EY Answers The AI-ROI Question With An Org Chart: One Office, One Budget

EY is creating a new corporate function, the AI Value Realization Office, to centralize the firm's AI spending and establish whether the technology is producing measurable impact ^[1]. Dan Diasio, EY's global consulting AI leader, told Business Insider the office should be fully operational within a couple of months ^[2]. The interesting part is not the name but the mechanism: this is a budget move dressed as a governance move, and budget moves have consequences that pilots do not.

The office's remit, per Diasio, is to govern AI spending, drive returns, monitor usage, decide which initiatives get scaled, and oversee how AI reshapes jobs ^[3]. His argument for pulling that authority out of the functions is structural. Corporate budgets are usually allocated along departmental lines, while AI spending and its payoff cut across IT, finance, sales, HR, and operations ^[4]. "If you fund by department, you end up addressing a bunch of use cases inside of each of the functions. And what that often means is that you're leaving a lot of value on the table," he said ^[5].

The number underwriting that claim is EY's own. Diasio cited EY-Parthenon research finding that 75% of the potential enterprise value from AI comes from horizontal value streams spanning multiple functions, against 25% from projects contained within a single function ^[6] - a three-to-one split ^[7]. Readers should note the source: the research belongs to the firm selling the reorganization, and the underlying methodology is not disclosed in the reporting.

The design also concentrates power in a way worth naming. The same office that decides where funding flows ^[8] will assess whether initiatives are changing business performance rather than running a dollar-for-dollar calculation ^[9], and will decide which ones scale ^[3]. A body that allocates the money and grades the outcome is a familiar governance problem, whatever the technology.

The precedent EY reaches for is real enough. Modern HR departments took shape in the Great Depression as firms looked for systematic ways to hire and fire ^[10], and treasury functions emerged in the 1970s when floating exchange rates turned currency movements into a material risk ^[11]. Both were responses to a cost or risk that no existing department owned ^[12].

The cost pressure here is documented. EY said in 2023 that $1.4 billion went toward building the foundation of its EY.ai platform ^[13], and changes to AI providers' pricing models have pushed professional services firms to use the technology efficiently ^[14]. EY has an "invisible" router behind some of its specialized tools that steers employees to the best model for a task, which it credits with cutting token consumption by 60% since April alongside other governance measures ^[15]. That is a cost result, not a value result, and it is the easier of the two to prove.

Clients are feeling the same bill. In EY's US AI Pulse survey released in July and fielded in April and May, 98% of 534 senior decision-makers said token spending had made them reconsider their approach ^[16]. So far, according to Diasio, companies have funded AI mostly by tightening budgets across the board and reallocating ^[17], and most clients are not yet setting up dedicated AI offices ^[18], though he expects the role to become a function over time ^[19].

Watch three things: whether EY publishes what the office defunded, not just what it approved; whether the 60% token reduction holds as usage grows; and whether clients buy the structure, given that EY pitches itself as "client zero" and then sells the advice ^[20].