Difficulty explaining their situation
Gap +20pp
AI comfort gap
06The AI comfort gap
AI isn’t a person, a team, or a customer experience strategy, it’s a tool. In customer contact, it’s only as strong as the people and processes wrapped around it.
That’s why the human safety net isn’t a nice-to-have, it’s intrinsic to the proposition.
The real opportunity is pairing AI with your colleagues’ judgement, empathy, and brand understanding so every interaction stays intuitive, flexible, and genuinely human when it matters most.
- Only 8% of consumers say they have no concerns about businesses using AI in customer service
- 59% feel comfortable with an AI receptionist that answers their call promptly
- 68% feel comfortable when AI can hand off to a human at any point
Customers are more comfortable with AI than headlines suggest, but that comfort has limits. It holds up well for routine tasks and starts to wobble when the stakes rise.
Where consumers are most comfortable:
Where comfort falls away:
The highest comfort score in the entire dataset isn't a task at all, it's a guarantee. When customers know they can be transferred to a real person at any point if needed, comfort is at its highest. That's the clearest signal that customers don't want AI to control the conversation. They want it to speed things up.
The perception gap: Businesses think customers are more ready than they are
When we tested specific AI receptionist scenarios, we asked consumers how comfortable they would feel with AI handling each task. We asked businesses how comfortable they believe customers would feel in the same situations. The gap shows the difference between customer reality and business assumption.
This is not about whether AI exists. It is about where customers are genuinely ready for it, and where they’re more cautious.
- UK Consumers
- Business assumption
Customer comfort vs business belief: AI receptionist scenarios
% saying they would feel comfortable
Books an appointment or reservation
+22pp
Answers simple questions or FAQs
+22pp
Answers the call promptly
+18pp
Handles a complaint
+18pp
Handles urgent or sensitive matters
+14pp
Age amplifies the risk
Concerns rise with age, especially around being able to explain what's needed and avoid dead ends. If a meaningful share of your audience is older, an AI experience without a clear human escalation is higher risk.
How concern grows with age:
Getting stuck in a loop
Gap +17pp
AI causing mistakes
Gap +12pp
Complex issues not handled properly
Gap +12pp
This indicates customers are more open to AI when it speeds things up, but they want to know they can reach a real person when needed.
The internal divide that can derail AI rollout
The comfort gap isn’t only between businesses and customers. It shows up inside organisations too, which is often why AI programmes stall, fragment, or get pushed live before teams have agreed how to handle the tricky, non-standard calls.
On the question of whether customers would feel comfortable with AI handling complaints, belief varies dramatically by department.
Belief that customers would be comfortable with AI handling complaints:
0%
IT
0%
Customer Service
0%
Sales
If teams aren’t able to align on customer readiness, they won’t align on where AI should sit in the customer journey, when it should escalate, or what “good” looks like in practice.
Treat AI as a front door, not a gatekeeper. Answer fast, route clearly, and always keep the route to a person open.
Start where comfort is already highest. Order updates, FAQs, and bookings are the strongest fit. Complaints and urgent matters need a human available, not an automated response.
Design the human hand-off before you automate anything else. Make it obvious, fast, and easy to find. For older customers especially, it’s not an edge case.
Be upfront about what customers are talking to. Disclose when it’s AI, be clear about what it can and can’t do, and treat data privacy as a front-of-mind concern, not small print.
Align internally before you scale. IT, Customer Service, and Sales often see customer readiness very differently. Close that gap before it becomes an operational problem.
Pete Hanlon
Chief Technology Officer
Moneypenny