The consulting firms that sold everyone on AI transformation just admitted they need a new department to make sure their own AI spending isn't lighting money on fire.
The Summary
- EY is launching an "AI Value Realization Office" within months to centralize AI governance, monitor spending, and ensure measurable returns across the firm
- EY-Parthenon research shows 75% of AI's enterprise value comes from horizontal initiatives that span departments, not siloed function-specific projects
- The office's mandate: decide which AI projects scale, track how AI reshapes jobs, and redirect funding to high-impact opportunities rather than departmental wish lists
The Signal
Big Four firms don't create new organizational functions lightly. When EY announces it needs a dedicated office just to make sure AI spending pays off, that's a signal about where the enterprise AI market actually is in 2026: past the proof-of-concept parade, deep into the "wait, what are we getting for this" phase.
Dan Diasio, EY's global consulting AI leader, frames the problem clearly: traditional budgeting by department doesn't work for AI. When IT wants agents for infrastructure monitoring, finance wants them for reconciliation, sales wants them for lead qualification, and HR wants them for recruiting, you end up with scattered investments and duplicated infrastructure. Each department solves its narrow problem while missing the compound value that comes from agents that work across functions.
"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."
The 75-25 split EY found between horizontal and vertical AI value is the data point that matters here. Three-quarters of the return comes from initiatives that cut across silos. That means most companies are currently optimizing for the 25% because that's what fits into existing budget structures and approval processes. The AI Value Realization Office is EY's answer to that misallocation problem.
What this office actually does reveals where AI deployment complexity lives now:
- Govern spending across functions instead of within them
- Monitor usage patterns to see what's working versus what's getting expensed and ignored
- Decide which pilots scale and which die
- Track how AI changes job structures and headcount needs
- Redirect funding based on cross-functional impact, not departmental squeakiness
This isn't about technical governance or model safety. It's about capital allocation and organizational redesign. The implication is that AI at scale requires rethinking how companies budget, measure success, and structure work itself.
The Implication
If EY needs a dedicated office to manage its own AI complexity, your organization probably does too, or will soon. The firms still running AI as an IT initiative or innovation lab project are setting up to waste money at scale.
Watch for this pattern to spread. Not just consulting firms, but any enterprise spending eight figures on AI will need someone whose job is asking "what are we actually getting" across the whole organization. That role doesn't exist in most companies yet. The people who can do it well, who understand both agent capabilities and organizational economics, are about to become very expensive.