AI Users — The Companies Putting AI to Work
AI Users is a Closelooknet index of 32 large caps that use AI inside businesses they already run — insurers, banks, retailers, logistics groups, drug makers and the IT firms that implement AI for others. Every member needs dated, public evidence of AI in production; price performance plays no part in who gets in.
What it tracks
Most AI money in 2026 went to the companies that build AI: chips, memory, power and the data centres around them. AI Users tracks the other side — companies where AI is not the product but makes an existing business run better. An insurer whose models draft every damage estimate belongs here. A software company that sells an AI product belongs in Agentic Winners 40 instead.
The index sits next to three others in the Closelooknet AI family. Rubin builds the AI factory, the Agentic Ecosystem operates it, Agentic Winners 40 sells AI products, and AI Users puts AI to work. No company sits in two of them.
Seven sections
Members are grouped by what AI does inside them, not by industry: workflow and service productivity, commerce and customer growth, supply chain, physical operations, decision intelligence and risk, research and development, and implementation. The last section holds the IT firms — Accenture, Infosys, IBM and others — that bring AI into client businesses.
The evidence test
Each company carries an adoption score from 0 to 4. A 2 needs AI in production in a core process with the scope, a number and a date. A 3 means the company lets AI act, not only answer, and ties it to a cost or revenue line; a 4 adds a financial benefit measured in two reporting periods.
A claim without a scope or a number stops at 1. Every score is backed by dated entries from filings, calls and company releases.
Point in time
The index is based at 1,000 on 30 December 2025, and the starting members were judged only on evidence published by then. That keeps hindsight out of the history. Three members without qualifying evidence at the base are marked PIT and are decided at the November 2026 review.
How it is run
Equal weight, reset at each quarter end, priced in US dollars. Members change at reviews on the third Friday of February, May, August and November. A company leaves when its adoption score falls to 1, or when its advantage over its own industry has gone in two reviews running.
How we read it
We measure the payoff against each company’s industry: how much faster revenue grows than headcount, and whether the operating margin keeps up. Our read is that markets pay for AI capacity before they pay for AI use, so the users’ benefit may show in their numbers before it shows in their share prices. The launch report has the first figures.