Everyone claims to be using AI. Earnings calls are full of it, press releases are full of it, job titles suddenly have "AI" bolted onto them. Paul Cheek, a senior lecturer at MIT Sloan, got tired of taking companies at their word and built something to check.
His project, called the AIDE Index, scores every company in the S&P 500 on four things: whether leadership actually understands AI, whether they talk about it in public, whether it shows up in their stated strategy, and whether it gets built into real products and workflows. The scores come from job postings, patent filings, earnings transcripts, and other public records, not from asking executives to grade their own homework.
The results confirm what a lot of people suspect but rarely see proven with data. Nvidia, Amazon, Meta, and the energy company SLB currently score highest, and the technology sector as a whole leads every other industry by a wide margin. That part is not a surprise.
What is more useful is a separate signal underneath it. A related MIT-led study of company filings found that only 11 percent of S&P 500 firms had truly built AI into their daily operations by 2025, even though the vast majority now mention AI whenever they talk to investors.
That gap matters for anyone running a business outside the tech industry. It means the loud AI talk you hear from competitors, vendors, and consultants is mostly just talk. If you have not deployed much AI yet, you are not behind some invisible curve, you are in the same position as roughly nine out of ten large companies.
The opportunity is that being in the small group that actually does the work puts you ahead of almost everyone, not just even with them. It also means procurement and partnership decisions need a sharper eye. A vendor's AI claims on a slide deck are not evidence of anything, and the kind of proof Cheek's index looks for, real hiring patterns, real patents, real changes to how work gets done, is the same proof any business leader can ask a supplier for before signing a contract.
The second part of this story is where it gets more personal for anyone managing people. Cheek has a book coming out that argues company org charts are about to change shape. Instead of a chart built entirely around human employees, he expects charts built around roles, or what he calls "nodes," that can be filled by a person or by an AI system doing the same job.
That idea sounds abstract until you translate it: some jobs currently done by a person on your team may soon be listed on an org chart the same way, regardless of whether a human or a piece of software fills the box. That does not mean every role is at risk right away. It does mean how companies plan headcount, budgets, and job descriptions is about to get rewritten, and the ones figuring that out now will have a real head start over the ones still waiting to see what happens.