Should AI targets be in retail executive pay?

What happened

Ralph Lauren has added AI goals to the factors that can adjust its executives' annual incentive payouts. Its fiscal 2026 proxy statement, sent to shareholders in June, ties the short-term plan to business priorities that "focus our executives on AI literacy and enablement", Reuters reports, and says strong performance on its AI metrics contributed to maximum payouts for top executives. Reuters cites Conference Board and ESGUAGE data showing about 12 per cent of S&P 500 companies disclosed AI-related goals in executive incentive plans this year, up from 5 per cent in 2024; a separate Pearl Meyer review of 2,500 proxy statements put the share of public companies across all sectors at 2 per cent. In retail it is still rare, Modern Retail reports.

Why it matters here

Australian boards are in the middle of AGM season, when shareholders vote on remuneration reports and two consecutive strikes trigger a vote on spilling the board. Most large listed retailers now tell investors that AI is part of their strategy. The obvious next step is to put it in the bonus scorecard, and the Ralph Lauren model is the one a remuneration committee will see first. Reuters' figures cover the S&P 500; I have not seen an equivalent count for the ASX.

The take

Putting AI in executive pay is fine. Paying executives for AI activity is a mistake.

"AI literacy and enablement" measures inputs: training completed, tools rolled out, teams with access. Those are worth doing, and they are also the easiest things in the business to hit. A scorecard that rewards them will be met, and the bonus will be paid whether or not anything changed for customers or for margin. It is the same trap as paying for the number of stores refitted rather than what the refits did to sales per square metre.

My view is that if AI goes in the scorecard, it should be measured the same way as every other investment: by what it changes in the profit and loss. Cost to serve per order. Stock availability. Conversion on the channels where an assistant is live. Hours of store labour moved from back-of-house to the floor. Pick one or two that the AI programme is supposed to move, set the baseline before the year starts, and pay on the result. If nobody can name the line it is meant to move, it is not ready for the bonus.

There is a second reason to hold the line. Most AI tools retailers buy are priced by usage. A goal that rewards adoption rewards spend, and it pays executives for driving up a cost they should be managing.

What to do now

If AI is going into next year's short-term incentive, ask the remuneration committee one question before it is approved: which number in the accounts will be better if this goal is met? If the answer is a training or rollout count, rewrite the goal. And if the business cannot yet measure that number, the first year's AI goal should be building the measurement, not paying for adoption.