Costs You Can't Attribute, You Can't Govern

The Cost-Center Mismatch™ framework applies a long-standing finance concept — cost-center mismatch — to AI consumption and adds a four-question attribution test. The core insight: the party creating the cost isn't the party paying, so nobody has a reason to ask for a limit. Question 1: Can you attribute AI consumption by team and use case, down to the application ID? Question 2: Who pays when the cost doesn't land where it was created? Question 3: How long is the gap between someone consuming and someone seeing the cost? Question 4: What would your business units learn from even a rough chargeback? Includes an illustrative Sev-1 incident cost model (20 engineers, 4 hours, ~$47 on Sonnet 5.5 at current published pricing), the model choice multiplier (~20x per token between Haiku 5.5 and Sonnet 5.5), and the practical stance: attribute first, cap second. Approximate allocation is cheap. An absent one is expensive. Developed from 13 years of enterprise API gateway operations and capacity planning. Keywords: AI cost attribution, cost-center mismatch, AI FinOps, AI spend, enterprise AI economics, AI governance, margin leakage, PE due diligence, AI budget, AI TCO, AI incident cost model, AI model choice multiplier, unowned cost category, CFO AI budget, AI OpEx.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24228679
Primary Topic
Information Technology Governance and Strategy
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Costs You Can't Attribute, You Can't Govern

DHAVALA DRUHIN
Zenodo (CERN European Organization for Nuclear Research)
Information Technology Governance and Strategy
article

Costs You Can't Attribute, You Can't Govern

DHAVALA DRUHIN
article en

Abstract

The Cost-Center Mismatch™ framework applies a long-standing finance concept — cost-center mismatch — to AI consumption and adds a four-question attribution test. The core insight: the party creating the cost isn't the party paying, so nobody has a reason to ask for a limit. Question 1: Can you attribute AI consumption by team and use case, down to the application ID? Question 2: Who pays when the cost doesn't land where it was created? Question 3: How long is the gap between someone consuming and someone seeing the cost? Question 4: What would your business units learn from even a rough chargeback? Includes an illustrative Sev-1 incident cost model (20 engineers, 4 hours, ~$47 on Sonnet 5.5 at current published pricing), the model choice multiplier (~20x per token between Haiku 5.5 and Sonnet 5.5), and the practical stance: attribute first, cap second. Approximate allocation is cheap. An absent one is expensive. Developed from 13 years of enterprise API gateway operations and capacity planning. Keywords: AI cost attribution, cost-center mismatch, AI FinOps, AI spend, enterprise AI economics, AI governance, margin leakage, PE due diligence, AI budget, AI TCO, AI incident cost model, AI model choice multiplier, unowned cost category, CFO AI budget, AI OpEx.

Zenodo (CERN European Organization for Nuclear Research)
Information Technology Governance and Strategy
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon