Customers Trust AI Twice as Much When a Human Is Within Reach

Most service organizations treat the route to a human agent as the expense that AI is supposed to remove. New consumer research says that framing is backwards. Customers are far more willing to let a bot try when they can see a person standing behind it, and far less willing when the door is hidden. Burying the human option does not push more contacts through automation. It pushes customers away from the automation altogether.

The Visible Exit Is What Earns Trust

Five9’s 2026 Consumer CX Report, published October 14 and based on an online survey of 3,000 consumers in the U.S., U.K. and Germany conducted with Hanover Research in April, puts a number on the effect. When a human option is visible, 55% trust AI to resolve their issue. When it is not, that falls to 26%. Seventy-one percent say it is very or extremely important to know when they are talking to an AI agent.

The gap is widest among older customers. CMSWire’s breakdown of the report shows baby boomer trust moving from 56% with a visible human to 16% without one, and the Silent Generation from 57% to 5%. Even millennials, the most comfortable cohort, drop from 58% to 33%. Removing the exit does not just cost a few points of goodwill. For some segments it removes almost all of it.

Adoption tells the same story. Only 20% of consumers use AI options most of the time or always, and among those who use AI, 52% say the main reason is that it was the only option on offer. Forty-eight percent try to avoid AI as much as possible, and about four in ten say they are less likely to use a company at all if it relies on AI for customer service. Usage driven by a lack of alternatives is not the same as acceptance.

Ten Minutes, and Then the Phone Rings

Customer patience with automation has a hard ceiling. Ninety-two percent give an AI agent less than 10 minutes before escalating to a person, and 46% say a bot that repeatedly misunderstood them is what drove the switch. Thirty-five percent often or always have to repeat themselves after the handoff, and about a third of those who escalated were moderately to extremely frustrated by the transition.

Where customers land after that is telling. Phone is the most preferred channel for every generation in the survey, and it outscores AI on satisfaction by a wide margin: 68% rated their most recent phone experience positively, against 46% for AI or chatbot interactions, the lowest of any channel. Leaders in Five9’s parallel decision-maker survey predicted that Gen Z would favor chatbots and messaging apps, yet Gen Z’s actual top three channels were phone, live chat and email.

Customers are also clear about which tasks they will hand over. More than a third are comfortable letting AI alone handle password resets and order status checks. For charge disputes, half want a human only, and for high-stakes escalations that rises to 60%. The routine work is available to automate. The judgment calls are not.

Leaders Think the Handoff Works. Customers Disagree.

The bigger operational risk is that leaders do not see the problem. CMSWire’s analysis of Five9’s Business Leaders CX Report, which surveyed 600 decision-makers alongside the consumer panel, found 96% of decision-makers say their organization preserves context effectively during AI-to-human handoffs. Yet 83% of consumers say they have had to repeat themselves after a transfer.

The same leaders acknowledge the mechanics that cause it. Thirty-seven percent say agents lack visibility into the prior AI interaction, 35% report technical transfer glitches and 34% report context or data lost when switching channels. Nine in ten report positive ROI across virtually every AI use case. When the dashboards say the program is working and the customers say it is not, the measures are probably counting containment rather than resolution.

Gartner’s own customer research points the same way. A survey of 3,566 B2B and B2C customers conducted in February and March found 87% say it is essential for companies using GenAI in service to provide an option to reach a human agent, while 50% say GenAI makes their interactions easier. Eric Keller, senior director analyst at Gartner, advised that “service leaders should not use GenAI as a mandatory first step for every issue.”

A Human-in-Reach Service Playbook

• Show the exit up front: Make the route to a person visible from the first AI interaction. Five9’s data shows that visibility alone moves trust from 26% to 55%, which is likely to raise AI usage rather than lower it.
• Escalate on a clock, not a complaint: With 92% of customers giving AI less than 10 minutes, set automatic escalation triggers on elapsed time and repeated misunderstanding instead of waiting for the customer to demand an agent.
• Route by task risk: Let AI own password resets and order status, where customers already accept it. Send disputes, refunds and high-stakes escalations to people by default.
• Pass the whole conversation: Give agents the AI transcript and customer record on the same screen at the moment of transfer. Thirty-seven percent of leaders admit agents cannot see the prior AI interaction.
• Measure what the customer felt: Track repeat-information rates, post-handoff satisfaction and resolution, not just containment. A 96% versus 83% perception gap is a measurement failure.
• Keep investing in the phone: Phone remains the preferred channel for every generation and the highest-rated one. Staff and equip it for the complex calls AI will increasingly send its way.

The lesson from this research is not that customers reject AI. Nearly half already find it useful, and many will let it handle routine work. What they reject is being trapped with it. Companies that keep a person visibly within reach will find customers more willing to try the bot first, and that is the only path to automation rates that hold up.