The Craft of a Real Apology
Customer service should sound human—and back its words with action.
I recently received a reply from Custom Ink after explaining that the experience had cost them my confidence, my future business, and my recommendations to clients.
The response told me the impact went beyond a single order. It repeated that losing my confidence as a creative professional and losing the referrals that came with it was serious. It even said trust has to be earned.
All the right sentiments. Nicely arranged. And somehow, it still felt as if nobody had really heard me.
Then came the sign-off: “Thank you again for choosing Custom Ink!”
I had just explained that I would not choose them again. That last line made the whole message feel less like a conversation and more like a system had found the “sympathetic customer response” button.
I can’t say for certain who wrote the email or how much AI was involved. I can say it reads as AI-washed: polished until the human details disappeared. And I can smell AI a mile away.
But this isn’t only about one company or one email. Customer service is often the one place where a person comes face to face with a company after something has gone wrong. That is exactly when the human touch matters.
Empathy isn’t an echo
Repeating a customer’s own words back to them doesn’t prove you listened. It can make the reply feel even more robotic, especially when those words are dressed up in formal sympathy.
A customer doesn’t need a company to restate the emotional impact of the problem. They need someone to understand what happened, answer the actual concern, and say what comes next. A response can be brief and still feel human if it is specific and honest.
A real response might say: “I reviewed your order and the messages. Here is what went wrong on our end, here is what I can do now, and here is what I’m doing to keep this from happening again.” Then the company has to follow through.
That’s not a magic apology formula. The point is that the reply answers the situation instead of performing concern.
When the words had to carry the weight
A much bigger example came in 1982, when seven people died after taking cyanide-laced Tylenol capsules in the Chicago area. The capsules had been tampered with after they left the factory. Johnson & Johnson could have limited its response to the affected lots. Instead, the company pulled Tylenol capsules from store shelves nationwide, urged people to use tablets, offered to replace capsules, and worked to bring the product back in tamper-resistant packaging. The Smithsonian’s National Museum of American History describes the decision as taking responsibility and pulling all capsules from stores nationwide. A contemporary Washington Post account reported that the company kept escalating its response and accepted the enormous cost of destroying and replacing the product.
The advertising matters here, too—but the words did not do the job by themselves. A TV spot produced by Compton Advertising featured McNeil’s medical director, Dr. Thomas Gates. He spoke directly about the tampering, what the company had already done, and what it would do next. One line was: “We will now work even harder to keep it.” The ad then told capsule users to use tablets and explained that capsules would return in tamper-resistant containers. Read UPI’s October 1982 report, which quotes portions of the ad’s spoken copy. You can also watch a Retro Report video about the Tylenol crisis and response, with archival footage.
Imagine if the spot had ended with, “Thank you for using Tylenol for your pain relief!” It would have sounded about as human as thanking a customer for choosing a company she had just told you she would never use again.
I haven’t found a source crediting an individual copywriter for that spot, so I won’t invent one. What we can see is the craft: the message named the danger, gave people useful instructions, explained the company’s actions, and made a promise the public could check against what happened next. The words had a real job to do.
Imagine replacing that with a generic, AI-polished paragraph: “We understand that recent events have affected your confidence. Your safety is our top priority, and we are committed to rebuilding trust.”
It sounds appropriately concerned. It also leaves out what a person urgently needs to know: Should I take the capsules? What is the company doing? What happens next?
AI could help write a clear version of the Tylenol message if it had the facts and a careful brief. A person still has to decide which facts belong, which promises are real, and what the audience needs first. That judgment is part of the writing. The craft of a well-written apology—or any message delivered at a moment of real concern—is an art.
Don’t wash the person out
AI can sort messages, find order details, and help prepare a first draft. Those are useful jobs. But when a customer asks for help, someone at the company has to take responsibility for the reply. They need to check whether it fits the situation, answers the question, and says something the company can actually stand behind.
A polished message can sound caring while doing none of the work of caring. It can reflect the customer’s frustration back in elegant language and still leave them with no sense that anyone took action.
Customer service should feel completely human—humanoid, even. Not a customer’s own words sent back in a corporate cardigan. Not a generic apology with the customer’s name dropped in. A real response from someone who understands there is a person on the other end of the email.
Thank you, Custom Ink, for giving me the idea for this blog post. I’ll be using this experience as an example.
For companies using AI in customer service: let it help with the work. Don’t let it wash the human being out of the reply.
Sources
Comments