InsightsPractice
PracticeP100People & GovernanceBy Fabrice Macarty · Founder, CEO

The Human+AI Performance Model

PwC 2026: top 20% of organizations capture 74% of AI economic value and generate 7.2x more performance than the average. Not from deploying more AI. From governing the Human+AI unit with unified measurement, investment, and accountability.

+40%
quality improvement for the Human+AI combined unit on tasks correctly allocated within the AI capability frontier. The same study: -19 percentage points when allocated outside it. (Dell'Acqua et al., Harvard / BCG, QJE 2025, 758 consultants)
13%
of workers feel rewarded for redesigning how work gets done when immediate results are uneven. This is the incentive gap blocking implementation of a combined performance model. (Microsoft Work Trend Index 2026, 20,000 AI-using workers)
AI & DigitalStrategy & GovernancePeople & Leadership
The decision stakes

The AI is not the performance unit.
The system is.

As AI and agents take on more of the execution, the opportunity is to expand human agency and redesign how work gets done.Satya Nadella, CEO, Microsoft, Work Trend Index 2026 launch, May 2026

The CFO and the CHRO who govern AI and human capital separately are not measuring the wrong things. They are measuring two things that are one. The governance of task allocation is the difference between a +40% quality gain and a 19-point performance loss from the same tool and the same team (BCG and Harvard, QJE 2025). The 4 moves below formalize that governance.

P×R=V: the Responsibility of governing the Human+AI combined unit (who owns it, what it measures, and who answers for every output) is the R that makes the combined performance commercially capturable. Without it, the performance premium is real but invisible to the board. Without allocation governance, the performance loss is also real and invisible until a client or regulator names it. The Fusion Equation. Force multipliers, not trade-offs.

The decision tool

Four moves.
One governance framework.

01

Unit Definition

For each of your 3 to 5 core business processes, define the Human+AI performance unit in 3 lines: what the AI executes (synthesis, volume, availability), what the human anchors (direction, judgment, accountability), and what the combined unit produces as its measurable output. This spec is the governance layer that determines whether AI amplifies or erodes value in each process.

The BCG and Harvard study (QJE 2025) documents 40% quality gain for tasks inside the AI frontier and 19 percentage points of performance loss outside it. Same consultants. Same AI. Task allocation was the only variable. The unit definition is the allocation governance.
02

Combined Measurement and Ownership

For each combined unit, establish 3 reporting elements: an output dimension (volume produced at scale by the combined unit), a quality dimension (error rate, client satisfaction, audit pass rate), and an ownership anchor (the named executive who bears client and regulatory responsibility for every AI-assisted output). Report all 3 as one line at COMEX level.

Without a combined performance dashboard, the board sees AI adoption rate and headcount separately. Neither captures what the unit produces together. The ownership anchor is not a role description. It is a specific named executive who can be questioned by the board, a client, or a regulator.
03

Investment Fusion

Fuse the AI deployment budget (DSI capex, software licensing) and the workforce development budget (CHRO opex, reskilling) into one envelope at COMEX level with a single named owner. The combined line funds the system, not the components. Propose the fusion at the budget review that follows the unit definitions. Review it quarterly against the combined unit performance.

When AI spend sits in DSI and reskilling sits in CHRO, no one owns the return on the combined system. The investment fusion is not an accounting reorganization. It is the governance signal that the Human+AI unit is a single value-creation system with one accountable owner.
04

Allocation Review

Redesign the quarterly performance review to include one question per core process: for which tasks did AI allocation improve combined output, and for which did it degrade it? The answer updates the unit definition (Step 01), adjusts the ownership anchor (Step 02), and informs the next investment envelope (Step 03). The review makes the system self-correcting.

PwC's 2026 study finds that only 28% of even AI leaders conduct portfolio reviews to terminate underperforming AI initiatives. The quarterly allocation review is not an audit. It is the self-correction mechanism that keeps the governance of the combined unit calibrated to its actual output. What the COMEX reviews determines what the organization optimizes.
The proof

PwC Global

2026 AI Performance Study. April 13, 2026. 1,217 senior executives, director level and above, 25 sectors, multiple regions. 60 AI management and investment practices measured. AI fitness index: AI use and AI foundations.

7.2x

more AI-driven revenue and efficiency gains for the top 20% of organizations vs. the average competitor. These organizations capture 74% of all AI economic value globally. The performance gap is not in how much AI they deploy. It is in the governance of the combined Human+AI system.

PwC's 2026 AI Performance Study is the largest current measurement of AI governance outcomes across the global enterprise. 1,217 senior executives at large, publicly listed companies across 25 sectors and multiple regions reported the revenue and efficiency gains attributable to AI in their organizations today. PwC analysed 60 AI management and investment practices grouped into two dimensions: AI use and AI foundations. The six foundations include strategy, investment, data and technology, workforce, governance and risk, and innovation. The finding is direct: 74% of all AI economic value is captured by 20% of organizations. These leaders generate 7.2x more AI-driven performance than the average competitor and carry 4 percentage points higher profit margins. The study is explicit about the differentiator: the gap is organisational, not sectoral. Leaders invest in workforce and governance as a unified foundation, not as separate budget lines. Joe Atkinson, Global Chief AI Officer at PwC, stated at the study launch: "Many companies are busy rolling out AI pilots, but only a minority are converting that activity into measurable financial returns. The leaders stand out because they point AI at growth, not just cost reduction, and back that ambition with the foundations that make AI scalable and reliable." A parallel 2026 study by Larridin (365 senior leaders, February 2026) documents the failure mode of the 80%: 58% cite unclear or fragmented ownership as their primary barrier to measuring AI performance. 75% lack AI governance frameworks entirely. The 20% who capture 74% of value have closed both gaps simultaneously.

The 7.2x performance multiplier is not an AI deployment premium. It is a governance premium. The organizations capturing 74% of all AI economic value govern the Human+AI combined system with unified measurement, unified investment, and unified accountability. That is what the 4 steps above build.

The full case

PDF · 5 slides