EY Launches Value Realization Office to Prove AI Investment Returns

The move is a quiet admission that buying AI is the easy part; proving it works is where most enterprises are getting stuck.

A spending split that is getting harder to ignore

New data from Ramp's August 2026 AI Index, authored by the company's lead economist Ara Kharazian, illustrates just how uneven corporate AI investment has become. The top 1% of US businesses on Ramp's platform are spending a median of $7,400 per employee per month on AI tools and services, based on July 2026 transactions. That figure has more than tripled since early 2024, when the same cohort was spending under $1,000 per employee monthly.

Step down the ladder, and the gap widens sharply. The top 10% of companies spent $650 per employee per month. The median US firm spent just $11.95.

Ramp describes this as a "whales-first" dynamic, in which a small group of heavy spenders drives nearly all growth in AI expenditure. The July figure of $7,400 actually represents a slight dip from earlier months around $7,500, and the report flags potential signs of spending fatigue even among the top cohort.

Why a dedicated office, not a committee

Here is the catch: most companies track AI spend, and almost none track AI outcomes in any disciplined way. EY's move suggests a recognition that procurement, IT, and individual business units have been purchasing AI tools in isolation, with nobody assigned the awkward job of asking whether any of it moved the needle on actual operations.

On Ramp's platform, the vendor picture is already shifting in response. Anthropic now claims a 43.5% share of businesses on the platform, while growth in OpenAI's adoption has slowed, a hint that even dominant providers are starting to feel buyer pressure for measurable results rather than impressive demos.

In practice, a value realization office tends to do three things well: it forces budget conversations around outcomes, it pushes for vendor consolidation, and it gives finance a real seat at the AI table instead of a spectator role.

What to watch next

For corporate leaders running their own AI programs, three signals will indicate whether this model has teeth:

• Reallocation, not just addition. If EY starts pulling spend from underperforming projects rather than layering on new tools, that is a genuine shift in posture.
• Talent sourcing. Where the office pulls its leaders from matters. Finance and operations backgrounds suggest accountability; purely technical hires suggest another innovation theater.
• Contract consolidation. A serious value office tends to reduce overlapping AI agreements faster than any procurement memo.

Spending $7,400 per employee a month is one conversation. Defending that spend is another, and that is the conversation EY is now formalizing.