Is it just me or is AI in customer service making life miserable?
Last month I found myself in a familiar kind of modern purgatory: trying to get a straight answer from my mobile data provider. It started innocently enough on WhatsApp. For twenty minutes, the chatbot asked me for the same details in slightly different ways. My issue wasn’t on its drop-down menu, and the bot couldn’t, or wouldn’t, understand that.
So I gave up and dialed their support line. An automated voice system escalated me to a conversational AI assistant. Another thirty minutes slipped by. I recited my account number, listened to cheerful hold music, and was continually routed back to the main menu because the software couldn’t bridge the gap between a standard keyword and my actual problem. I was getting nowhere. After nearly an hour, frustrated, I hung up.
I’m an AI enthusiast. I use these tools daily and remain genuinely fascinated by all that they make possible. This is why AI in customer service bothers me. If companies adopt AI to improve customer service, why do so many of these interactions leave us fuming?
The AI trend is not going away. And some would rightfully argue that there are benefits to be had. The challenge is that technology decisions have consequences far beyond immediate cost savings. When efficiency is prioritized over human connection, the tool meant to help often becomes the barrier.
Lost in translation
AI is now the front door for customer support. For routine requests like tracking a package, resetting a password, or checking a balance, an AI chatbot works brilliantly.
The friction occurs when a customer’s reality doesn’t match the bot’s script. You type, “I need to speak to a person.” It asks you to describe your issue again. You simplify your wording, only to get another automated menu. The software is built to classify and clear tickets, so it cannot recognize frustration or nuance. When it can’t classify, it stalls.
This isn’t just an occasional glitch. Research shows a growing gap between corporate implementation and customer expectations:
- A Gartner survey of 5,728 customers, run in December 2023 found that 64% would prefer companies not use AI in customer service. Their top concern, at 60%, was that it would make reaching an actual human being significantly harder.
- The business fallout is tangible, too. A survey of 1026 US shoppers in February 2025 by e-commerce marketing automation platform Omnisend found that 39% of online shoppers had abandoned a purchase mid-way due to frustrating AI interactions, including poor chatbot experiences and inaccurate recommendations.
It makes you pause. If the attempt at efficiency drives customers away before a single purchase, what are those savings really costing us?
Klarna: AI in its proper place
In February 2024, Klarna shared that its AI assistant had taken over two-thirds of its customer service chats, 2.3 million in its first month. It was widely cited as a headline moment for automated support. Yet, in May 2025, Klarna began hiring human agents again. Its CEO acknowledged that the emphasis on automation and cost reduction had lowered overall service quality.
Klarna didn’t scrap AI. Instead, they refined where it belonged. They kept AI for speed and volume and brought people back for complex and sensitive cases.
That split is the key takeaway. Automated systems excel at speed and volume, but struggle with nuance, empathy, and complex problem-solving. Knowing which task to automate makes all the difference.
What Hospitality shows us
Hotels are people-intensive, operationally demanding, and run on tight margins. Naturally, hoteliers have good reason to adopt smart technology and streamline operations.
At the same time, hospitality rests on something fundamental to human nature: how a place makes us feel. We all have a basic human need for recognition, welcome, and ease when we enter an unfamiliar space.
A comprehensive study conducted in 2026 by Mews, a hotel management platform, captured this dynamic across two perspectives:
- Hoteliers: Out of 500+ properties surveyed, 98% had adopted AI in some capacity over the previous six months. Yet, 59% of hoteliers strongly believed that the front desk welcome and check-in should remain human-led, and that view was strongest at properties already using AI extensively.
- Guests: In a parallel survey of 3,250 guests across eight countries, respondents expressed a clear preference for human interaction across everyday touchpoints. Human presence was valued highest in bar and restaurant service (72%), the front-desk welcome (66%), and initial check-in (58%).
This doesn’t mean travelers are anti-tech. In fact, 89% preferred a hybrid stay or AI working in the background. 59% wanted AI to handle repetitive tasks, specifically so staff would have more time to connect with guests.
Perhaps most telling, 75% of guests noted that true hospitality depends on human staff, regardless of how advanced the underlying tech may be.
Both Omnisend and Mews are technology companies, so take their findings with a grain of salt. But their findings still broadly reflect the Gartner data.
A simple test: chore or concern
AI in customer service should not be a foregone conclusion. In fact, customers can tell us exactly where AI belongs. Sort every customer interaction into two buckets:
- Chore – resetting a PIN, tracking a parcel, confirming a booking – if it is routine, automate it.
- Concern – the customer is nervous, anxious, confused, or has a problem not listed in the drop-down menu – get a human to solve their problem quickly.
Think of an express checkout kiosk and a concierge. A kiosk is perfectly acceptable, and may even be quicker, after a routine stay. But if your flight is cancelled, you want to speak to a concierge. Trouble starts when the hotel starts relying on the kiosk for more intricate requests.
Customers accept AI for functional tasks, but resist it for personal decisions and concerns. The Mews data confirms this.
The humans behind the screens
Customers notice how organizations treat the people who work for them. The Mews survey also noted that 67% of guests expressed concern about service staff losing their livelihoods to automation. When a process feels stripped down solely to cut headcount, it subtly alters how we view the brand.
There’s also a practical question of mentorship. If entry-level service roles are completely automated away, where does the next generation of leaders learn the art of reading a room, de-escalating a difficult moment, or showing empathy? Nobody learns those skills from a manual or a prompt. They are internalized through daily practice alongside experienced peers.
Three questions to ask before you automate
AI is at its best when it empowers people to serve other people better. Before you automate, ask yourself:
- Can a customer reach a person in two steps or fewer?
- Does the bot pass the full conversation to the agent, so nobody has to repeat themselves?
- Is anyone measuring where customers give up?
Most of us are glad to let an automated tool handle a quick confirmation or answer a straightforward question. We don’t need a deep conversation with a human to reset a PIN. Technology can organize data, clear queues, and speed up routine logistics. But when an issue grows complex, or when a customer feels anxious, frustrated, or simply stuck, the ability to reach a thoughtful person shouldn’t feel like winning a prize at the end of a maze.
Before the next rollout, perhaps a more useful question an organization can ask is:
Does this choice help us serve our people better? Or are we just making it harder for them to reach us?
At Bento Coach, we train teams in the skills no bot can replace: reading a room, easing tension, and showing care. Our Customer Service training programs and Train-the-Trainer workshops build those skills. Contact us to learn more.
Hi, I’m Ripudaman, a coach, trainer, and facilitator with experience in customer service and people development across hospitality, QSR, BPO, and retail industry. I help organizations build high-performing cultures through bespoke leadership and management development programs, customer service training, and executive coaching. I am passionate about developing leaders and teams who deliver outstanding results while creating memorable customer experiences. Connect with me on LinkedIn.