What changed
Torsten Slok, chief economist at Apollo Global Management, published a note on October 24 showing AI delivered zero financial payoff to the S&P 500 outside the seven biggest tech companies. Revenue per worker at Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla rose from around 550,000 dollars in 2022 to nearly 800,000 dollars in 2026. At the other 493 companies, that number stayed flat at 550,000 dollars since 2019. Profit margins tell the same story: up for big tech, flat for everyone else. Slok found no sign AI is lifting productivity for the firms that buy it, only for the firms that sell it.
What you can do with it now
Pick one task your team does every week that takes two hours and costs nothing to get wrong. Customer follow-up emails, meeting summaries, first-pass social media captions, or inventory counts. Give it to AI for one month. Track the time saved and whether someone still has to fix the output. If you save an hour a week and the output is usable, keep it. If you're spending the same time fixing mistakes, stop paying. Most teams see gains in tasks that are repetitive and low-risk. Start there, measure it, and ignore what the software company promises in the demo.
What to keep in your own hands
Don't hand AI a task where a mistake costs you a client, breaks a rule, or requires judgment you can't write down. Pricing a custom quote, deciding whether to grant a refund, hiring someone, or answering a complaint that's already escalated. If the task depends on knowing the client, the context, or the unwritten rules of how you work, a person does it. AI can draft the answer; you decide if it ships. The companies making money from AI today are the ones selling the tools. The companies that will make money from it tomorrow are the ones using it on tasks where speed beats perfection and the stakes are low.