British consumers collectively spend more than 445 million hours a year waiting on hold with customer service departments, equivalent to 56 million lost working days, according to new research from ServiceNow conducted with ThoughtLab.
The study, titled The CX Shift: Customer Expectations in the AI Era, surveyed 34,000 executives, service representatives and consumers globally, with 2,465 respondents in the UK. It found that poor customer service costs each British consumer the equivalent of 9.7 hours a year spent navigating long wait times, repeating information, or dealing with slow systems. Nearly half (41%) rated current service as average, poor, or terrible, and 53% said they would switch to a competitor after a single bad experience.
The gap between AI investment and customer outcomes is a recurring thread in the data. Almost half of UK consumers (49%) said AI had improved customer service, and 56% credited it with better after-hours support. But 45% said chatbots still fail to understand their queries, and 51% cited lack of empathy as their top frustration. There is a disconnect between what customers want and what executives prioritise: empathy ranks as a top concern for 51% of UK customers but only 20% of executives. Similarly, 49% of customers are frustrated by being transferred between departments, while just 39% of executives recognise this as a significant issue.
Consumers across EMEA are losing entire working days to service experiences that should take minutes. The root cause isn’t a lack of AI investment — it’s that most CRM systems were built to record interactions, not resolve them.
Part of the problem sits behind the scenes. UK service agents spend just 49% of their working week on customer issues, with the remainder absorbed by administration and system-hopping. Seventy-nine per cent must log into three to five systems to resolve a single query, and 37% cite inconsistent customer data as a major challenge. Fewer than half of UK organisations (42%) have integrated data across silos into a single source of truth, and only 22% have enterprise-wide AI strategies that cross departmental boundaries.
The research was conducted between September and October 2025 in accordance with ISO 20252 standards, spanning 18 countries and eight industries.
