Section
Coluna de Rafael Cordeiro sobre como empresas transformam inteligência artificial em operação, decisões melhores e resultado mensurável. Rafael é cofundador da OperOS AI.
Saving thousands of hours does not guarantee efficiency, quality or growth. In the author's view, incrementality takes on a new dimension with AI, and how freed-up time is used is a business decision.
From the fall of the barrier between technical and non-technical to the risk of not training the next generation of seniors, Rafael Cordeiro wrote in 2026 about what AI changes in work and careers. This collection brings together five notes, three of them from an October sequence.
When agents start researching, comparing and buying, who makes the final decision changes, and so does how retail measures results. This collection brings together three notes by Rafael Cordeiro, from February to September 2026.
From January to September 2026, Rafael Cordeiro wrote short notes about the same bottleneck. AI impresses in the demo and stalls when it meets the real company. This collection brings together six of them.
Hours saved and costs cut are the most obvious part of the return. In the author's view, the more interesting value lies in work that only became worth doing because AI now exists.
Users, prompts and tokens show how much AI is being used. In the author's view, the right question is what happened because of AI that would not have happened without it.
SaaS sped up digitization and created an invisible dependency. In the author's view, AI coding is redrawing the line between buying and building, and companies need to keep control of the intelligence of their own business.
Large companies have more capital, data and specialists. In the author's view, midsize companies have something that weighs more at this stage, the speed to redesign entire processes around AI.
Manufacturers and retailers already exchange sales, inventory and demand data, and the misalignment persists. In the author's view, before connecting agents across companies, the decisions inside each one have to become visible.
Controlling AI's access to company systems solves only part of the problem. The author argues for governed business definitions before trusting the answers agents give.
Five capabilities turn the gateway into the piece that lets a company grow its use of AI with shared policies, visibility and standards.
Agents can reach thousands of tables and documents and still not know how the company works. The author proposes a seven-layer architecture to fix that.