01 · THE AI MANDATE
Scope moved faster than structure
91%
Say AI expanded their responsibilities in the past 12 months
57%
Of those whose scope expanded significantly are very confident the role persists — against 33% of those reporting no change
86%
Are confident overall that the role exists in three years
HOW AI CHANGED THE ROLE, AGAINST CONFIDENCE IT SURVIVES THREE YEARS
Categories under n=5 omitted.
The finding underneath the finding
Executives whose responsibilities expanded significantly are the most confident group in the study, and they are not paid more. Their median base of $305,000 sits below the $347,500 reported by the no-change group — though that group is small (n=18) and skews toward larger organizations. Roughly two thirds of these organizations added AI accountability to an existing data leadership seat rather than funding a separate one. The executives absorbing that scope read it as job security. The market has not yet repriced it.
External corroboration: the EY Chief Data Officer Role Study (second edition, April 2026) reports that only 33% of CDOs hold full authority over AI while 80% share the responsibility, and that AI-inclusive titles now account for 31% of data leadership roles against 26% still titled Chief Data Officer. Both point the same way — scope has moved faster than structure.
02 · WHAT PREDICTS PAY
Span of control, then scale
We fit base salary against nine characteristics of the seat, then removed each in turn to see how much unique explanatory power it held. The first column shows what each factor explains on its own; the second shows what is lost when it is taken away and the other eight remain. They answer different questions, and the ranking is the finding rather than the values.
SHARE OF BASE SALARY VARIATION EXPLAINED
OLS on ln(base salary), n=235, R² = 0.47, HC3 robust standard errors. The two columns answer different questions and neither sums to 100%. Ranked by unique contribution.
Number of people managed leads on both columns. Industry, geography and company revenue follow. Executive level ranks fifth, below all three. Reporting line ranks seventh and work arrangement last — the two things executives ask about most often sit at the bottom of the list.
No factor explains more than a fifth of the variation on its own, and the list falls away quickly after the top four. In a mature function you would expect a tighter relationship between scope and price. This market has not converged on one, which means individual benchmarking and individual negotiation still carry real weight.
03 · ORGANIZATIONAL DESIGN
Who you report to matters less than what you run
Reporting line is the question these executives ask most often and the one the compensation data answers least. It sits near the bottom of the factor ranking. Span of control sits near the top. Two executives with the same title — one reporting to the CEO with forty people, one reporting to the CIO with four hundred — are not paid alike, and it is not the reporting line doing the work.
This cuts against how the role is usually discussed, so it is worth stating carefully. A seat at the CEO table is a real thing: it changes what the executive can decide and how fast. It is not, on this evidence, what sets the number. The EY study reaches a compatible conclusion from the other direction — it finds organizational structure and role clarity correlate more strongly with data-office performance than technology does. Reporting line predicts whether the executive can execute. Span of control predicts what they are paid.
WHERE THE CAPABILITY SITS
Who each title reports to
Reporting line barely moves compensation, but it tells you a great deal about how the organization is built. Split by title, the picture is not the one the job titles imply.
Only 27% of chief data officers report to the CEO or the board; 69% report to another C-suite executive. Chief information and technology officers are the mirror image at 58% to the CEO. For most organizations the data office is still a function inside someone else’s remit rather than a seat at the top table.