The screen was always a workaround
A user interface exists because a human being and a machine could not otherwise negotiate. Buttons, menus, forms, dashboards — every one of them is a translation layer, built because software could not simply be told what to do. We got extraordinarily good at that translation. We hired for it, we ran design systems for it, we made it the defensible part of the product.
Once agents can act on your behalf, the translation stops being necessary. Not gradually, and not everywhere at once, but in the places where the interface was pure overhead: booking, comparing, filing, reconciling, procuring. Nobody ever wanted the form. They wanted the outcome the form stood in for.
What replaces it is a negotiation
The interesting part is not that humans stop clicking. It is what the traffic becomes. When my agent talks to your service, the exchange looks less like a page view and more like a negotiation between two parties who need to establish, quickly: who is this, is it allowed to act for the person it claims to represent, what is this request worth, and should it be served at all.
That is not a design problem. It is an identity, trust and settlement problem — and it happens on the wire, not on the screen.
When the interface disappears, the value does not vanish. It relocates to whoever sits between the agent and the thing it wants.
Which is why I am watching Cloudflare
Cloudflare already sits in that position for an enormous share of the internet, and for reasons that have nothing to do with this shift. It terminates the connection. It already answers the question "is this a human, a good bot, or something else" millions of times a second, because that was the job it was hired for. It has the identity primitives, the edge footprint, and — crucially — a commercial relationship with the sites on the other side.
If the post-UI world needs a layer where agent traffic gets authenticated, rate-shaped and paid for, the incumbents at that layer did not set out to build it. They are just standing where it will have to go. That is usually the more interesting bet than the company that announces it is building the thing.
I hold no view here worth acting on financially, and this is not investment advice. It is a thesis I am watching, and the kind of question I find myself asking on behalf of clients: if the screen stops being where your product is experienced, what exactly is your company still the best in the world at?