AI in Customer Service: Getting the Balance Right Between Automation and Empathy

Think of the last time you asked for help online. Chances are you were greeted by a chatbot before speaking to a person. That’s the new normal. Retailers, airlines, and hotels all push AI to the front of the queue because it’s fast, consistent, and always awake.

And customers? They’re divided. People do like speed. But time and again — whether we’re speaking with a frustrated hotel guest or an ecommerce manager we support — the reality is simple: customers want to feel understood, not just processed.

One stat reinforces this clearly: PwC found 59% of customers believe companies have lost touch with the human side of service (PwC CX, 2018).

That’s the tension we’re talking about here. At Propeller Digital, we’ve built digital journeys for over 20 years across hospitality, travel and retail. What we’ve learned is that AI is powerful — but only when its role is framed correctly.

The business case makes sense:

• Customers expect 24/7 service.
• Volumes are bigger than ever, particularly post‑pandemic in ecommerce and travel.
• Leadership teams watch costs carefully.

AI ticks all three boxes. It’s tireless, it answers instantly, and it does the repetitive admin without complaint.

Brands also see opportunity in personalisation. New Epsilon research found 80% of people are more likely to buy from companies offering tailored suggestions. AI makes that scale possible.

But speed and data aren’t the full story. Without human judgement woven in, AI risks alienating the very people it’s meant to serve.

Picture a traveller stuck in a foreign airport at midnight. The bot can tell them the flight is delayed. What it can’t do is empathise, calm, and think creatively about solutions.

We’ve seen this in hospitality too. A couple’s honeymoon request falls flat when the chatbot misses the sentiment and just lists upgrade prices. It’s tone deaf rather than delightful.

How to solve it: Use AI as a front filter for factual replies, but route emotional‑charged queries fast. Create training phrases that flag words like “anniversary”, “angry”, or “urgent” and trigger human handover.

AI thrives when the question is sharp and single. Problems start the moment queries are layered: “Can I change the order, apply my voucher, and reroute delivery?”

What happens next? Customers get stuck in a polite but maddening loop of “I didn’t quite understand.” We’ve seen retailers lose customers not from slow response — but from bots locking customers into dead‑end flows.

How to solve it: Design what we call “escape ramps”. Bots admit their limits: “This looks a bit complex — connecting you to my colleague so you don’t repeat yourself.” That saves frustration and builds trust.

Customers don’t just want quick answers; they want to know what you’re doing with their data. AI systems often pull purchase history, browsing patterns, or previous destinations. If customers weren’t told about that, it feels creepy.

A study found 85% of consumers want more transparency over how data is collected and used.

How to solve it: Be straight. Label AI agents clearly (“I’m Bot Ava, here to help”), explain when and why data is being used, and link to privacy in plain English. Clarity equals credibility.

We coach brands to expect one thing: AI will fail sometimes. Heavy accents, slang (“my parcel’s knackered”), or stress in tone can derail bots. We’ve seen fashion brands lose conversions when refund bots lock up, or hotels scramble because a chatbot mishandled adjoining room requests.

A research showed only 17% of companies say their bots can resolve customer issues without human support.

How to solve it: Stop seeing handoffs as failure. Monitor bot “drop outs,” analyse common sticking points, and equip human staff with the conversation history so customers don’t repeat themselves when they switch channel.

The best service experiences blend automation and people:

• AI clears the repetitive noise — order checks, receipts, FAQs.
• Humans resolve the meaningful — frustration, loyalty requests, complex changes.
• AI supports humans — surfacing background details so calls start informed.
• Teams get trained — not just on empathy, but on how to use AI cues effectively.

That balance lets companies cut volume without cutting heart.

• Hotels and restaurants: Guests like bots for “what time’s breakfast?” When a wedding party’s room isn’t ready, they need a person who can fix problems with care.
• Online retail: Bots are great on tracking parcels. Wrong size at Christmas time? That’s where human reassurance matters more than logistics.
• Airlines and travel: Bots update on boarding times. Ground staff resolve the emotional fallout of cancellations.

1. Start low‑risk. Use bots where failure has low emotional cost (e.g. address lookups).
2. Be transparent. No one wins when customers guess: label AI clearly and explain data usage.
3. Measure relentlessly. Track handoff rates, customer satisfaction, and iterate.

We’ve applied this approach with clients across hospitality and ecommerce, helping them cut inbound volume by up to 30% without denting experience.

Here’s the blunt truth: people don’t fall in love with your chatbot. They trust your brand because of how your people make them feel.

AI in customer service is here to stay. Done well, it clears background noise so staff can focus on what matters. Done badly, it turns customer care into customer churn.

At Propeller Digital, we’ve seen both sides. Our approach? Keep humans at the centre, let automation amplify, and create experiences that customers actually remember. That balance is where brands win.