Deloitte spent six weeks last fall, August through September 2025, surveying 3,235 board members, C-suite leaders, and VP and director level executives across 24 countries. Half of them ran IT. Half ran the actual business.
The headline number is easy to like. 66% say AI is already delivering productivity or efficiency gains, the single most cited benefit in the whole report. Thats the number that shows up in the press release.
Heres the number that doesnt. Only 34% of those same companies say theyre pursuing real business transformation with AI. The other two thirds are optimizing what already exists. Faster drafts. Shorter meetings. Cleaner reports. Not new products, not new revenue lines, not a changed business model.
Same 3,235 executives. Same survey. Two numbers that dont agree with each other. Most companies arent lying about either one. Theyre just measuring two different things and calling both of them “AI progress.”
What “productivity gains” actually means in the data
66% report productivity or efficiency gains. 53% report better insights and decision-making. 40% got cost reductions. 38% say customer relationships improved. Those are all real. Theyre also all efficiency metrics, not growth metrics.
Only 20% of companies report AI actually increasing revenue. Meanwhile 74% say they aspire to revenue growth from AI eventually. Aspiration and result are two very different columns on this survey, and right now theres a 54-point gap between them. Nobody puts that gap on the slide that goes to the board.
“Only 34% of companies say theyre pursuing real business transformation with AI. The other two thirds are optimizing what already exists.”
The transformation gap, and why strategy looks confident anyway
42% of executives believe their strategy is highly prepared for AI adoption. Fewer say the same about infrastructure, data, risk management, or talent. Strategy is always the part that looks finished on a slide. Its the part underneath, the actual plumbing, where the confidence drops off.
Thats the honest read on the 34% number. Companies arent lacking ambition. Theyre lacking the infrastructure and governance to move past pilots into something that changes how the business actually makes money. A strategy document cant fix a data pipeline, and it cant assign accountability when a model gets something wrong in production.
Governance is the part nobody wants to talk about
Only 20% of companies in the survey have a mature governance model for autonomous AI agents. Agentic AI use is about to increase sharply over the next two years, according to the same report. Most companies are about to hand more autonomy to systems they havent finished governing. Thats the order most rollouts get backwards: give the agent more scope first, write the accountability rules later, if ever.
Thats not a hypothetical risk. Its a documented, measured gap, in a survey of 3,235 senior leaders who presumably know their own companies better than anyone. Figuring out where your own organization actually sits on that gap, not where the strategy deck says it sits, which is exactly what an AI Assessment for companies is built to surface, is the difference between the 66% and the 34%.
The skills problem sitting underneath both numbers
Deloitte names insufficient worker skills as the single biggest barrier to real AI integration. 53% of companies are prioritizing workforce education. 48% are running upskilling programs. 36% are hiring specialized AI talent. Only 33% are actually redesigning career paths around it.
Thats the pattern across HR, Compliance, and Legal at almost every company we talk to. Everyone trains people on the tool. Almost nobody rebuilds the job around it. Career paths dont move because nobody owns moving them, so the org chart stays exactly as it was before AI showed up, just with a faster version of the old job underneath it.
The 66% is real. The 34% is what happens after the pilot budget runs out and someone has to redesign the actual workflow.
Frequently Asked Questions
What does this mean for a Chief AI Officer?
It means the productivity number wont protect you at the next board meeting. A Chief AI Officer needs to be able to show which 34% bucket the company is actually in, transformation or optimization, with evidence, not a strategy deck.
Why did only 20% of companies report a revenue increase from AI?
Most current AI use sits in efficiency work: drafting, summarizing, analysis. Those tasks cut cost and time but dont on their own create new revenue. Revenue growth requires new products or business models, which Deloitte found only 34% of companies are actually building toward.
How does an AI Assessment for companies relate to this Deloitte data?
An AI Assessment for companies, like the ones Silicon Valley Certification Hub runs, is built to find exactly the gap Deloitte measured at the industry level: where a company’s AI use is real efficiency versus where it’s still just a pilot with a good story attached.
Whats the actual risk in the 20% governance number?
Agentic AI use is set to increase sharply over the next two years, per Deloitte, while only 20% of companies have mature governance for autonomous agents today. Companies are about to give more independence to systems with less oversight, which is exactly backwards.
What should executives actually do with this data now?
Stop reporting the 66% number on its own. Ask which bucket your company is really in, the 34% building transformation or the 66% optimizing existing work, and get an honest answer before the next planning cycle, not after.
Want to know how this applies to your company?
At Silicon Valley Certification Hub, we help you align AI + Strategy. Our team works directly with your directors and teams to assess AI readiness, identify gaps, and build a clear path forward, tailored to your business context.
Book a time with our CEO, Alejandro Cuauhtemoc-Mejia
Silicon Valley Certification Hub
3000 El Camino Real, Building 4, Palo Alto, CA
0 Comments