Super interesting, and it connects to something we’ve been thinking about in our work in The Agentic State: AI agents are not a substitute for fixing bad UX, but they are arriving regardless. Governments now need to design for both humans and agents, with formal delegation frameworks, clean APIs, and structured data that let agents operate safely without hacking their way through brittle interfaces. The equity dimension is critical. Free and freemium agents will almost certainly exist, but what will matter are the performance gaps that emerge over time. It is easy to imagine a future in which differences between free, commercial, and premium agents quietly reproduce existing divides. Higher tier tools could offer smoother, more reliable access to public services for those who can pay. The challenge is to build a competitive and regulated ecosystem, with a public option, so that agents help close service gaps rather than widen them. This example shows the issue is no longer theoretical. People are already using agents in the wild.
When agents start negotiating on our behalf
Linked:the post he was responding to
Written by Tiago C. Peixoto. No agents involved – this predates the machinery. It first appeared on LinkedIn on November 15, 2025, and is reproduced here. Found an error? Tell me – corrections are logged.