A young lawyer once told me she preferred not to deal with humans.
At the time, I found it funny. Perhaps even understandable. Anyone who has worked with clients, customers, bosses, suppliers or service counters will know that human beings can be wonderfully complicated creatures.
But lately, I have been thinking about the other side of that statement.
What happens when we need a human being — and the system gives us an app, a chatbot or a help page instead?
I got into a Grab car recently and noticed a handphone on the seat. The driver, flustered, remembered his last passengers — a woman and her elderly mother, whom he had fetched from a clinic.
He took a small detour to check if they were still nearby. They were not. He said he would return after dropping me off.
What struck me was not just his effort. It was what he said next: there used to be a support line he could call for situations like this. Now everything is automated. He did not know who to contact.
The story ended well. The phone owner eventually called.
But what if she had not?
The second story did not end as cleanly.
A travel app showed a limited-time offer — last two rooms, timer running. After I entered my bank details, the cursor kept spinning. Uncertain whether the payment had gone through, I tried again.
It had.
I had made a double booking.
I immediately tried to cancel one. The bot said it could be cancelled but not refunded — policy. When I tried to explain what had happened, it replied: “I do not understand your issue.”
A bot acting like a bot is forgivable.
The next day, I contacted the app company, the hotel and the bank.
The app company said it would look into the matter and check with the hotel. The bank said it could not do anything as the issue was between me and the app company. The hotel understood perfectly, but said the app company would have to handle it.
About a week later, the app company sent an email stating that it would cancel the booking because I had changed my mind. I replied that I had not changed my mind, and explained once again what had happened.
Another email, signed by another person, arrived in my inbox saying he would look into it. Then came a third email, from yet another person, saying the booking could be cancelled but the money would remain in the app for me to use on another booking. In other words, the booking could be cancelled, but the payment would not be returned to me as a refund.
I was not cancelling the trip. I was requesting that one unnecessary booking be reversed because the booking process had failed at a critical moment. There is a difference between offering store credit and taking responsibility for what went wrong.
That, to me, is the real problem. Not automation itself, but automation without human accountability.
Apps are useful when everything goes according to script. But when the script breaks, someone still has to take ownership, exercise judgement and decide what is fair.
A company can spend heavily on branding and promotions to attract customers. But when something goes wrong, can it also respond with basic goodwill instead of hiding behind process?
This is the gap that automation has not closed.
Not routine transactions. Apps handle those well enough.
But moments that require judgement, context and the ability to say: “I understand what happened, and I am going to help make this right.”
Is automation useful? Of course.
Until something unexpected surfaces — something that needs empathy, discretion or simply a person willing to listen.
Apps can be automated. Processes can be automated. But accountability cannot be automated away.
Some problems still need a human voice. The irony is that the more systems try to remove that voice, the more valuable it becomes.
Note: Images used in this post are AI-generated illustrations created with ChatGPT.