AI companies are betting that agents will make everyday tasks easier, but many people are wary of giving them the access that would make them useful. Axios reports that survey findings point to a sizeable gap between industry ambition and public willingness to hand over email, travel bookings or money.
Watch Desk analysis
What happened
Axios’s 30 September report argues that agent adoption faces a practical hurdle: people may not want to reorganise their lives around AI, or grant software broad access to their digital accounts. It points to Meta’s Muse, Microsoft Copilot and OpenAI’s Dots as examples of companies pushing agentic products towards consumers and businesses.
The figures Axios cites are striking. In an international Thales poll, 13% of respondents said they would let an AI helper read their emails, 11% would let it rebook travel and 7% would allow it to move money between bank accounts. YouGov found that 56% would not let an agent shop for them at all; only 10% trusted one with more than $25 without approval.
The report also describes uneven use. It cites St Louis Fed figures showing AI use across many occupations, but with much higher adoption in some screen-based professional jobs than in roles such as nursing, reception and animal care. A Pew finding cited by Axios says 51% of Americans avoided AI chatbots; among that group, 79% cited privacy concerns.
Why it matters
An agent that cannot access a person’s accounts may be less useful, but access is precisely where trust gets tested. The gap is not just about whether the software can book a trip or sort an inbox. It is about who is comfortable letting it act, and who has jobs and routines that fit tools built around screens.
That matters for companies counting on agents to become everyday products. Early enthusiasm among people already comfortable with AI does not, by itself, show that the wider public wants an assistant embedded in daily life.
Our read
The useful takeaway is not that agents are doomed, or that one poll has settled the matter. It is that adoption depends on more than clever demonstrations: people need a reason to grant access, and a clear say over what happens next. “It can do the task” is only half the pitch; “I’m happy to let it” is the other half.
What to watch
- Whether agent products make approval and access controls clear enough to earn trust.
- Whether adoption grows beyond people already using paid AI tools.
- Whether agents become useful in jobs and daily tasks that do not happen mainly on a screen.
Discussion spark: Should AI agents be allowed to act across people’s accounts by default, or should every meaningful action require approval, even if that makes the assistant less convenient?
Sources and evidence
- AI agents have a normal-people problem (30 September 2026, 09:30 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.