What Still Gets Paid When AI Changes the Interface?

Summary

This opinion and investment analysis asks which economic bottlenecks remain valuable when AI agents take over parts of software interfaces. Its central distinction is between a company’s interface, which may become easier to replace or reduce seat growth, and the scarce data, workflows, payments, or distribution underneath it. Experian is presented as a relatively strong example: agents may change how banks make credit decisions, but credit histories, identity data, and fraud signals remain necessary; the company reported 8% organic revenue growth and 15% Benchmark EPS growth in FY26. RELX is viewed similarly because legal, scientific, and risk information remains necessary for lawyers, researchers, and the agents they use; underlying revenue rose 7% and operating profit 9% in the first half of 2026. Intuit is more ambiguous: QuickBooks owns accounting history and important workflows, but an agent could perform much of the accounting in the background, making payments, payroll, and tax services more strategically important. Adobe’s Q3 revenue grew 13%, but the author sees AI as attacking the creation and editing functions Adobe has historically monetized and is less certain about what remains scarce. MSCI and S&P Global are treated as less clear examples because their earlier valuation gains may explain some recent stagnation. The analysis therefore favors Experian, RELX, and Intuit over Adobe, while leaving the broader question open: after AI changes the interface, which underlying asset still commands payment?