Murata Manufacturing will pay part of Accenture's AI fee only on results, not hours
The Japanese component maker will pay part of Accenture's fee only when AI performance indicators are met, days after Accenture's own research showed falling token prices will not cut AI bills. Outcome pricing just entered the mainstream of enterprise AI consulting.
Richard Tang · 3 min read
Accenture+4.07% — Accenture, up 4.07 percent today has agreed to be paid for AI results, not AI hours: on October 28 Murata Manufacturing, Japan's ceramic electronic-components maker, announced a three-year partnership with Accenture's Japanese arm in which part of Accenture's fee on selected AI initiatives is tied to how much agreed performance indicators are achieved 12. As explicit an outcome-pricing clause as has surfaced in enterprise AI consulting, it tells buyers and sellers where consulting pricing goes next.
If you buy enterprise AI services, this clause is built for you. If you sell them, Accenture just accepted the risk you wanted them to avoid.
Falling token prices do not mean falling AI bills
The timing is pointed: one day earlier, Accenture Research published findings that falling AI token prices will not cut AI bills. Average token prices should fall about 19% over two years while consumption rises about 78%, so "lower unit costs do not necessarily translate into lower overall spend" 3. Those numbers make the consulting problem plain: per-token prices fall, yet usage grows, so a buyer's AI bill can rise even as the underlying technology gets cheaper.
A buyer holding that research has an obvious answer, pay consultants for achieved outcomes rather than hours; Murata's clause is that answer in contract form, with shared targets and a fee that moves when the indicators do.
The fee covers the harness, not the model
The deal's scope shows where the real work sits: data-driven quality control, IT architecture and maintenance, governance for protecting data and intellectual property, and training the people who run AI, not just model deployment 1. Accenture has said only 10 to 20% of enterprise workflows justify frontier-model capabilities at all 4; the rest is integration, governance and training, and Murata is now paying Accenture only when that work delivers. Accenture wins margin if the model scales; it bears the cost if outcomes miss.
Accenture's own survey explains why buyers will keep demanding this: if token prices fall further, only five percent of organizations would keep the savings as savings, while most would pour them straight back into more AI use 3.
Almost none of the savings would be kept; most become more AI use
Data
| Value | |
|---|---|
| Retain the savings | 5 |
| Expand AI to more employees | 42 |
| Fund previously too-expensive apps | 35 |
Lower prices fund expansion, not refunds, which is exactly the dynamic the outcome clause protects Murata from.
The week's other moves price the work, not the hours
The week's other Accenture news points the same direction: a $151 million fixed-price ICBM engineering task order running to 2030 5, plus acquisitions of UK customer-intelligence firm Sagacity 6 and Miami construction-tech consultancy Avicado, adding more than 300 named specialists rather than billable ambiguity 72. Watch two things: whether the Murata clause's selected-initiative scope widens to the whole partnership, and whether Cognizant+0.39% — Cognizant, up 0.39 percent today or a peer signs a comparable outcome fee. The first movers set the benchmark buyers will demand from everyone.
Get Deals by email, free
The acquisitions, rounds and contracts that move AI money. In our daily email, every weekday.
