54% of Customers Just Proved Your AI Chatbot Is Losing You Money: The Holiday Shopping Report

Okay, so basically the holiday shopping numbers just dropped, and they're wild. Not in a good way.

Liveops surveyed 1,000 U.S. shoppers on December 2, 2025, asking about their AI customer service experiences during the holiday rush. Here's the thing - 78% of customers used AI or automation for customer service. Sounds like a massive win for AI, right? Wrong.

Only 29% said AI actually improved their experience. And get this - 54% said human agents are still doing better than AI. That's not a typo. More than half of your customers just told you they'd rather talk to a person.

The Math That Should Scare You

Let me give you an example with real numbers from recent research compiled in November 2025.

The average small business loses $126,360 per year from missed calls. Not slow response times. Not bad service. Just... missed calls. Calls that go straight to voicemail because nobody's there to answer.

Now here's where it gets interesting. When you miss a call, 85% of people won't call back. They just move on to your competitor. And if they get voicemail? 80% would rather contact a competitor than leave you a message.

Let's do the back-of-the-envelope math. Say you run an HVAC company. Average service call is $450. You miss just two emergency calls per week because it's after hours or your guys are in the field. That's $46,000 in annual revenue just... gone. And you probably don't even know it's happening.

What's Actually Going Wrong with AI Customer Service

So basically, companies spent 2025 throwing AI at customer service like it's the answer to everything. But recent data from Twilio's November 2025 survey shows a 31-point gap between what businesses think is happening and reality.

90% of business leaders believe their customers are satisfied with AI. Only 59% of customers actually are. That's a massive disconnect.

Here's what customers are actually experiencing right now:

The Loop Problem: 40% of customers get stuck in repetitive loops where the chatbot asks the same question over and over. You explain your problem, the AI responds by asking you to explain your problem again. It's designed to frustrate you until you get tired and give up.

The Comprehension Fail: Recent December 2025 data shows 45% said AI failed to understand their problem. The chatbot's inability to understand context is what customers find most annoying.

The Resolution Gap: Only 35% say a chatbot usually solves their problem effectively. Meanwhile, 85% say their problems typically need a live human to resolve.

Not exactly a ringing endorsement.

The Cautionary Tales (These Actually Happened)

Air Canada - February 2024: Their chatbot told a customer he could apply for a bereavement discount within 90 days after his flight. He booked based on that advice. Air Canada later denied the refund, claiming the chatbot was wrong. A tribunal ordered them to pay $650.88 plus fees. The ruling? Companies are liable for what their AI tells customers. Full stop.

DPD - January 2024: After a system update, their chatbot started swearing at customers, wrote poems about how awful DPD is, and told customers "I would never recommend them to anyone." The post went viral - 1.7 million views. Massive brand damage from one system update they didn't test properly.

Cursor AI - April 2025: Their AI support agent started telling developers they were limited to one device per subscription due to "security features" - a policy that didn't exist. Mass cancellations followed. The AI just made stuff up.

What Actually Works (And the Numbers to Prove It)

Here's the thing - AI isn't the enemy. Bad AI is the enemy.

Bank of America's Erica handled 3 billion interactions with a 98% problem resolution rate. Virgin Money's Redi hit 94% customer satisfaction. What's the difference?

They use AI to augment humans, not replace them. Clear escalation paths. When the AI hits its limit, it hands off to a human. No loops. No dead ends. No customers screaming "REPRESENTATIVE!" at their phone for 15 minutes.

And yes, that's real - as of December 2025, nearly 80% of consumers admit to screaming "Representative!" at automated phone systems. 77% have looked up hacks online to bypass AI and reach humans.

Your customers are literally Googling "how to avoid your chatbot." Let that sink in.

The Opportunity for Small Businesses

So basically, while big companies are fumbling AI implementations, there's a wedge for small businesses that get it right.

The play isn't AI-only. The play is AI that knows when to loop in humans. 24/7 availability for the simple stuff - booking appointments, answering basic questions, capturing leads. But the second it gets complex? Human takes over.

Recent case studies from 2025 show home service companies going from 40% booking rates with traditional answering services to 95% with AI that actually works. That's a 137.5% improvement in booking rate.

For a plumbing company missing two emergency calls a week at $450 per ticket, that's $46,000 in captured revenue per year. The AI service costs under $75/month. A human receptionist costs $3,100/month. Do the math.

What December 2025 Data Tells Us About 2026

Only 17% of shoppers want companies to increase their use of AI next year. 32% want less AI in customer service.

After experiencing the first fully AI-powered holiday shopping season, customers aren't asking for more automation. They're asking for better human contact, with AI handling only the simple stuff and getting out of the way for everything else.

The brands that win in 2026 won't be the ones with the most AI. They'll be the ones with the smartest AI - AI that knows its limits, escalates appropriately, and treats customer service like what it actually is: a chance to build trust, not cut costs.

The Bottom Line

Distribution is everything. But if your distribution is bringing customers to a chatbot that can't help them, you're just efficiently disappointing more people.

The 2025 holiday data is clear: AI customer service is a revenue problem disguised as a cost-saving solution. 54% of your customers just told you they prefer humans. Maybe it's time to listen.

Get the foundation right - AI that captures every lead, answers the simple stuff instantly, and seamlessly hands off to humans when things get real. That's not playing startup. That's get revenue quick.

Learn how Julya's AI handles calls 24/7 while knowing exactly when to bring in humans