80% of customer service will be automated. And 64% of customers would rather it wasn't.
The two predictions come from the same field of research and seem to contradict each other. They don't: taken together, they tell you what will separate businesses that use these systems well from those that use them to cut costs.
The prediction says that by 2029 AI agents will resolve 80% of common customer service issues on their own, cutting operational costs by 30% (Gartner, press release of 5 March 2025). The word “common” carries the whole sentence, and almost every quotation drops it.
The same field of research produces a second prediction that gets far less airtime: 64% of customers would prefer businesses not to use AI in customer service. Both are true, and they're about different things: the first about what the technology will be able to do, the second about how people react to the way it's used today.
The number that decides what to do isn't the prediction: it's the last hundred enquiries you received, sorted into three groups (same answer every time, needs a look in a system, needs judgement). The ratio between the three is the only figure that concerns your business.
The most important word is “common”
The prediction is almost always quoted in cut-down form, and the cut always falls in the same place. It doesn't say that 80% of customer service will be automated: it says that 80% of common issues will be. It's a specific category, and every business already knows what its own are: where are you, what time do you open, where's my order, does this product contain gluten, can I move my booking.
They're the questions someone is answering today for the thousandth time. They need no judgement, no negotiation, and no knowledge of the customer. They're also, almost always, most of the incoming volume by number and a clear minority of its value.
Put that way, the prediction becomes far less spectacular and far more useful: in a few years it will be odd to keep someone waiting for a question whose answer is always the same. It's less a revolution than a shift in the threshold, as happened with opening hours on websites or with booking without phoning.
The other half of the data
The same body of research that produces that prediction produces another, which gets far less airtime because it sells far worse: 64% of customers would prefer businesses not to use AI in customer service. And more: half of the businesses that will have cut staff counting on AI expect to rehire by 2027.
It's worth holding the two together, because both are true and they aren't about the same thing. The first describes what the technology will be able to do. The second describes how people react to the way it's used today. The 64% have no opinion on the architecture of agents: they have the memory of the last time they tried to reach a human being and couldn't.
The rejection isn't about automation, but about automation used as a wall rather than a door.
In practice, the difference comes down to three choices
They're three design decisions, all made before building, and all free:
- The way through to a person is always visible. Not hidden four steps down, not dependent on phrasing the question the right way. A system that makes it hard to talk to someone isn't saving money: it's shifting the cost onto the customer, and every customer notices.
- The system says what it is. No made-up names, no fake hesitations to seem human. Transparency here isn't just a matter of style: the AI Act is heading in the same direction, and a system built honestly today won't need rebuilding tomorrow.
- What it doesn't know, it doesn't make up. An assistant that answers only from the business's own documents, and hands over instead of improvising when a question falls outside its scope, gets far less wrong. The mechanism is called RAG, and its main job isn't to seem cleverer, but to know when to stay quiet.
This is a prediction, not a measurement: it describes a 2029 scenario set out in 2025, and it's the kind of statement that should be treated for what it is. Predictions about adoption rates have a documented history of optimism, and this one also mainly concerns organisations much larger than an Italian SME.
It has no value as a promise. Nobody can tell you that 80% of your enquiries can be automated, for the simple reason that they don't know what enquiries you get. It has value as a sign of direction, and as a prompt to do something you can do straight away: count how many of your last hundred enquiries got the same answer.
The count that replaces the prediction
Industry predictions help you decide whether to look at a topic. They don't help you decide what to do, because they describe an average of businesses that aren't yours. The number that decides what to do, in this case, takes an afternoon to get.
Take the last hundred enquiries you received, from any channel, and sort them into three groups: those you gave the same answer every time, those that needed someone to look something up in a system, and those that needed judgement. The ratio between the three groups is your figure, and it's the only one that counts.
If the first group is large, you have a problem that can be solved and you know its size. If the third group is large, the honest conclusion is that right now an automated customer service system isn't the first thing you should do, and anyone proposing one is doing so without having counted.
Questions and answers
Will 80% of customer service really be automated by 2029?
The prediction says something more precise: by 2029 AI agents will autonomously resolve 80% of common customer service issues, cutting operational costs by 30% (Gartner, press release of 5 March 2025). The word “common” carries the whole sentence, and almost every quotation drops it.
Common enquiries are a specific category every business recognises: where are you, what time do you open, where's my order, does this product contain gluten, can I move my booking. They make up most of the incoming volume by number and a clear minority of its value.
Is it true that customers don't want AI in customer service?
64% of customers would prefer businesses not to use AI in customer service, and half of the businesses that will have cut staff counting on AI expect to rehire by 2027.
But the rejection isn't about automation itself. The 64% have no opinion on the architecture of agents: they have the memory of the last time they tried to reach a human being and couldn't. It's about automation used as a wall rather than a door.
How do you build an automated assistant customers will accept?
Three design decisions, all made before building and all free. The way through to a person is always visible, not hidden four steps down. The system says what it is, with no made-up names or fake hesitations to seem human, which is also the direction the AI Act is heading in. What it doesn't know, it doesn't make up: it answers from the business's own documents, and when a question falls outside its scope, it hands over.
How do I know how many of my enquiries could be automated?
By counting them, because the prediction describes an average of businesses that aren't yours. Take the last hundred enquiries you received, from any channel, and sort them into three groups: those you gave the same answer every time, those that needed someone to look something up in a system, and those that needed judgement.
If the first group is large, you have a problem that can be solved and you know its size. If the third is large, the honest conclusion is that an automated customer service system isn't the first thing you should do.
Sources
- Gartner, Gartner Predicts Agentic AI Will Autonomously Resolve 80% of Common Customer Service Issues Without Human Intervention by 2029, press release of 5 March 2025. The source of: 80% of common issues and the 30% cut in operational costs.
- From the Gartner customer service surveys cited in coverage following the prediction: the 64% of customers who would prefer no AI in customer service, and the share of businesses expecting to rehire staff by 2027.
Count your enquiries before you automate them.
The Diagnostico asks about this too: what happens to an enquiry that arrives outside working hours, and how many times a week you rewrite the same information. Twenty questions, five minutes, no spam.