More than eight in ten UK adults would rather deal with a person than an AI agent when contacting a business, according to a survey that raises questions about how quickly companies are automating customer service.
The 2026 UK customer service survey from contact-centre software supplier TCN questioned 1,000 adults. It found that 84% preferred speaking to a human agent, against 8% who preferred AI.
That does not mean customers reject all automation. The same research suggests many will accept AI for routine enquiries, provided it does not stand between them and someone who can solve a difficult problem.
What the TCN survey found
TCN’s published executive summary gives the clearest indication of where people draw the line.
- 84% preferred talking to a human agent, compared with 8% choosing an AI agent.
- 33% said they immediately try to reach a human when they encounter an online chatbot while resolving a customer service issue.
- 37% said AI was helpful when contacting a business, while 29% found it unhelpful.
- 35% would be comfortable with AI directing them to the correct department, but just 9% were comfortable with it collecting payment information.
- 53% said they did not trust providing personal data to an AI agent.
Speed and a satisfactory outcome also matter. In the same survey, 38% rated solving an issue quickly among the most important parts of customer service, while 36% highlighted being able to reach a live agent easily.
Customers are not necessarily against AI
Separate UK research points to a similar distinction. In July, Deloitte reported that 61% of consumers were less inclined to engage with customer service if they knew they were talking to a generative AI assistant. But the answer changed depending on the task: 30% were comfortable using AI to cancel an order, against 11% when making a complaint.
The communications regulator Ofcom said in September that some people welcome AI because it lets them avoid an unnecessary conversation, while others worry about empathy, errors and difficulty reaching a person. Its research focused on phone, broadband and pay-TV services, so it should not be read as a verdict on every industry.
The dividing line appears to be less about whether a computer is involved and more about whether the customer is still in control of the process.
Businesses are spending more on automation
While consumers express reservations, investment is moving ahead. In an international survey of 199 service and support leaders, Gartner reported in August that spending on AI had risen by 38%, compared with a 2% rise in overall customer service budgets.
That is a measure of the organisations Gartner surveyed, not a verified increase across all UK businesses. It nevertheless illustrates the commercial pressure to introduce AI while keeping service costs under control.
Another Gartner survey of 3,566 business and consumer customers found that 50% believed generative AI made interactions easier, but 87% considered access to a human essential. Gartner’s advice was to avoid forcing customers through unsuccessful automated exchanges before allowing them to speak to someone.
What protections do UK customers have?
In guidance published on 9 March 2026, the Competition and Markets Authority reminded businesses that consumer law applies whether a customer is dealing with an employee or an AI agent. Companies remain responsible for the information and actions of AI systems they use, including systems supplied by a third party.
The CMA also advised businesses to make clear when customers are dealing with an AI agent where failing to do so could mislead them. That does not create a blanket legal right to demand a human representative for every query, although particular sectors have additional requirements.
Ofcom has separately warned telecoms providers against using AI as a gatekeeper that obstructs access to required support channels. It also said it had found limited evidence of systemic harm to telecoms customers from AI so far.
The real test is whether the problem gets solved
A useful automated assistant can answer an opening-hours question, find an order or direct a call without leaving anybody on hold. A poorly designed one can add another obstacle for someone trying to challenge a bill, explain a sensitive issue or correct a mistake.
For businesses, the choice is not simply between people and technology. It is whether AI helps staff resolve problems or becomes a substitute for being reachable. The surveys suggest customers can tell the difference.
About the research
The 84% and 8% figures come from question 21 of TCN’s UK executive summary, which identifies a panel of 1,000 adults. TCN also published a shorter online summary reporting 81.2% and 9.65% for a similar human-versus-AI comparison. The reason for that difference is not explained in the material reviewed. These are reported survey preferences, not independently measured service outcomes.
Sources
- TCN, State of Customer Service 2026: UK executive summary (PDF)
- Deloitte, Digital Consumer Trends, 6 July 2026
- Ofcom, AI and telecoms customer experiences, 25 September 2026
- Gartner, customer access to human agents, 4 August 2026
- Gartner, AI customer service spending, 26 August 2026
- CMA, complying with consumer law when using AI agents, 9 March 2026
