What good AI looks like to your customers

UK consumers are not hostile to AI. But they have clear, consistent rules about where it belongs and where it does not. New Moneypenny research of 2,000 UK consumers, conducted by Censuswide in June 2026, shows that people welcome AI for booking, scheduling and routine information. They want a human for anything sensitive, consequential or personal. The gap between those two positions varies significantly by sector and age and getting it wrong carries a real commercial cost.

• 40% of UK consumers are comfortable using AI for general information requests across healthcare, legal, care homes, automotive and GP services.
• 45% of consumers want no AI involvement at all when contacting a care home.
• 38% of consumers want no AI involvement in legal services — the highest sector-wide resistance.
• Only 13% of consumers are comfortable using AI to query an illness or medication with their GP.
• 47% of 25–34 year olds are comfortable using AI to call in sick; only 20% of over-55s say the same.

Most businesses approach AI the same way. They pick a tool, deploy it as broadly as possible, and assume customers will adapt.

Some get away with it. Others discover through complaints, churn, or silence, that their customers had already decided this was not the kind of business they wanted to deal with.

The difference between those two outcomes is rarely the technology. It is whether the business understood what its customers were actually comfortable with before it made the decision.

Until now, that data has been hard to come by. Moneypenny commissioned independent research from Censuswide, surveying 2,000 nationally representative UK consumers in June 2026. We asked them exactly which parts of business communication they would be comfortable handling with AI and which parts they would not. The results are specific, sector-by-sector, and in several cases, sharply at odds with how businesses are currently deploying AI.

New Moneypenny research involving 2,000 UK consumers shows that people are not hostile to AI, but they have clear rules about where it belongs. They welcome AI for booking, scheduling and general information. They want a human for anything sensitive, consequential or personal. And the gap between those two positions varies significantly by age and sector.

How are businesses using AI to improve customer experience?

The headline finding is not that customers reject AI. They do not. What they reject is AI in the wrong place, handling the wrong things.

Across every sector we looked at, including health and personal care, care homes, legal, automotive, and GP services, consumers showed a consistent willingness to use AI for certain types of interaction. General information, booking and scheduling, and routine administrative tasks all scored well.

But that willingness dropped sharply as interactions became more personal, more sensitive, or more consequential. A significant proportion of consumers, ranging from around 31% in care homes to 41% in automotive, said they would not want AI involved in any part of their communication with that type of business.

Related reading: 10 examples of AI in customer service.

What are customers comfortable using AI for?

Across all sectors, three types of interaction consistently attracted the highest comfort levels.

General information requests

Opening hours, visiting hours, surgery times. Around 40% of consumers across sectors were comfortable getting this from AI. It is factual, low-stakes, and does not require judgement or empathy.

Booking and scheduling

GP appointments (34%), care home visits (29%), automotive servicing or MOTs (34%). The task is transactional, the outcome is clear, and AI does not need to understand nuance to do it well.

Routine administrative tasks

Settling a bill, ordering a repeat prescription, completing a questionnaire. Lower than booking or general information, but meaningful. Customers are comfortable when the task is routine and getting it wrong would not cause significant harm.

Consumers value AI most when it delivers speed, convenience and efficiency without requiring emotional intelligence, professional judgement or personal reassurance.

Source: Moneypenny / Censuswide, 2,000 UK consumers, June 2026.

Where customers still want a human

Comfort levels fall quickly once interactions move beyond the routine.

Lowest-comfort interactions by sector

Source: Moneypenny / Censuswide consumer research, 2,000 UK consumers, June 2026.

These are not edge cases. In many scenarios, they represent the majority position. Businesses that treat AI as the answer to every customer interaction will find that in several sectors, more customers are uncomfortable than comfortable. Find out about he different types of answering services available here.

The generational divide businesses cannot ignore

The aggregate numbers matter. The generational breakdown matters more.

Younger consumers, particularly Gen Z (18 to 28) and Millennials (29 to 44), are significantly more comfortable with AI across every sector and every task type. In some scenarios, the gap between the most and least comfortable age groups is more than 20 percentage points.

Using AI to ring in sick to an employer

One scenario that reveals the generational gap most clearly.

47% of 25 to 34-year-olds are comfortable using AI to call in sick

20% of 55 and over arecomfortable using AI to call in sick

Source: Moneypenny / Censuswide consumer research, June 2026.

This matters for two reasons. First, your customer base is not homogeneous. A communications strategy built around the preferences of one generation will alienate another. If your customers skew older, as they do in many sectors including legal, healthcare, and financial services, treating AI as the default channel is a real commercial risk.

Second, expectations are shifting. The youngest consumers in this survey have grown up with AI-assisted interaction as the norm. As they move through life stages and take on more of the buying decisions, their expectations will become the dominant ones. Businesses that build good AI habits now will be far better placed to meet those expectations as they evolve.

What this means sector by sector

🏥 Health and personal care

Around 41% comfortable with AI for general information. Booking and changing appointments: 31%. "No parts" response: 31%. There is a clear role for AI in routine administration, but clinical or treatment-related queries (18%) are much more cautious territory.

🏠 Care homes

Visiting hours and general information: 40%. Booking visits: 29%. Checking on a loved one's wellbeing: 18%. Nearly half (45%) want no AI involvement at all. This sector requires particular care in how AI is positioned and deployed.

