Proved, Probable, Possible: A Board Reporting Framework for the Return on AI Efficiency Gains

Management usually reports the return on an investment in artificial intelligence to its board as a single figure. That figure carries two different things at once: benefit that has reached the accounts and benefit that management expects to reach them. When management combines the two, a board cannot tell how much of the reported return has been banked, how much depends on decisions not yet taken, and how much rests on capacity that no one has acted on. The Australian Institute of Company Directors now advises boards on how to assess AI return on investment but supplies no form of report in which to do so. This paper proposes a reporting framework that separates those quantities. It classifies the return on an AI efficiency gain into three classes. Proved return has met a four-part recognition rule and is tied to a ledger line. Probable return is supported by a dated conversion commitment from the owner of the affected cost centre. Possible return is verified released capacity with a feasible path to the income statement and no commitment. Released capacity with no feasible path is reported as stranded and carries no value. The structure is adapted from mineral reserves reporting, in which confidence in a quantity and the conditions for converting it into economic value are assessed on separate axes. The framework reports return as a band of three figures: a floor on proved return alone, an expected figure that adds probable return adjusted by the organisation’s own record of commitments kept, and a ceiling that adds possible return. It adds two measures that, to the author’s knowledge, existing AI return methods do not report. Leakage is the spend that continued in the period on capacity that AI had released and that was not converted. Conversion yield is the share of past commitments that were later proved, which calibrates the expected figure from the organisation’s own history in place of an assumed realisation rate. An AI Return Statement brings the classes, the band, a capacity roll-forward, leakage and yield onto a single page, with attestation divided between the technology function, the cost centre owner, finance and internal audit. In a worked illustration, a portfolio with a conventional return of 473% reports a floor of 17%, an unadjusted expected return of 183% and annualised leakage of about $402,000. Carried over three reporting dates with constructed events, the same portfolio shows the floor rising to 83%, the ceiling rising to 300% after a change in the operating model, and the expected return falling as a lapse and a reversal take the conversion yield from 0.85 to 0.65. The framework is a design artefact evaluated analytically and by constructed illustration. It has not been applied to an organisation’s records, and five propositions state how it could be tested.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-20
DOI
https://doi.org/10.5281/zenodo.24398641
Primary Topic
Accounting and Organizational Management
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Proved, Probable, Possible: A Board Reporting Framework for the Return on AI Efficiency Gains

M Maruf Hossain
4 links
Zenodo (CERN European Organization for Nuclear Research)
Accounting and Organizational Management
preprint

Proved, Probable, Possible: A Board Reporting Framework for the Return on AI Efficiency Gains

M Maruf Hossain
preprint en

Abstract

Management usually reports the return on an investment in artificial intelligence to its board as a single figure. That figure carries two different things at once: benefit that has reached the accounts and benefit that management expects to reach them. When management combines the two, a board cannot tell how much of the reported return has been banked, how much depends on decisions not yet taken, and how much rests on capacity that no one has acted on. The Australian Institute of Company Directors now advises boards on how to assess AI return on investment but supplies no form of report in which to do so. This paper proposes a reporting framework that separates those quantities. It classifies the return on an AI efficiency gain into three classes. Proved return has met a four-part recognition rule and is tied to a ledger line. Probable return is supported by a dated conversion commitment from the owner of the affected cost centre. Possible return is verified released capacity with a feasible path to the income statement and no commitment. Released capacity with no feasible path is reported as stranded and carries no value. The structure is adapted from mineral reserves reporting, in which confidence in a quantity and the conditions for converting it into economic value are assessed on separate axes. The framework reports return as a band of three figures: a floor on proved return alone, an expected figure that adds probable return adjusted by the organisation’s own record of commitments kept, and a ceiling that adds possible return. It adds two measures that, to the author’s knowledge, existing AI return methods do not report. Leakage is the spend that continued in the period on capacity that AI had released and that was not converted. Conversion yield is the share of past commitments that were later proved, which calibrates the expected figure from the organisation’s own history in place of an assumed realisation rate. An AI Return Statement brings the classes, the band, a capacity roll-forward, leakage and yield onto a single page, with attestation divided between the technology function, the cost centre owner, finance and internal audit. In a worked illustration, a portfolio with a conventional return of 473% reports a floor of 17%, an unadjusted expected return of 183% and annualised leakage of about $402,000. Carried over three reporting dates with constructed events, the same portfolio shows the floor rising to 83%, the ceiling rising to 300% after a change in the operating model, and the expected return falling as a lapse and a reversal take the conversion yield from 0.85 to 0.65. The framework is a design artefact evaluated analytically and by constructed illustration. It has not been applied to an organisation’s records, and five propositions state how it could be tested.

Zenodo (CERN European Organization for Nuclear Research)
Accounting and Organizational Management
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