Gartner: 87% of Customers Demand Human Agent Access in GenAI Service

What happened

A new Gartner survey has found that 87% of customers say companies deploying generative AI in customer service must still provide access to a human agent. The research, released by Gartner, signals that while organisations are accelerating GenAI adoption across service channels, customers are drawing a clear line: automation is acceptable, but the removal of human escalation paths is not.

The findings arrive at a moment when many enterprises are actively replacing or supplementing frontline service roles with AI-powered chatbots, virtual assistants and automated resolution tools. Gartner's data suggests that the pace of that shift may be outrunning customer readiness — and that organisations treating GenAI as a cost-reduction lever rather than a service-enhancement tool risk a measurable erosion of trust.

Why it matters

For customer experience leaders, this research reframes the GenAI deployment question. The issue is not whether to automate — most customers have accepted that AI will play a role in service interactions — but whether the architecture of that automation preserves meaningful human access. In behavioural economics terms, what customers are expressing is a demand for optionality: the knowledge that a human is reachable, even if never actually used, materially changes how safe and respected a customer feels within a service journey. Removing that option, even when the AI performs well, triggers a sense of being trapped — a well-documented driver of dissatisfaction and churn.

For service designers, the implication is structural. Human escalation cannot be an afterthought bolted onto an AI-first model; it must be a visible, low-friction feature of the experience from the outset. Organisations that bury escalation options behind multiple automated steps, or that eliminate them entirely in the name of efficiency, are likely to find that short-term cost savings come at the expense of long-term loyalty.

By the numbers

• 87% of customers surveyed by Gartner say access to a human agent is a requirement when a company uses GenAI for customer service.

The Renascence take

The headline figure is striking, but the more important insight sits underneath it: customers are not rejecting AI — they are rejecting the feeling of being handed off to a system with no exit. That is a design failure, not a technology failure, and it is one that many organisations are currently building into their GenAI rollouts at scale.

What most operators will miss here is that the 87% figure is not a vote against automation — it is a vote against powerlessness. The behavioural principle at work is perceived control: customers tolerate friction, waiting and even imperfect resolution far better when they believe they could have chosen differently. Smart service designers should treat human escalation not as a fallback cost but as a trust signal — one worth making prominent, even when the AI is performing well. The question to ask in every GenAI service deployment is not "can we remove the human?" but "does the customer know the human is there if they need one?"