⚖ Legal

The highest sector-wide resistance. Nearly 4 in 10 want no AI involvement at all. Initial enquiries (29%), questionnaires (27%), and requesting updates (22%) have room for AI, but the overall approach must be human-first with AI in a supporting role.

🚗 Automotive

Booking a service or MOT is the clearest AI opportunity at 34%. Test drives: 24%. Buying or selling a vehicle attracted very low comfort scores. AI can own the operational layer, but should stay well clear of the sales conversation.

🩺 GP and doctor services

Appointment booking (34%) and finding opening hours (36%) are solid. Ordering repeat prescriptions (29%) is credible. Clinical queries, such as what a medication does or what a symptom might mean, drop to 13%. Customers want a qualified human for these.

When do customers want a human instead of AI?

Look across all of this data and a single principle emerges. Customers are drawing a line between two types of interaction.

Booking, information, scheduling, payment. Speed and accuracy matter more than empathy. AI can genuinely deliver a better experience than a phone queue or an unanswered email.

Clinical queries, legal matters, emotional conversations, high-value decisions. The cost of a wrong answer is high, context matters, and customers want to feel heard by a person.

The businesses that get AI right use it to handle the first category brilliantly, and use that to free up their people to excel at the second.

The research points to a clear conclusion: customers favour businesses that combine AI efficiency with human expertise. The most effective approach is to use AI for routine enquiries and administrative tasks, provide fast access to a human when conversations become more complex, clearly communicate when customers are interacting with AI, ensure escalation routes to a human are simple and immediate, and focus AI on improving customer experience rather than replacing human interaction.

What does good AI look like in a business context?

Based on the research, good AI in a business communication context has a few consistent characteristics.

✓ It handles the predictable, not the personal. The interactions that score highest for comfort are ones where the task is clear, the outcome is defined, and the customer knows exactly what they are getting.

✓ It knows when to hand over. A well-designed AI Receptionist that routes a legal query to an experienced team member in seconds is a better customer experience than either a slow human or an AI that tries to handle something it should not.

✓ It sounds like the business, not like a bot. Customers are more comfortable with AI when it reflects the brand's tone, understands the context of the service, and avoids robotic, templated responses.

✓ It does not leave people in the dark. Customers who are uncomfortable with AI in legal or clinical settings are often rejecting the idea of being unable to reach a real person. Making it clear that a human is always available changes how AI feels entirely.

✓ It creates data, not just answers. Good AI captures what was asked, how it was answered, and what was escalated. That information improves the AI over time and gives human teams the context they need when they step in.

The findings suggest that successful AI adoption is less about replacing people and more about redesigning customer journeys. For most organisations, the biggest opportunities lie in answering routine enquiries faster, managing bookings and appointments, handling repetitive administrative tasks, improving availability outside normal business hours, and supporting customer service teams during peak demand.

The research also highlights the importance of maintaining human support for conversations that require empathy, judgement, expertise or reassurance. The businesses most likely to succeed with AI will be those that understand where automation improves the customer experience and where people continue to make the greatest difference.

Key takeaways from the research

• UK consumers are not hostile to AI, but apply a clear distinction between routine tasks where AI is welcome and sensitive interactions where it is not
• The "no AI at all" response ranges from 31% in health and personal care to 45% in care homes
• Booking, scheduling and general information are the tasks most consistently suited to AI across all sectors and demographics
• Younger consumers (Gen Z and Millennials) are significantly more comfortable with AI than those aged 55 and over
• The businesses that get AI right use it to handle the predictable and free up their people for the personal

Frequently asked questions

• What did the Moneypenny AI research involve?

Moneypenny commissioned Censuswide to survey 2,000 nationally representative UK consumers in June 2026. Respondents were asked which parts of business communication they would be comfortable interacting with AI for, across five sectors: health and personal care, care homes, legal, automotive, and GP services.

• Which sector had the highest resistance to AI?

Legal services had the highest proportion of consumers (38%) who said they did not want AI involved in any part of their communication with a law firm.

• What are consumers most comfortable using AI for?

Consumers are most comfortable using AI for general information requests, appointment booking, scheduling and routine administrative tasks. These interactions are typically low-risk, predictable and focused on efficiency.

• What are consumers least comfortable using AI for?

Consumers are least comfortable using AI for legal matters, medical questions, wellbeing enquiries and high-value decisions where expertise, trust and human judgement play an important role.

• Does age affect AI comfort levels?

Yes, significantly. Younger consumers, particularly those aged 18 to 34, are considerably more comfortable with AI across every scenario. Consumers aged 55 and over show consistently higher resistance, often by 20 percentage points or more.

• Should businesses replace customer service teams with AI?

The research does not support a fully automated approach. Consumers consistently favour a blended model where AI handles routine tasks while people remain available for complex, sensitive or high-stakes conversations.

• How should businesses respond to this research?

Use AI where it performs well, on routine, predictable, low-stakes tasks, and keep people front and centre for interactions where context, judgement or empathy matter. Making sure customers can always reach a human is particularly important in sensitive sectors such as healthcare and legal.

• What is Moneypenny's approach to AI?

Moneypenny brings together real people and AI technology designed to work together rather than in opposition. The AI Receptionist handles routine and predictable call types, while a team of trained professionals steps in for more complex or sensitive conversations. The aim is to give every caller the right experience, not simply the fastest one.

‍

Research conducted by Censuswide on behalf of Moneypenny, June 2026. Sample: 2,000 nationally representative UK consumers.

‍