Five9 Research: Consumers Accept AI in Service, But Want a Human Option

What happened

Five9, a cloud contact-centre software provider, has released new research indicating that consumers are receptive to artificial intelligence in customer service, but that their trust in the experience still depends on being able to reach a human. The finding, as reported by Yahoo Finance, is that openness to automated help and the need for a person to be available are coexisting expectations rather than competing ones.

The reporting available to us is limited to the headline and lede. It does not set out sample sizes, methodology or detailed survey results, so this briefing does not go beyond the core finding. Readers should consult the original Five9 research for those specifics.

Why it matters

The research arrives as AI moves from pilot projects into front-line service operations. The tension it describes, comfort with automation alongside a need for a human fallback, is the central design problem for any organisation deploying virtual agents, conversational AI or automated call routing. When an automated tool resolves a query well, customers accept it. When it stands between a customer and a person, that acceptance can turn into frustration quickly. For leaders in experience, AI and digital transformation, this suggests that model capability is only part of the case. Escalation paths, the quality of handover and the visibility of a human option all shape whether customers trust the channel at all.

The finding also points to a shift in how contact-centre roles could be understood. Human agents are not only a cost to be automated away. They may also act as the trust mechanism that the wider automated journey depends on.

The Renascence take

Most organisations will read this as a prompt to add a "talk to a person" button. That misses the point. Trust is built earlier and more quietly, in whether the AI recognises when it is out of its depth, whether customers have to repeat themselves after a handover, and whether the person who picks up has the context to resolve the issue.

Customers are not asking for humans instead of AI; they are asking for confidence that a human will be there if the automation fails them. The underlying principle is that perceived control drives trust, so the design goal is to make the exit visible and effortless before anyone needs it. Operators should audit the handover first: track how often customers repeat information after escalation, and check whether the human agent actually receives the AI's context. A clear, easy route to a person is a trust signal in its own right, not a concession.