Skip to content
Advertisement
UBS warns AI agents could upend internet business models built for humans

UBS warns AI agents could upend internet business models built for humans

The bank's Americas wealth CIO says agentic AI threatens digital advertising and software-as-a-service revenue that depends on human attention

For about as long as the commercial internet has existed, its economics have rested on one quiet assumption: somebody is actually looking at the screen. UBS now argues that assumption is starting to crack.

Ulrike Hoffmann-Burchardi, chief investment officer for the Americas at UBS Wealth Management, said in a note and podcast released Monday that AI agents acting on behalf of people undermine two of tech’s most lucrative revenue engines. Those are digital advertising and enterprise software sold as a service.

The eyeball economy meets the bot that doesn’t blink

Her core argument is simple. Websites, checkout flows and software dashboards were all designed around a human clicking through tabs.

“Since its inception, the internet has been built for human eyeballs,” Hoffmann-Burchardi said.

AI agents that shop, compare and click for a user break that design premise, she argued. If the agent does the browsing, the human never sees the page, the banner or the carefully tested button color.

Advertisement

That matters because digital advertising is fundamentally a business of selling human attention. An ad placed in front of software that reads the page for its data, not its vibes, is a harder product to price.

The software side of the warning works differently. Many software-as-a-service companies tie revenue to headcount and usage, meaning more human users generally means more licenses sold.

UBS research flags that model as exposed. If an agent can handle work that once required several people logged into a dashboard, the per-seat math starts to look shaky.

From B2C to B2A

UBS frames the shift as a move from business-to-consumer to what it calls business-to-agent, or B2A. In plain terms, companies may increasingly be selling to software that represents a customer, rather than to the customer directly.

According to the research, UBS analysts believe financial markets have not yet fully priced in that transition. The bank sees the change as underappreciated, particularly for business models built around user metrics and licenses.

The report also points to a possible migration of value. Revenue that traditional companies capture today could shift toward new platforms built around agents, UBS suggests.

A theme UBS has been building for months

Monday’s note is not a sudden pivot. Hoffmann-Burchardi has been making versions of this case publicly throughout 2026.

She discussed agentic AI’s disruptive potential on CNBC in June and again in October. In those appearances, she compared its likely impact on the internet to the shake-up AI coding tools have caused in software development.

UBS has been tracking these trends in its research since April 2026, according to the findings. Over that period, the bank has adjusted its view of how agentic AI could reshape the digital economy, including its effects on SaaS pricing and revenue structures.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.