EY Creates AI Value Realization Office to Make Its $1.4bn AI Investment Pay

EY's New AI Value Realization Office

Ernst & Young is creating a dedicated unit called the AI Value Realization Office, designed to concentrate authority over AI spending, usage and returns in one place. Dan Diasio, who leads AI within EY's consulting business, told Business Insider that the office should be fully operational within several months.

The new office will decide where AI money goes, oversee which initiatives get scaled, monitor how the technology is used and track how it changes work across the firm. That mandate extends across IT, finance, sales, human resources and operations — areas EY says currently fund AI in fragmented ways.

EY's internal research from strategy arm EY-Parthenon found that 75 percent of AI's potential enterprise value sits in horizontal value streams spanning multiple functions, while only 25 percent comes from projects confined to a single department. Diasio said the office will push funding toward the largest opportunities rather than judging initiatives purely on cost-benefit arithmetic.

The move comes as EY and its Big Four peers position themselves as "customer zero" for AI. EY previously disclosed a US$1.4 billion investment in the foundations of its EY.ai platform, and the firm has already cut token consumption by 60 percent since April using an internal router that directs employees to the most suitable AI model.

Why EY Is Pulling AI Spending Out of Departmental Silos

The lesson behind EY's structure: horizontal value beats departmental budgets

Diasio's core argument is that corporate budgeting works against AI's economics. When functions fund their own use cases, they optimize locally; EY wants a central office that can see cross-function value and move resources toward the biggest opportunities. The 75/25 split from EY-Parthenon is the analytical backbone. It is EY's own research, so the precise ratio is self-serving, but the underlying logic matches a broader enterprise shift toward cross-functional AI ownership.

EY explicitly frames the office as the modern equivalent of HR departments during the Great Depression or treasury units in the 1970s: a new corporate function created when a new kind of cost or risk required coordinated management.

Token pricing turns AI from experiment to cost problem

The immediate trigger is cost. Diasio says AI providers are shifting away from subsidized pricing, leaving enterprises with the real bill for tokens. EY's US survey of 534 executives found 98 percent said token spending made them rethink their approach. EY's own 60 percent reduction in token consumption since April, achieved with an internal routing layer, is the proof point it can show clients.

What EY gains from being "customer zero"

By centralizing its own AI governance, EY creates a packaged case study for consulting clients who are now asking how to organize AI scaling. Most clients have not yet built their own AI offices, according to Diasio, which leaves room for EY to sell the model. Competitors in the Big Four face the same pricing pressure, but the first mover with a named internal unit can turn internal cost control into a revenue story.

What Enterprise AI Buyers Should Take From EY's Move

• Review whether AI spend is trapped in departmental budgets. EY's rationale rests on its EY-Parthenon finding that 75 percent of potential AI value spans functions. If your own AI projects are funded locally, map the use cases that cut across IT, finance, sales, HR and operations before increasing budgets.
• Treat token consumption as a board-level cost, not an engineering detail. EY reduced token consumption by 60 percent after April by routing employees to the most suitable model. Ask your AI team whether model routing, prompt efficiency or usage tiering can deliver a comparable reduction without sacrificing output quality.
• Prepare for the end of subsidized AI pricing. Ninety-eight percent of the 534 executives in EY's April–May survey said token costs made them rethink their approach. If your pilots were built on artificially cheap models, re-price the business case now before scaling.
• If you buy AI consulting, push for proof of cost control. EY will likely market the AI Value Realization Office as a blueprint; clients can ask for concrete metrics — not just framework slides — such as token-per-task reductions and cross-functional value attribution.