Consumer Appetite for Agentic AI Could Transform Network Traffic

Artificial intelligence agents are poised to become a significant "invisible" network user category by the end of the decade, according to new research from Ericsson ConsumerLab. The study indicates that 38% of current smartphone users expect to delegate their daily tasks to autonomous AI agents by 2030.

Currently, only about 4% of smartphone users utilize agentic AI capabilities—where the AI independently manages tasks typically performed by the user—according to the research focused on communications service providers (CSPs). However, surveyed users across 27 markets anticipate a rapid increase in adoption, with 38% expecting these features to be a standard part of their device experience within the next few years.

Approximately 10% of smartphone owners globally are currently classified as "AI super users," individuals who engage heavily with multimodal AI tools. This demographic is expected to triple to about 30% by 2030. The most common tasks expected to be delegated to these agents include product comparisons, scheduling, health monitoring, and the oversight of recurring purchases, with respondents anticipating they will save up to an hour of manual browsing time daily.

The report highlights that current AI super users spend roughly 2.5 times longer utilizing AI apps and services than the average user, a gap that is expected to widen to four times longer by 2030. The findings also point to high consumer expectations regarding the role of CSPs in managing AI agent identity, network capabilities, and overall mobile security.

“We have seen many times that early adopter behavior and expectations - in this case, the AI super users - are a strong indicator of much wider-scale adoption to follow. We believe this will be the same with the use of agentic AI,” said Jasmeet Singh Sethi, Head of Ericsson ConsumerLab. “Multimodal and agentic AI growth will not only create network quality expectations for consumers in terms of their experience but will also result in significantly increased background actions in the networks that need to be handled securely and efficiently.”