AI Agents / POV

AI Agents as Users: What Changes When Software's Customers Aren't People (Part 2)

If AI agents become software's primary users, interfaces change: predictable outputs, clear boundaries and identity matter more than screens. Part 2 of a series.

What changes if AI agents become the main users of software? An agent needs to know what it can do, what inputs to send, what result to expect and when it must stop. Other systems need a way to find and call these functions, which is where ideas like the Model Context Protocol fit in.

I use GitHub Copilot and customer support to show what that looks like in practice, with people stating intent while systems coordinate the steps. The part I would not skip is accountability. When an agent acts on someone’s behalf, authentication, authorization and audit trails need to be part of the interface design, especially in regulated industries such as banking. Even if agents make execution cheaper, someone still needs to define the task and judge whether the result makes sense. That requires knowledge of the work itself.

Summary of an essay originally published on Indexnine. Read the original ↗