This is the 30% that gives you 5-star reviews

Written by Signe Bergström, July 2026
Interview with Max Riis Christensen, CEO MakesYouLocal

It's easy to answer in any language through a chatbot - but when an unhappy French customer calls, you need a native speaker on the other end. Someone who can listen to the problem with a real  understanding and actually say "je suis désolée.".

We sat down with our CEO, Max Riis Christensen, to talk about something we discuss a lot, and something most of you are probably weighing right now too: how much of your customer service can you automate before it starts costing you customers?

 

 

If you're wondering whether AI can replace customer service, the honest answer is no, not entirely, and not the part that matters most. AI can safely handle the majority of routine enquiries,  but the conversations where a customer has something real at stake still need a human. Here's why, according to Max.

The 30% of customer service you shouldn't automate

how much of your customer service can you automate before it starts costing you customers?

This is one of the question we get asked constantly, on stages, at conferences, in meetings with clients who are convinced the answer is "automate as much as possible."

It's also the one we have had some of our best conversations about over the years.

The instinct to chase a higher automation percentage makes sense.

It is  also exactly where a lot of companies quietly lose customers without noticing.

Done well, automation can handle up to 70% of customer enquiries, but the remaining 30% is the part many companies overlook once they have seen how "easy" it can be to automate.

The remaining 30%, and it's the part many companies overlook once they have fallen in love with the savings from automation, is where something is at stake for the customer. And at a time where one of the most frustrating things you can encounter is a chatbot asking "was this answer helpful?" when you just wrote "can I please speak to a REAL PERSON."

Easy questions like "where's my tracking code" works, but "I got sent the wrong item, and I need the right one before Saturday at my sisters birthday" or "I have tried to reach you to file a complaint, but I need some help".

A mistake has usually already happened by the time a customer reaches this kind of conversation. And sometimes a mistake needs an individual solution, not a standard reply. That's the real difference between the 70% and the 30%: it needs to be more nuanced.

A customer should be able to reach a human, when they are in need of a human.

 

"Customers don't dream of trying out a shiny new chatbot to get their problem solved," as Max puts it. They just want it to fixed right away. That's why this 30% deserves more attention than its size suggests. It's small in volume, but it's where loyalty is either won or quietly lost.

The easy 70%: fast, simple, and safe to automate

Let's move on to the part that makes the finance team very happy.

According to Max, you can comfortably automate somewhere between 50 and 70% of all customer service enquiries without harming a single customer relationship. Saving money and time, wuhu!

That covers the classic "where's my order" enquiry, tracking codes, return policies, and other standard questions. Nobody minds getting that answer from a bot instead of a person, as long as it's fast and correct.

Key insight

Delivery status is the top customer service question despite proactive measures like tracking updates and order confirmations. This is because delays, incomplete tracking, or ambiguous updates leave customers uncertain and seeking reassurance. Businesses must ensure updates are not only timely and accurate but also proactive when unexpected things happen.

Speed matters here. Customers have been treated well with fast responses over the past few years, Max points out. We expect a tracking code right now, not tomorrow in an email.

So automation is solving a problem people didn't know they had: a shrinking tolerance for waiting.

You should automate this part. These repetitative questions should be automated and 

Why a mistake, handled well, can win you a more loyal customer

This is where the service recovery paradox comes in, a concept Max brings up as the core reason human agents matter, a lot.

When something goes wrong and a person takes ownership and finds a solution, the resulting experience can end up stronger than a flawless, fully automated process. It's the difference between a one-star review and a four-star review that reads "they made a mistake, but they sorted it out to perfection."

 

Max touches the peak-end rule. In short, we don't remember an experience as an average of everything that happened. We remember the peak moment (good or bad) AND how it ended.

So if the ending is a person who owns the mistake and sorts it out properly, that's the memory that sticks, not the fact that something went wrong in the first place.

And we all know that reviews are one of the biggest drivers of trust. And now you know that the ending of the story is what gets written down. And that's where our headline earns its place, because those 30% are what stand between you and a 5-star review.

Losing that chance is expensive. Really expensive. A good conversation might cost you 20 kroner. Acquiring a new customer through marketing typically costs 200 to 300 kroner. The maths speaks for itself.

What this means for your business

If you're trying to figure out what should stay human versus what can be automated, the principle Max keeps coming back to is simple: use automation for what's repetitive and predictable. Use people for what's hard and really counts.

 

And give your people the mandate to solve the problem, instead of locking them into scripts they can't step outside of.

That's where trust, and the good reviews, will balance the 70% with the 30%.


3 quick answers with Max: