+ responses across financial services, retail, healthcare, telecom, and manufacturing
Follow Rajesh on LinkedIn or X for daily enterprise AI insights.
